Joining the Riskgaming podcast this week is Bill Gurley, a legendary venture capitalist who for more than two decades invested at Benchmark in such defining startups as Uber and Zillow. He’s just published a new book titled, Runnin’ Down a Dream: How to Thrive in a Career You Actually Love. It’s based on his own personal story of how to avoid career regret, of sticking with a job way past the point of boredom.

He’s joined in conversation by Lux’s own Josh Wolfe. The two talk about the book as well as the intersecting points of their long friendship and careers.

Josh Wolfe:

Let’s start at the beginning. You’ve said that, after school, you landed a job, I think, at a famous tech company, and you were surprisingly bored. What was that moment of realization like, and how close did you come to just staying on that conveyor belt?

Bill Gurley:

I don’t think there was much possibility of that. I fell in love with computers at a young age, like many people in our industry did. So I went and got a computer science degree, and I was blown away. Writing software in college was so much fun. The fact that you could make stuff that would actually go do things was, I thought, magic.

But I got out and was in the world and working on product cycles. And we started our third product, which looked like the second and looked like the first.

In the meantime, I had started reading industry trade magazines. I had read one on Wall Street by Peter Lynch, and I had opened a trading account and I was trading stocks as a hobby. I wasn’t doing it as my main gig. But one of the things I talk about in the book is that maybe your hobby should be your main gig. And so, I had these other things going on in my brain that seemed more interesting than what I was doing.

It is important to highlight that engineering may be someone else’s passion. I just came to a realization it wasn’t mine. I did this exercise, which I only reflected on as I wrote the book. I said to myself, “Do I see myself doing this 30 years from now?”

I didn’t know what I wanted to be doing, but I knew it wasn’t what I was doing. So I went to business school, which is an easy place to re-pot from. I thought about venture while I was there. I knocked on doors and it didn’t seem like there was any reasonable path to venture as a business school student. I started reading a lot of Fortune and Forbes articles about these gentlemen at Goldman Sachs — their tech team back at the time was incredible. Their sell-side analysts were being asked the questions about these industries I was most fascinated by. I loved my corporate strategy class. I thought maybe I could be one of those guys.

I went and knocked on doors here in New York City, and met the Goldman team. Even though I was a silly first year student at University of Texas, they let me in the door and I eventually got a job as a sell-side analyst.

That’s where I became even more aware of what was going on in Silicon Valley, the young companies that were disrupting things. I had that same discussion with myself as a sell-side analyst. About two, three years in, I walked around the office late one night about 10 P.M. I paused in front of each of the corner offices — the offices of people who were lifers in their role — and thought about, “Is that what I want?” And again, I got to “no” pretty quick and started moving in a different direction.

So those experiences helped paved the way for where I would spend a quarter-century.

Josh Wolfe:

Right. Your book itself was a decade-long project. You’ve said it started in part when you noticed that there was this pattern across three different biographies: of a restaurateur, a basketball coach and a folk singer. You connected these threads.

Bill Gurley:

As a blogger, one of the things I tried to do is tear apart some new thing that’s happening in venture and create a framework for understanding it. And so, when I finished that third biography, I had this aha moment. I was like, look, these people are all in very different fields. They’re all in fields your parents would probably tell you not to go into. They all started on the very bottom rung and they all did a lot of similar things. They used similar processes and principles that could be borrowed by anyone.

I got asked to speak at my MBA school and I asked, “Can I do this presentation?” They posted it on the internet. Some people noticed. One of the people who reposted was James Clear. I don’t know how he found it, but he’s probably the most successful writer in the category by far with Atomic Habits. And eventually, a couple different people really got into my head as this is what you have to do. You have to go do this. And I give them both credit.

There’s this saying that life begins where your comfort zone ends. And writing a book about venture or investing wouldn’t have pushed me to that place. This pushed me to a different place, and I approached it as a new hard project. So I hired a researcher, a co-writer. We read 100 biographies, we went through all the academic literature. We ran our own academic study with Wharton. We took it very seriously.

Josh Wolfe:

I like this point you’re making, which is that it was at the edge of your comfort zone. You’ve talked about this idea where something like 60% of adults are basically unsatisfied with what they do, so they’re not happy.

Bill Gurley:

Yes. We did a survey with Wharton and said, if you could go back and start over, would you choose a different career path? And 60% of people said yes.

Josh Wolfe:

And so, you call it career regret, right? There’s a consistent through line in Bezos’ framework of regret minimization that people choose wrong or that the systems never really let them choose or…

Bill Gurley:

I think it’s all those things, and I think it’s gotten worse. The college matriculation industrial complex has gotten very grindy. I think everyone knows it. Most would admit that people used to have way more free time when they were young to explore and learn and make mistakes and challenge themselves. Whereas today, by the time you’re in sixth grade, a parent is worried about your college application and they’re filling your schedule immensely.

And then you and I have discussed many different forms of psychological biases. One of those is sunk costs. If you grind from sixth grade to college graduation to get a degree in computer science, you feel pretty invested in it. I also meet a lot of young people who come out of college burnt out on learning. I think that’s a problem as well because the very best and the brightest, including the folk singer and the restaurateur and the basketball coach, are lifelong learners.

Josh Wolfe:

I think it’s also really important — and you highlight a few of them — but being exposed to a panoply of heroes, a pantheon of people that you can look up to and be like, “Oh, I could be like that.” My mom wanted me to be a doctor or a lawyer when I was young. And it wasn’t until I started to be exposed to these heroes that I was like, “Wait, I want to be like that. I want to do what they’re doing.” And you find this passion.

Bill Gurley:

I have a chapter on mentors. What you just described is what I call aspirational mentors. I borrowed this from Danny Meyer, the restaurateur, whose uncle convinced him to completely pivot. His salary dropped by 10x to go work in a restaurant. But he made a list of 10 people he felt were changing the restaurant industry at that point in time and started studying them.

If you are on an intentional high-agency career path, AI is a super weapon. You can find YouTube videos and podcasts with these heroes and then you can even just tell AI, pretend you’re this person and you start throwing questions at them. It’s unbelievable what you could go do and what you could go learn in this modern world.

It’s ironic. If you’re in a job you don’t like, if you feel like you’re a widget, AI is a problem. But if you’re crafting your own path, it could accelerate you. I don’t know how to get people on a high-agency path if it’s not something they’re fascinated with. It’s got to feel easy to run at it.

Josh Wolfe:

Just as people have different genetic predispositions for risk-taking — or for fitting in or for standing out — so many of the entrepreneurs that I’ve backed, you’ve backed, are very comfortable standing out. They’re very comfortable, basically failing. They’re comfortable disagreeing with the consensus. And some of that might be wired, some of it might be youth. I say “chips on shoulders put chips in pockets.”

Another of your principles is to chase your curiosity, which is different from other advice where people tell you to chase your passions. In venture, you see people who are obsessively curious, but you also see people who confuse curiosity with restlessness. How do you tell the difference between those two?

Bill Gurley:

Yes, so the first principle is to chase your curiosity. I admit that the hardest part of figuring this whole thing out is just identifying that. And a lot of people don’t get there. You have to be willing to explore and admit and test and not view that as failure. I think when you get there, you’ll know. But I like to throw out the idea of comparing it to binge-watching your favorite Netflix show. Would you go read about this instead of doing that? And I think for the people that are the most curious about certain fields, they do that constantly.

Josh Wolfe:

Would you rather? Yeah.

Bill Gurley:

There’s also a principle about owning your craft. There’s a principle on mentors and there’s a principle on peer groups. And in all of those cases, as you start to climb the ladder and get to a higher and higher rung, I think it’s a more difficult course. It’s not a 101 class, it’s more of a 400 level course, but you start to expand your learning, your peer network and your mentors outside of your field. It’s noisier. It’s harder to gather insights, but when you do, they might be more impactful.

Josh Wolfe:

I personally have such a wide range of interests that come from... It doesn’t come from a virtuous pursuit. It comes arguably from an insecurity that I want to be smarter than the next guy. Sometimes that is a liability. Sometimes I come off like a jerk or arrogant or something, but when somebody knows something I don’t know. Oh my gosh, I got to know what they know.

Bill Gurley:

One thing you said right there: there is a mindset of FOMO about knowledge. And one thing I would tell very young people is, especially when you’re on that first rung and that second rung, if someone says something you don’t understand, tell them right then. Say, “What do you mean by that?” There’s an instinct that you’re going to look like you don’t know and therefore, you’re dumb. But all the greats talk about this: Ask the question. They’re going to respect you for asking it.

Josh Wolfe:

Another of the other principles you have is to go where the action is. And some of that requires a little bit of EQ. There’s a difference between the people who really know where the action is, the A-plus-caliber people versus the transactional hucksters. But, either way, it’s arguably great and easy advice for the 22-year-old. But what do you say to the 40-year-old or the 44-year-old that...

Bill Gurley:

Well, let me make sure I ground it properly. The title of the chapter is “Go to the Epicenter.” But it basically says if you are entering a career that has an epicenter, theater in New York or movies in Hollywood or songwriters in Nashville, or advertising’s probably in New York, or Silicon Valley for high-tech, if you take yourself to that place, things don’t get easier by a little bit. They get easier by maybe 10x or more.

Many people who are wildly successful will tell you two or three stories about a chance encounter that led to something that led to something and that they wouldn’t be where they are if that didn’t happen. And some people can push back on that and say, “Well, that’s just luck.” But the number of opportunities for luck go up tremendously if you’re in the battlefield where everyone else is.

Josh Wolfe:

So you’ve seen older entrepreneurs, you’ve seen career entrepreneurs, you’ve seen young people. Do you think Gen Z is more or less equipped to follow your playbook than my generation or older folks?

Bill Gurley:

If I read what Jonathan Haidt’s written, not just in The Anxious Generation, but in The Coddling of the American Mind as well, people are not being set up for high agency (Editor’s Note: check out our interview with Haidt). And I do think the book requires a pivot towards high agency. You have to believe you can be successful. If you’re completely of a mindset of being a victim — of I can’t win, everything’s stacked against me — I don’t know that you’ll make it to the end of the book.

Josh Wolfe:

I want to talk about a few of the characters in the book, and then I want to ask a few more questions. One of the things that I thought was interesting was how you open with Danny Meyer’s visceral memory of this meal he had before the LSAT. He’s doing really well, he is making a lot of money, and then he quits to open up this restaurant, and you open the book with this anecdote. What does it tell you that people remember the moment they almost took the wrong path more vividly than the moment that they might’ve taken the right one?

Bill Gurley:

Well, when we go to bed at night, we ruminate a bit about things. When we make mistakes, we let ourselves off the hook, but when we miss out on something, it weighs on us, it weighs on us heavily.

On the Danny story: He cared about the details at a level most of us wouldn’t. And if you are obsessive about something, those details are super interesting to you. Steve Jobs, in his graduation speech at Stanford, talked about the calligraphy class he took and how much it weighed on his brain. My point is it’s possibly a sign of what you should go and do if the details of something matter to you.

Josh Wolfe:

I want to ask about not just chasing curiosity but chasing purpose. I’ve got a friend who said, for Shabbat, “we say a prayer every Friday and we wish our kids a life not of success but of meaning.” And I thought that was a really profound end goal. But I want to ask you, your parents, what did they expect of you? What did you feel they expected of you? What did you expect of your kids?

Bill Gurley:

My parents did not put overt pressure on me in any way, shape or form.

Josh Wolfe:

What did your mom and dad do?

Bill Gurley:

My mom was a substitute teacher and mostly a stay-at-home mom after that, but then really got into civics. So the last 20 years of her life, she was an alderman in the city and ran the school board and all those things.

My dad had a dream job. He fell in love with gas-powered airplanes when he was a kid. Fell in love with them and got an aeronautical engineering degree and was working in a wind tunnel in Langley Air Force Base. So he was doing it. He was doing what he loved. And then he took a flyer — a version of go to the epicenter. 50 of the people at Langley got offered a job at the beginning of NASA in Houston, and he took it.

I don’t think he ever once said to me, “Oh, go take a leap like that.” Most of my friends didn’t leave the state. And I thought nothing of it. I think if you live in a family that’s been in a place for a very long time, you probably don’t feel like you have the permission to leave, or you might be intimidated by the notion of it. And I didn’t have that. I do think that mattered.

Josh Wolfe:

And for your expectations of your kids?

Bill Gurley:

I’m trying to live the book as much as I possibly can, so we’ll see. I don’t think any of them are right on top of their dream job yet. But one of them got a MEC-E degree at Carnegie Mellon and came to us the summer before his graduation. He said he’s convinced he wants to teach, and he’s in Teach for America now.

Don’t miss

Will predictions become reality — or can reality be made to conform to predictions? That’s just part of the conversation I have with Dustin Gouker. Dustin is the writer of “The Closing Line” and “The Event Horizon” newsletters covering prediction markets and the sports betting landscape.

We talk about why prediction markets remain a small sliver of betting, how new underwriting models are taking market share from incumbents, the interface between betting and parametric insurance (because why not?), why sports will always dominate the industry, how performativity is increasingly interacting with international relations, and whether betting markets can be optimized for propaganda value.

This interview has been edited for length and clarity. For the full conversation, please check out our podcast.

Danny Crichton:

You’ve covered prediction markets for a long time. They’re going through a transition from being very niche maybe two–three years ago, to being broader. And this year, it seems like they’re poised to break out and become a nationwide phenomenon. Every retirement village is going to skip out on Bingo night and have Polymarket nights to bet on the future of Iran or whatever.

But what are you seeing? What are your predictions going into 2026?

Dustin Gouker:

The big thing, I think, is on the information side of things. The big idea of prediction markets is that people are trading on information. It’s not the gambling part, it’s not the sports part. People believe in this as a disseminator of information — I largely believe them.

I always come back to Intercontinental Exchange, the owner of the New York Stock Exchange, which basically said, “We don’t care about the sports stuff. What we care about is the information we are going to glean from the rise of prediction markets and disseminating that to partners.”

That’s one of the themes we’re going to see driven home this year and into the future: What’s the value of that information? Where else does this go? It also, of course, creates some controversy around insider trading and whether we actually need to be able to trade on some questions. But that information is the key piece.

The example I always go to is the federal government shutdown. It’s always been hard to place probabilities on whether the government will shut down. But we got that from prediction markets. Is the government going to shut down? Last year, the traders thought it would, and it did. And then, how long will the government shutdown go on? We were able to price that in real-time in a way I don’t think was possible before.

Danny Crichton:

You go back to the mid-2000s and there was this goal of crowdsourcing knowledge. The idea was that it wouldn’t be experts. I’m thinking of folks like Philip Tetlock, who showed that if you ask experts about political events, they’re not just wrong, they’re consistently wrong.

There were some startups around this in the late 2000s, early 2010s, but it petered out. And now you have Polymarket, Kalshi, and others that came out in the last couple of years and have done extremely well. I don’t know if you have an opinion on this, but I’m curious: Why the success today when, theoretically, prediction markets have been around for decades.

Dustin Gouker:

I started coming into prediction markets along with PredictIt back several election cycles ago. You could bet on elections — the presidency and down-ballot races. Polymarket and Kalshi have been growing up in parallel, but differently.

Polymarket tapped into something from a crypto standpoint. This is all on-chain. This is crypto native. And it’s international. Wink wink, it’s not in the United States.

Kalshi was founded in 2018, but it didn’t really launch until much later. It wasn’t doing a whole lot — a little election betting, started fighting the CFTC under the Biden administration to allow election betting, won in court. In 2024, they start blowing up, with hundreds of millions of dollars getting bet on the Harris/Trump election. And then early last year, sports betting rolls around, and that’s been most of what’s happened at Kalshi.

A lot of people talk about the other stuff, but 90% of the volume at Kalshi is sports contracts, between football, basketball, golf, you name it. Everything else has been kind of de minimis. There’s still hundreds of millions being traded on this stuff, but it’s dwarfed, absolutely dwarfed, by all the sports stuff.

And now we’re at a point where there are all these other players, and it’ll be fascinating to see what happens. There’s tons of crypto projects. There’s the blending of on-chain and off-chain, like Kalshi’s trying to do. There’s going to be a lot happening in 2026.

Danny Crichton:

When we go from sports into politics, culture, and society, these markets seem to have a huge amount of influence, or at least they do in my world. Polymarket just signed this partnership agreement with Dow Jones, the owner of Wall Street Journal, Barron’s, and a bunch of other products. Presumably they’re going to start to include some of that betting — the tickers and predictions — in articles.

Is this just, here’s another vertical, another opportunity to bet? Or is there something deeper there that I’m not seeing?

Dustin Gouker:

Bloomberg Terminal has added a bunch of functionality to ingest and then put back out Polymarket and Kalshi data, which is fascinating. People were asking, “Who’s going to create the Bloomberg Terminal of prediction markets?” It’s going to be Bloomberg!

The sports betting part of this is interesting because we have an ingrained world around sports betting and how the media does it. We might get a little too in-the-weeds, but the lifetime value of a customer at a sportsbook is way higher than the lifetime value of a single customer at a prediction market. So a sportsbook will pay more to acquire a customer or retain a customer than a prediction market ever will. That’s not going to change, I don’t think, even with competition.

So are we going to see this on ESPN and all the other networks? Arguably not, because they’re ingrained within the sports betting industry. I’m not saying it could never happen, but there are reasons why the sports ecosystem is not necessarily going to jump into this. Now, we saw the NHL do seven-figure deals with Kalshi and Polymarket to become their marketing partners and official partners. There’s something there, but at the same time, I just don’t see it being adopted in the same way.

The difference is that we now have these probabilities for things we never had probabilities for. Temperature, for example. Before, you were looking at weather apps or to your local meteorologist. Instead, we now have what traders think the temperature is going to be, how many tornadoes there are going to be, how many hurricanes we think there are going to be. All those things — that data and those probabilities — are what we haven’t had.

Could we see a partnership with the Weather Channel? Sure. Why not? Obviously, the Weather Channel probably really believes in its meteorologists, but at the same time, here’s another data point we can use to validate what we think is going to happen on any given day and in any given season.

Danny Crichton:

Is the Weather Channel still owned by IBM? IBM purchased them at some point back 15 years ago. [Editor’s note: IBM sold the Weather Channel in 2024.] IBM Watson is really what’s getting fired here, getting replaced with a prediction market.

I think of parametric insurance, this model that’s taking insurance markets from traditional insurance — where you’re guaranteed around damage — to a parametric model where you say, “Look, if winds hit 135 miles an hour sustained for more than five minutes in your area, you get a payout automatically, whether there’s damage or not.” Parametric models dramatically lower the cost of administration. It sounds subtle, but this is literally tens and tens and tens of millions of dollars of lawyer fees that go back and forth between insurance companies to figure out who’s at fault and who’s not. So it’s interesting to me to think of a future in which prediction markets will determine your parametric home insurance rates.

Dustin Gouker:

The insurance stuff is definitely interesting. I’m skeptical that average retail — you and me trying to insure our homes — are going to do it through a prediction market. I could definitely see institutional capital coming in trying to hedge against bad outcomes and hurricanes and things like that, though. That seems like a pretty likely outcome.

Polymarket just did a deal with an on-chain housing site called Parcl to start launching prediction markets for housing prices. Real estate is this huge market that, for you and me, the only way we intersect with it is buying and selling houses. I can sell my house or I can take out a home equity line of credit. It’s limited.

If we have markets where you can hedge on the price going up or down over a day, a month, a year, a quarter, whatever, then you have a real way to protect against bad outcomes related to your house.

Danny Crichton:

I want us to take a step back, because one of the goals with prediction markets was that they cover phenomena in the real world and you could get these bets that were, let’s call them objective, or crowdsourced. But now there’s a loop — actions in the real world can be altered based on what the bets are showing.

We’ve already seen some major scandals in sports, where players were trying to hit their own betting targets or trying to throw a game — whatever the case may be. We’re starting to see this a little bit with politics, although I don’t think we would ever blow up Iran just to hit a Polymarket deadline.

But I’m thinking of a little tempest-in-a-teapot scandal, where a press conference went an extra couple minutes. Say there was a prediction that the press conference would go for longer than 60 minutes, and so is a press secretary going to be looking at their clock and being like, “If I hold another minute, I can make $500 on Polymarket”?

How real is the feedback loop, and is that something you’re concerned about long term?

Dustin Gouker:

So far it’s been — I think most people believe it has been — on the up and up. Sometimes there’s insider trading, where somebody who has information is trading and profiting off of it.

But the Venezuela markets are the one that’s been ... I’ve talked to a lot of mainstream media folks about this, where somebody who was probably an insider at Polymarket traded on three other markets right ahead of the attacks in Venezuela on Maduro being out. It’s hard to believe it was a coincidence.

Now, is somebody going to strike a country because of a prediction market? Probably not. What I think is actually more interesting is, if these get to scale, would you make a bet to send false information. That doesn’t seem impossible to me. There are already, I’m certain, governments using Polymarket to inform decision-making about world leaders and things like that. Khamenei in Iran, that’s a huge one that people are keeping tabs on. What do traders think? Are there insiders trading on it who know more than we do? So I don’t think there’s this impossibility of using markets to send information.

You’re definitely going to see it in politics, with people trying to manipulate markets by putting a lot of money on X, Y, or Z to make it look like somebody should win when they don’t. That’s impossible to do at the presidential level with the amount of liquidity you’re going to have to beat down, but on smaller, down-ballot races, it’s not impossible. Now, Kalshi makes a point of prohibiting that kind of betting, but I think there’s some skepticism about how well they can actually police it. On Polymarket, and anything that’s on-chain, you’re anonymous. You’re linked to a wallet, but we don’t know who you are.

Danny Crichton:

You’re getting at the concept of “performativity”: Do markets conform to the metric or do metrics conform to the market? Hypothetically, metrics come out of the market. But I think there’s a lot of work, particularly in the last 20–30 years, that shows it’s actually quite the opposite, that expectations set the market more than the actual market itself.

You can certainly imagine that in the context of polls, where a lot of people just care that their side wins, or they don’t want to vote for the loser and so they just don’t show up at the polls.

The other piece I think is really interesting is this: I’m on the Polymarket homepage right now. There’s a lot of stuff on Iran, but the volumes are low. To your point earlier, some of the Iran markets are $700,000. The Super Bowl is $700 million. Okay. So you’re probably not going to win a lot trying to influence the Super Bowl outcome. But if you put $30,000, $40,000, $50,000 on a $700,000 market, yes, it will adjust. If the media were to start reporting that Polymarket says there’s an 85%, 90% chance of strikes in the next month, does that put pressure on a president? If something doesn’t happen, now it looks like TACO ... Trump always chickens out.

Dustin Gouker:

The CEO of Kalshi at one point was talking about the idea of prediction markets informing policy. I want to say it was a podcast with the Coinbase CEO where he said, “Should we be taking prediction markets and then using that to inform policy?”

That’s could definitely happen. We can see approval numbers or how races are trending or what people think about GDP. Can all of that turn into information people use to inform their policy decisions? For sure. I don’t think we’re going to see the collapse of democracy. But I do think there are concerns here.

And again, the lower liquidity the market is, the worse this is, and that’s why they want to get the scale. If there’s enough money, you’ll get hammered in the market. There’s proposed legislation about insider trading that may or may not help all of this. But this has all happened so fast that I don’t think a whole lot of people have stopped to think about some of the knock-on effects.

Danny Crichton:

We already talked a lot about 2026, but let’s project out. It feels like Polymarket and Kalshi have really built up the category. They’re likely to continue to grow. And we have these additional societal effects. How does this all come together over the next couple of years?

Dustin Gouker:

In the short-term, it is all about sports.

In two to three years, we’re going to have a Supreme Court case about all of this, and whether it’s legal and whether federal laws preempt all the state gambling laws or not. We get a decision there, then we have clarity, one way or the other. This is either here to stay or it’s gone.

The other one I remind people about is a presidential regime change in 2028. If the Democrats come in, or even if the Democrats win in midterms, there’s going to be, I think, more coming out against the sports part of this. I’m not overstating it to say that there are dozens of states — there are attorney generals, both red and blue — who have said, “This is sports gambling. Can’t have it here.”

So all the political and legal contours are going to shape all of this. That might be years down the road.

But longer term, I don’t know. If I’m pricing it, it’s a coin flip that sports betting survives. And then how much during these two to three years has everybody pivoted to elections and other stuff? Elections are going to be huge no matter what, but getting people to trade on what’s the price of Bitcoin, financial markets, cultural markets? How much adoption do they get?

For Kalshi, it has been pretty de minimis so far. It’s been sports, and then everything else is a rounding error. 10% of trading is everything else that’s not sports. Polymarket has done a good job of getting people to care about all these other things and put their money on it. So if they can replicate that here in the United States, or anybody else who comes along, that’s the part that is interesting and is going to be here no matter what.

Don’t miss

Lily Boland wants to help policymakers get a greater handle on these future risks, both to understand them individually and how they intersect with each other. She is the Strategic Foresight Fellow at the Converging Risks Lab of the Council on Strategic Risks. In that role, she designs unique foresight games that bring people together to explore alternative realities and their implication for our own.

We talk about the techniques of foresight and how it differs from forecasting, what can be learned through games, how and why people change their mind, and what are the most under-reported risks that we should all pay more attention to.

This interview has been edited for length and clarity. For the full conversation, please visit our podcast.

Danny Crichton:

Before we get going, let’s talk a little bit about the work you do.

Lily Boland:

The Converging Risk Lab is an institute at the Council on Strategic Risks. We have three different institutes. They all kind of cover different topical foci. So we have a Jane E. Nolan Center on Strategic Weapons. They cover weapons of mass destruction: nuclear, chem, bio, radiological. And then we also have a Center for Climate & Security, which was actually our first institute and is really focused on bridging those two worlds. And then at the newest lab, we merge subjects and look at converging risks across the other two institutes. So we’ve done projects looking at the intersection of climate, national security and nuclear energy.

We’re convening people from all of those different fields to do things like exercises and games. Overall, our mission is to prevent strategic surprise by identifying points of risk convergence.

Danny Crichton:

How did you get into wargaming? It’s a very narrow category of people that stumbles into this!

Lily Boland:

Yeah. I feel like I’ve reached an even narrower niche now by blending foresight and gaming. There are very few of us doing that multimodal approach. I’ve heard it called all sorts of things: mixed methods, people blending, forecasting, foresight.

I didn’t necessarily start off that way. I got into wargaming through crisis management simulation. I did two master’s degrees: a dual master’s program with Sciences Po in Paris and with King’s College in London. And by the time I had hit that point in my education, I wasn’t really sure career-wise where I wanted to go.

My first semester at Sciences Po, I took a crisis simulation course run by a former NATO general. He did a course for the whole year of my master’s degree in Paris, where there was not a second where you were not technically in role play. He could email you at 10:00 PM and ask you to prepare all these documents for the next day.

He created a scenario: a fictional war, a fictional country, fictional leaders. And it was such a detailed opening scenario that he later confided in me that he had spent almost a year designing it. But I found it to be a really fascinating way to apply what I had spent years training for. I had no idea that was its own kind of career, like crisis management, simulation designing, that kind of thing.

By the time I was preparing to move to London to do the second part of my master’s at King’s College, London, I was really excited to see that they had a whole bunch of courses on wargaming, led by David Banks, who was my wargaming professor and my mentor and everything.

Danny Crichton:

King’s College is a massive hub for wargaming. The United States has kind of pulled back over the last couple of years, but King’s College has really picked it up. Lots of academics focus on it, really trying to build it into a field.

You say you focus on this world of foresight. Maybe I do also, but just don’t realize it. So can you explain a little bit more? When you say the field of foresight, what does that look like? What are some of the techniques you’ve used from that field in your own work?

Lily Boland:

At its core, foresight is just a systematic way of looking at the future. Call it future studies. There’s a whole lot of lingo used to describe it. And not everyone agrees whether they’re a foresight practitioner or not. It often gets conflated with forecasting. It is not forecasting.

Forecasting has some predictive quality to it, where you are looking probably on a much smaller timescale and you are using more quantitative outputs. And you’re claiming that because you’re taking in all this data, you can predict to some extent that something will play out in a certain way.

Foresight doesn’t try to acknowledge any kind of predictive qualities or anything. You’re really just projecting outward and trying to think about all of the plausibilities of how the future could play out. So foresight’s often a better tool than forecasting in cases where people are coming in with a really strong conviction about what they think will happen and you want to challenge the scope of their understanding of the future. You pick apart those assumptions through different foresight exercises.

Foresight’s a very systematic approach, and that’s why I like to attach it to the front end of game design, because it has a step-by-step process for actually assessing factors and things that are pushing and pulling the future in different directions. It also wants you to look at history to see what events might have shaped where we are now, and then pushing that forward.

You are still grounding yourself in data and evidence. Sometimes that gets lost if you’re only doing one exercise, like a scenario planning exercise or using one foresight tool in isolation. But similar to wargaming, you don’t really want to just play something once and have that be the end of it. It’s important to have more iteration and more play tests so that you get a variety of results.

Foresight’s really more of a field of structured brainstorming techniques and learning how to challenge people and challenge their assumptions and get them better at anticipating things they may not otherwise expect. It’s a way of managing uncertainty.

Danny Crichton:

I think it’s interesting. Being a venture capital investment firm, we at Lux are not this formal, but I do think we do a decent job of what you’re describing. We do pre-mortems, we do pre-stories. We’re saying like, “Well, what would take this and make it a 100x company? What were the factors that would have come together that we said, if A and B and C combined, that’s where the multiple is going to come from and this is going to be an exponential return.” And then on the flip side, even at the growth stage, we say, “Well, what would have been those factors that we would have been shocked by? Would we be shocked by a regulatory change? Would we be shocked by technology competition?” We want to be able to get ahead of surprises. You started out with this when you went over CRL’s mission in general, which is to reduce strategic surprise.

But I go back to something you said a couple of minutes ago — obviously some people come into these scenarios with a lot of experience, but they have one narrative in their head. They have a very specific direction they want to go. We had Daniel Kahneman, the Nobel Prize winner, on the podcast a few years ago and he said, “No one ever changes their mind.”

I’ve hosted enough games to believe people do change their mind, or at least change a personal opinion. But we also know folks who don’t. What is the strategy? Is there a magic toolkit?

Lily Boland:

I hate to use the phrase “case by case,” but it often is case by case. My metric for a successful game or exercise is if someone has remarked that their understanding of a topic has expanded.

These policymakers are presented with so many different things in a day, if I can get them to acknowledge that they’ve expanded their understanding of a topic or they’ve expanded a definition of something to include something new — that to me is a huge success. And I have had that happen in really poignant ways.

One that comes to mind is that we ran a multifaceted crisis game in South Korea on this renewable energy project we’ve been working on. A couple of people fought me on the premise of the scenario of that game: we had projected outward into 2045 or something, and the world’s a lot warmer. People are really upset. There’s a global greening movement and tens of thousands of Koreans mobilize to protest. Everyone fought me on that scenario in the room.

In that moment, I would always want to humble myself. I was with a bunch of people who are experts on things I’m not an expert on. I’m in their country. I want to listen. But the dispute was just based on the fact that they didn’t believe Koreans would take to the streets against their own government. And flash forward eight months later, martial law was declared, and that’s what happened in December.

But in that room, they were fighting something so hard because they hadn’t seen it happen in so long, and they couldn’t really conceive of it. But we can’t predict how people are going to respond if all of a sudden we don’t have winters. If all of a sudden we don’t have water access, if groundwater depletes. And that’s really the foresight coming through: I want you to think about what is plausible. So for a moment, don’t fight me on it anymore, just try to consider it.

At the end of that exercise, the same people who did not want to do it came to me and said that, through the game, they understand now that when they talk about national security, they also need to talk about climate. The game expanded their definition of what national security is for South Korea. Huge win.

It is really challenging to change people’s minds. But to me, if you’re a good game designer, you’re getting people to play out of their comfort zones and into that magic circle of immersion, where they don’t want to leave.

Danny Crichton:

I like the idea of the magic circle because I do think immersion matters a lot. You have to have that kind of immersion that comes from gaming to really consider the full scope of the changes that are underway going into the future.

But let me turn to some of the policy topics you work on. Maybe we’ll start with nuclear because that’s one of your passions. I can’t think of a year in the last decade or two in which more has changed in nuclear. Massive shifts, thanks to the Ukraine war and Russia, massive shifts in East Asia, between Japan, Korea, China and others. What’s going on? What’s under-reported?

Lily Boland:

The news coming out of APEC and the deal to give Korea the technology to build a nuclear-powered submarine is fascinating. If we look at it in the context of AUKUS — a foundation of that agreement was to give Australia the same thing. But you have Korea that already has the technology. They’re one of the top leaders in nuclear energy exports and production, although the new Lee administration does not necessarily want to rely on nuclear energy too heavily.

So I’m interested to see how his energy plans, which are more heavily based on renewables, shift with Korea taking on this new agreement. And I’m interested in what kind of impacts we’ll see, diplomatically speaking, across the Asia-Pacific region, especially if it’s going to be seen as some sort of new threat.

It’s hard not to be depressed about the way the world’s going when it comes to new nuclear technology in general, though. I got a little bit excited about the Korea situation only because they are innovators in nuclear, and I’m interested to know how they take that on.

We just put out a report on tactical nuclear weapons and their implications, which debates what tactical even means and whether we actually have tactical nuclear weapons. I personally — and throughout this conversation, I’ve not been representing CSR’s views and especially now in the nuclear space — I would say I don’t think they’re going to be stabilizing or anything remotely or solely tactical. To me, anything related to nuclear is strategic and has strategic implications.

We did a whole foresight report on that and potentially plan to turn it into a game. With tensions being high, everything with the Ukraine war, it’s hard to remain optimistic. But I don’t want to go only into doomsday talk.

I will say, news coming out of Japan: I’m not really sure how either Japan or South Korea are going to go. I don’t think Korea’s in a position where it wants to piss off China. They have their own special economic relationship with China that’s outside of ours. They have their own offensive and defensive military compared to Japan, which does not. And I just heard some Japanese parliamentarian a few weeks ago was questioning the new prime minister, Takaichi, whether or not she considers Japan to be a colony of the United States.

Interesting things all around happening over there. I think it’s a tense situation to introduce yet another type of nuclear power technology.

Danny Crichton:

My solution to nuclear threats is to just live in Manhattan, since we will be the first to be bombed. And then everyone who survives — the 20 million people left — can figure out all the implications when the rest of us are all gone.

We’re almost out of time, but you obviously cover a lot of issues at the Council. So when you think about all of them, what is one that doesn’t get enough attention, enough funding, people don’t care enough, but you are very concerned?

Lily Boland:

Ecological security.

Danny Crichton:

Meaning like wildlife, wildlife trafficking or ...

Lily Boland:

That’s all part of it.

Danny Crichton:

Forests, forest management?

Lily Boland:

Yeah, all part of it. I would say ecological security isn’t a phrase that’s widely used. It’s kind of where climate security was about a decade ago. Ecological security covers everything that’s not related to the climate. So we’re looking at ecosystem collapse, biodiversity loss, algae blooms, all these other things.

And they have these tipping points. So something like groundwater depletion or reduction: full depletion is that tipping point. And it’s something that scares me a little bit more than climate.

If we were to surpass one of these thresholds, like the Amazon die back, for example… if we go past the point of no return burning the Amazon up, it’s going to start to be the largest producer of emissions instead of being a carbon vacuum. Those feedback loops that we could trigger — and we will trigger them as humans unless we intervene — is what’s keeping me up at night.

Don’t miss

The search for efficiency and its disappearance is at the heart of Brian Potter’s new book, The Origins of Efficiency. Through the book by Stripe Press and his popular Substack newsletter Construction Physics, Brian has tried to explain to an angry if curious public how construction actually works in the real world and why it’s an industry both ripe for innovation yet also mired in antiquated techniques.

We talk about the challenges of construction, why the variability of site selection is a huge problem, the lack of economies of scale in construction, and the regulatory burdens plus NIMBYs that make building so difficult. Then we talk about why Brian doesn’t think aesthetic uniformity has improved efficiency over time before talking about his writing process and how he does such in-depth research.

This conversation has been edited for length and clarity. For the full conversation, please visit our podcast.

Danny Crichton:

I feel like you’ve become the singular ambassador between the world of bits and the world of atoms. Why are people so interested in this subject and what got you into it?

Brian Potter:

My background is in the physical world: in the construction industry. I spent the first 15 years or so of my career as a structural engineer designing buildings and parking garages and water treatment plants and those sorts of things. The industry always seemed extremely backward and inefficient to me. Buildings are so slow to go up, and the way they go up is not that different from how we built them a hundred years ago.

And then in 2018, I had a chance to join this construction startup called Katerra. This was back when SoftBank was writing checks to anybody who walked in the door. They had given Katerra several billion dollars in venture capital to revolutionize the construction industry — to take what seemed like a very backward industry and move it into the 21st century, to build things in factories … the same way everything else made efficiently is built.

I was very excited to join this company because I thought their mission was spot on. To the extent I had a thesis, it was that construction is locked into some sort of bad equilibrium where the incentives force people to do things in suboptimal ways. And to disrupt it, you need a sufficiently big jolt to the industry. I thought, “Okay, well these guys have $2 billion, that’s enough for a pretty big jolt. They’ll be able to come through and establish this new, more efficient way of building buildings.”

The same way Henry Ford came along and established a new, more efficient way of building cars … I thought something similar would happen with construction. But it didn’t work out that way at all. They burned through their funding, and in a few years, they had to start laying people off. I left after five or six rounds of layoffs. After I left, they ended up going bankrupt. In the aftermath of that, I wanted to understand what had gone so wrong.

The thesis they were operating under was that you just need to move these operations into factories — that’s what you need to do to make processes more efficient and be able to produce things more inexpensively. That idea wasn’t really enough, though. Lots of people had tried that exact same thing, and it had never really worked. There’s a huge graveyard of companies that tried similar ideas. And I wanted to understand why that was and what about the construction industry was so hard to change.

I came to the conclusion that if we keep doing this thing and it keeps not working, we clearly don’t understand something about what it takes to make a given process more efficient. And so I started writing this newsletter, “Construction Physics,” about the construction industry. And this idea that we need to understand what it takes to make a process generally more efficient was the genesis of the book, The Origins of Efficiency.

Danny Crichton:

I’m sure a lot of our listeners have seen the viral chart graphing real cost differences between the early 2000s and today. If you look at, say, chips, those have massively improved in performance while costs have gone down, following Moore’s law. And then you have categories like construction, hospitals, education, and some other human services that have actually gone up in cost over time; they’re far more expensive today than ever before.

I guess the question is why the origins of efficiency and not the origins of inefficiency?

Brian Potter:

So for building stuff — for construction specifically — the thesis I’m operating under is that there are specific things you can do to make some process more efficient, and if you can do those things, great. Whatever you’re making, you can find ways to make it cheaper over time. If you can’t do those things, then you’re kind of screwed. It’s very, very hard to make whatever it is that you’re doing more efficient if those paths are blocked.

With construction, it turns out that all these strategies available to make some process more efficient are very, very hard to do in construction. And so you have these paths to efficiency that are blocked or extremely narrowed. And at the same time — and construction is not unique — we’ve made it increasingly difficult to do the processes that we have. There’s increasing regulation and increasing administrative and bureaucratic overhead in general. Things take more steps than in the 50s and the 60s, when we could go a little bit more by the seat of our pants.

Danny Crichton:

One of the theories I’ve always heard is that general contractors are sort of managing the construction project at a high level, bringing in subcontractors. You bring in electricians, you bring in duct work, you bring in folks who are doing plumbing, whatever the case may be. And one of the challenges is that there’s a lot of thrashing, and building sites are not full all the time. Just on my block in New York City alone, there are three building sites. There hasn’t been work done on any of them in three months. They’re just sitting there. One is a three-story townhouse. I think we’re in year four of construction for a building that is roughly 2,000 square feet.

Are there organizational behavior changes on the construction side, whether it’s the site, the general contractors, subcontractors that make it hard to be efficient?

Brian Potter:

Yeah, it’s interesting. You talk about the contractor breakdown, where each task is done by a separate subcontractor. They’re responsible for one little thing. Some mechanical engineer or mechanical company will come and install the HVAC system or whatever, and they’re a separate company than the guys doing the plumbing and the guys doing the framing.

But that’s not necessarily that different from how other industries operate, right? Ford doesn’t manufacture every single part in their cars, right? They have hundreds of suppliers, they have tier-one suppliers and those tier-one suppliers have their own suppliers and so on down the line. But when you’re doing very repetitive operations, it’s much easier to coordinate these steps, so everything moves a lot more smoothly from one thing to the other.

It doesn’t work like that in construction, for a variety of reasons. Part of it is regulatory, and part of it is that these are all separate businesses. But buildings themselves are also unique. You’re building something different every time. And so it’s hard to get your process dialed in, and you’re finding surprises, or someone puts a pipe in the wrong place because this is their first time building the building. And so there are all these things that conspire to prevent you from having a really swift and smooth process.

Danny Crichton:

When we look at housing in particular, there is this spectrum from a completely prefabbed house — like a trailer or double wide or something like this — to, when you get into cities, almost every apartment going up is unique. Every building has its own design.

How much do you think about the aesthetic concerns? I mean, there’s NIMBYism, and that shows up in historical preservation, but why can’t we be a little more process oriented? Is it purely aesthetics or is there something deeper?

Brian Potter:

Yeah, that’s a good question. I think the aesthetic aspect is a little bit overrated. People often theorize that you can’t have mass-produced housing because people demand too much uniqueness in their homes. The house is a reflection of them, and they want to have a house that works for their specific needs.

I don’t really think that is true. If you look at the housing that actually gets delivered, it is quite repetitive. Large-scale home builders have a library of a relatively small number of houses, and they’ll plunk those houses down in developments all over the country.

I live in a housing development built five years ago by a small local developer. It has four or five different floor plans, and those floor plans are just copied over 80 different houses. I live in a long line of houses, where every single one is the exact same floor plan, just sometimes mirrored, sometimes with slightly different finishes or whatever. It did not stop anybody from buying up those houses. It has not had any effect on the valuation of those houses.

In general, I think people are willing to accept a fair amount of uniformity in their housing and in their buildings, especially if that uniformity is associated with a significant cost decrease. I think people would be very on board with that.

The lack of uniformity has more to do with other factors — like different sites are different. They’re different shapes and utilities come in from different places and the ground has different properties and different parts of the country have different wind forces and different earthquake forces and different temperatures or whatever.

Back when I was in the construction industry, I would sometimes work on projects where we’d have a basic design for a building, but we’d need to adapt it to the specific site. You do save a decent amount of time because you’re not redoing the whole thing from scratch, but you couldn’t cut and paste it from one place to the other.

Another big part of it is the market is not quite big enough to accommodate a really huge amount of uniformity. Part of that is due to transportation costs — you can’t move stuff a super long distance. Housing is not like semiconductors, where you can make a billion of them in a single factory in Taiwan and then ship those all over the world. You’re building a relatively small number of units, and you can transport those only a relatively short distance. If you can’t get huge economies of scale, the returns from having a completely uniform product are not nearly as high.

Danny Crichton:

Are there examples in construction or manufacturing where there has been a dramatic revolution that really improved efficiency? And then on the flip side, are you cynical about certain parts of the industry where it seems like it will be impossible to see any change going on in the next 10, 20 years?

Brian Potter:

This is not an especially interesting answer, but the stuff where there’s a ton of regulation or other constraints makes it hard to have the flexibility to do new things or try something and fail, and use that as feedback.

So nuclear power is an example. Where it’s so regulated, it’s difficult to go in and try something and see how it works, and if you fail, learn from that failure. You’re prevented from doing that basic thing in nuclear power construction, especially in the United States. Other countries have managed to do it better. I’m not an expert on South Korean nuclear power regulation or Chinese regulation, so I couldn’t really speak to that. But in the United States especially, you’re so constrained from experimenting that it makes it really difficult to see where you can improve.

Danny Crichton:

When you think about the rise of AI over the last two years, the United States is bringing on tens of gigawatts of additional power in order to power all this. It seems like we have a construction boom. As some estimates say, half of last year’s GDP growth was driven by AI investments. Are there lessons from there to port to the rest of manufacturing and construction?

Brian Potter:

The amount of data center spending is now as much as — or more than — commercial office spending. So we’re building more data centers than we are office buildings, on some level of measurement.

One thing you’re seeing with this data center build-out is that data centers have been a relatively popular thing for jurisdictions to build. Local governments or whatever wanted these centers because they came in and paid a decent amount of property tax revenue, but didn’t place many demands on local services, since there were not that many people in these things. They needed water and power, but they’re not using extra police force capacity. You’re not stressing the school capacity, you’re not stressing road infrastructure or stuff like that.

So it’s like, “Oh, great, these guys will come in and build this big thing and pay me a bunch of property tax and it won’t incur a bunch of extra local government costs.” But now you’re starting to see a lot of local opposition. NIMBYism, which historically wasn’t really a factor in data center construction, now is much stronger. You’re really starting to see local opposition because these things are so big and are creating such big demands for electric power.

Danny Crichton:

The epilogue in your book looks at the future of construction manufacturing, taking lessons from the origins of efficiency to the rest of the 21st century. If you’re a policymaker, what are the top two?

Brian Potter:

One is this: It is very hard to predict what is going to be important and what is not going to be important. And so you want to give yourself enough flexibility to take a variety of approaches. This is a high-level idea that encompasses a lot of different things.

Kind of like I talked about earlier: It’s nice if you can experiment and learn from those experiments. It’s nice to be able to sort of solve problems in a bunch of different ways. For any given technology or process, there’s often a large variety of ways to implement it, and it’s not necessarily obvious what the best one is, so it is good to have flexibility.

The second one is that being able to achieve economies of scale is quite important. We see this in shipbuilding. There are lots of ideas for revitalizing the shipbuilding industry, which is good and admirable. But people don’t often come up with an obvious path to make it competitive in a world market where there’s already these other companies operating at huge scales and have poured huge amounts of money into these shipbuilding facilities. Are we really going to be willing to bite that bullet and make similarly huge investments?

Don’t miss

One person who has made it his mission to fix America is Charles Yang. He most recently served in the Biden administration at the Department of Energy, and spun off with the change of admins to start a new think tank called the Center for Industrial Strategy. He’s not just focused on research and publishing, but also building a network of likeminded souls who have the skills needed to bring industrial tech discussions into Washington. Through the Knudsen Industrial Strategy Fellowship, he is constructing cohorts of sophisticated and fervent believers that America can manufacture the future once again.

We discuss Charles’s transition from government service into Silicon Valley, the persistent cultural divide between engineering and politics, how to balance being a generalist versus a specialist, what think tanks really do, and how experiential tools like Riskgaming can change the policy discourse. This interview has been edited for length and clarity. For the full conversation, please visit our podcast.

Danny Crichton:

You spent a couple of years in government, how was your experience?

Charles Yang:

I was originally at an AI startup in the Bay Area. I was an ML engineer, building models on top of custom silicon, and I decided I wanted to do something more. There were a lot of exciting things happening at the time, and still now, in DC. The more you pay attention, the more you see there’s always something going on.

I was at the Department of Energy as a civil servant, helping stand up our new manufacturing office, helping do portfolio strategy for $10 billion of investments in manufacturing and critical minerals. And then later on, because of my background, I just got roped into helping stand up our AI policy office as well.

Despite all the challenges and bureaucracy, my takeaway is that everyone should spend some time in public service. It’s good for people to understand how our government works. It’s good for people to have some sense of ownership — or equity, you could say — in the fate of our country.

Danny Crichton:

You have quite a bit of depth in energy and these new financial mechanisms. To what degree do you want to stay specialized — think, “I go to a think tank, I focus on the research around this. I get better at it, I go back into government. I try to get it down the line” — vs a generalist — like, “Hey, being a generalist is actually really important. Being able to integrate different types of information, being good at politics, being able to do communications, that’s the key skillset.”

Charles Yang:

There is something to being a better generalist than all the other “generalists” out there. We can get into the new think tank that I worked on, because I’m starting to think a little bit more broadly beyond just, “Oh, I’m an expert in this thing. I can do X, Y, and Z thing.” When I came into the Energy Department, that was my mindset. I was like, “Look, I have a technical background. I understand how industry works to some degree, and I want to learn about how government works and serve my country.”

I did that for a little over two years. And at the end of it, my scope expanded a little bit more to thinking about, “How do we build movements? How do we crowd in talent to work on the problems that are important for reindustrialization in meaningful ways?” So hence the more generalist frame, more institution building.

Danny Crichton:

Well that leads us to one of the projects you’re working on, a new think tank called the Center for Industrial Strategy. What was the genesis of that idea?

Charles Yang:

It kind of started when I was at the department working two jobs. I’m like, “I’m doing a lot, but we need way more people doing this kind of work,” and, “Where do those people come from? Who is going to hold the banner for this?”

People often ask, “What do think tanks do?” which is a great question because it’s very amorphous. But one way to think about think tanks is that they are essentially vessels of different strains of ideology or schools of thought. So there’s varying shades of the right, tech right, different angles on that. Similarly on the left, there are different sorts of bundles of topics.

Looking at the landscape, when I thought about what we need to do to build a movement for reindustrialization, I didn’t feel like there was anyone building that vessel.

Danny Crichton:

How do you describe your think tank’s thesis? There’s industrial policy strategy all the way to the 1980s developmental state. How would you define it today in 2025?

Charles Yang:

The short tagline we use for CIS is “building a bipartisan coalition around industrial policy.” The “bipartisan coalition” part is important. “Bipartisan” because if this becomes a partisan issue, it just will not work. We will just ping-pong back and forth between different policies. And “coalition” because we are not the only player in the game. There is an ecosystem, and it’s an important part of our theory of change that you need to bring everyone else along for the ride. If you continue to be the only person banging a drum about something, it is very, very hard to get anyone to really listen. But if you convince many other people to follow in the same theme, then you’re actually able to effect greater change.

Danny Crichton:

When you think about think tanks, they have different motivations and different goals. One is policy advocacy. Policy development is a different version of that, coming up with new ideas, filtering them into political offices. Then there is the field team model, either because your party is out of power or you’re looking to develop a new group of people and it’s really a human capital management strategy. And then you have a media model, the strategy of saying, “Look, we got to popularize this.” How do you balance these motivations?

Charles Yang:

We very much don’t want to be a holding pen. There are a few think tanks who really specialize in that. And I think the benefit of being bipartisan is we can sort of flexibly play on both sides.

We’re very much focused on coalition building and getting a broader set of folks on board with this, and then helping shape the discourse in DC. One takeaway from my six–seven months so far is that the think tank space is very performative, and the actual shaping of discourse is actually much more subtle most of the time. It’s much more illegible, but most of the performative nature of reports or publishing is simply to tell your donor that you did it in a way that’s provable.

But we’ve done a lot of private dinners, and we’re doing some other work, building coalitions, for instance, with the Abundance folks (note: listen to our episode with the New York Abundance leaders from last year), to the point around getting other camps on side with reindustrialization and thinking about what that intersection looks like. We’ve done some other things that are more for fun. We did a rave for nuclear energy with a few friends, which was the after-party for the American Conservative Energy Summit.

Danny Crichton:

The donor management part of all this must be really interesting. In general, donors are generations ahead. They’re people who have built wealth, so they tend to be older (often by decades) than fellows and people who are going into politics. I’m thinking about some of the folks who sponsored my research over the years. I can’t imagine proposing a rave.

Charles Yang:

Part of my model is thinking about America’s new generation of wealthy stemming from arguably the greatest wealth creation event in history, which is Silicon Valley. They are much younger and they are much less integrated with traditional institutions. They’re more online, they’re more interested in doing things with some of their money that are not as traditional.

You see this showing up in this new age of think tanks that have websites that look different, that are doing micro sites, that are doing raves.

Danny Crichton:

We’re talking about the power of influence, influence of power. You are coming at it from a novel direction. Obviously, there are a lot of incumbent institutions that still have a lot of resources. How do these all integrate? Are these just different levers? Are they different layers? Are they targeting different people?

Charles Yang:

When I think about power, power is people. It is people who change policies, it’s people who write policies, and it’s people who are the ones implementing them. So if you think about stocks and flows of people, the Trump admin certainly has done very interesting stuff. Like with DOGE, for instance, bringing in a new class of tech talent that otherwise would’ve never, ever thought about government work.

I think this is what continues to amaze me about DC, and why I think it’s so beautiful: you will always meet people from every walk of life who just come because they really care about some issue. People you would’ve never expected are in the reserves or working on nuclear security or working in their local representative’s office. How you shape the mindset of the people who are coming in is key.

Danny Crichton:

There’s a whole hierarchy — social networks that run across the legislative branch, the executive branch, people in and out of government, programs that connect folks together in unique ways. Where do you start? Where’s the untapped lever of change?

Charles Yang:

It’s a good question. The ecosystems are all different: there’s Congress, the executive branch, the agencies. They are distinct sets of people, and their modes of operating or theories of change, the type of work they do, are all quite distinct.

With CIS, it’s really starting from where I am and then where we can go from where we started. That’s why, for instance, we did this bipartisan dinner series because I worked with a lot of folks who were Biden political appointees on industrial policy, and I knew some of the new folks who were coming into the Trump admin.

That’s sort of one example of where we started with just what we had and where we could go from there, and I think everything else was sort of opportunistically building off of that.

Danny Crichton:

I’m in a bunch of strategic planning meetings, which I’m guessing your think tank may have. But this gets at a very challenging topic, which is... As much as I work at a free-market think tank, we have plans that go for multiple years. They’re almost five-year plans because you have strategic direction, you have to match donors, you have a budget, you have dozens and dozens of staff. There’s a lot of work to build up a project, get it out there, get it into the hands of the right people, et cetera.

But then there are those inspirational moments of like, “Hey, there’s a policy topic. We can host a conference next month.”

What I am curious about, because you’re focused on industrial policy, which from my perspective is one of those things with a very long time horizon, is how you balance between planning and improvisation.

Charles Yang:

We are very actively testing out the can-you-just-do-things theory in DC. This is not a traditional way of starting a think tank or running a think tank, but we’ve been very opportunistic, and we’re seeing how far that can take us.

Danny Crichton:

Talking about people, you have this Knudsen Fellowship. Tell me about the selection process for that. How do you think about that expertise?

Charles Yang:

When I was at the Department of Energy working on industrial policy, I can tell you, for many of the manufacturing and critical mineral projects — projects with outcomes of $10 billion plus — decisions were made on the basis of the three or four people we had in the department who had real industry background. If those three or four people hadn’t come in, who knows where those billions of dollars would have flowed. I think that’s true of every kind of industrial policy program we’ve seen.

The pipeline of talent for people with actual industrial experience is very small. Frankly, it doesn’t exist. I was also inspired by a lot of the tech policy fellowships that have now sprung up and have been very successful in shaping AI policy and tech policy more broadly and bringing in technical talent, software engineers, AI practitioners into government. I was like, “Well, where is that for industrial policy?” It didn’t exist, so we made it.

In Dan Wang’s book (note, we had Dan on the podcast last year to discuss it), he talked about how engineers are given political responsibility early in their career, and given progressively more and more responsibility over time. So to your earlier question about how people learn, I think you learn by doing. Obviously, the political world is a very different one. I had to go through the learning process, but there’s no reason someone can’t do it. We try really hard to teach lawyers about industrial policy and how industry works, but maybe let’s just try an experiment where we teach some engineers how policy works and just see how that plays out. That is a big part of the Knudsen Fellowship, where we are taking a lot of founders and engineers.

Danny Crichton:

Policy and politics are different. But on the policy side, it’s mostly about trade-offs, and engineering’s fundamentally about trade offs.

Charles Yang:

Yeah.

Danny Crichton:

You are constrained by physical features, you’re constrained by budget, you are constrained by the talent available, the construction costs, whatever the case may be. If you’re building a building or building a software project, you work with infinite levels of trade-offs. That is exactly what government does, too.

Charles Yang:

100%. Oftentimes, the trade-offs are technical, and you really want an engineer in the room when you’re thinking about how the margins across different parts of a supply chain influence the type of capital that we use. I don’t think you learned that in law school.

Danny Crichton:

I mean, the lawyers have to be there at some point. But I do think, in many cases, programs haven’t been designed from the bottom-up in the right way because no one has thought about the users, the customers, the margin structures.

Charles Yang:

I will say – and it’s something I wanted to say earlier — that in my ideal state, we would probably have a lot fewer think tanks. We would just have the state capacity to really analyze tradeoffs for ourselves.

But that comment aside, I’ll just tell you what we’re doing for Knudsen Fellowship and you can infer. We see this as teaching engineers and founders how to engage with policymakers, and the policy memo is one of the basic primitives of that. So we’ll be publishing all the policy memos that come out of the fellowship on our Substack, but the point is really for them to learn how to engage a set of policy ideas and communicate them clearly.

Danny Crichton:

I was in a PhD program focused on industrial policy in 2014. Someone came up to me, a professor, and was like, “That is the dumbest thing I’ve ever heard. Industrial policy does not exist. It will never exist again.” I would’ve graduated in 2020, 2021, but I dropped out. And it’s very ironic because industrial policy seemed to come back right as I was looking to do it.

But one of the things I’m curious about is to what degree is this a field in which you need new ideas, new research, new topics versus, “Hey, we have this huge stack of ideas and a lot of it is about organizations and politics.”

Charles Yang:

I started a YIMBY group in my hometown in a previous life. The things they ended up doing are not groundbreaking, it’s simply doing the work. And that’s sort of the spirit, I think. Around national policy, there are still new ideas to be unearthed, but I agree that there aren’t that many. There aren’t 36 in a policy playbook that are all brand new. Sometimes it’s bringing back things that are old, but are now relevant once again. A lot of it I think should be learning from how other countries have done this successfully and how it applies to our political system.

And then, again, to the people-centered approach here, it is just getting the right people in the right places who are willing to drive outcomes and take responsibility within the broader policy apparatus or build connections between different political groups. People ask me sometimes, “What industrial strategy should we have?”, and usually my response is, “Almost any strategy will do just as long as we stick with it.” And the “sticking with it” part requires coalition building and getting people to share a common sense of goals and outcomes in the political system. That’s one of the key things we’re also working on.

Don’t miss

1.

First up, Europe may be a technological backwater, but it is the center of the geopolitical world. Here’s Laurence Pevsner chatting with Marko Papic, the chief strategist at BCA Research, about Europe’s adaptability in the face of threats like Russia.

Pressure makes diamonds. That’s what I would say. I think our two oceans have made us fat and lazy in many ways, geopolitically at least… I think that pressure makes diamonds. I mean this is what human history tells us.

2.

Europe, and particularly the European Union, seems like one big stream of regulatory ooze. Well, America is hardly immune to the same depredations as Yoni Appelbaum makes clear. In his book Stuck, he describes how housing growth has halted in vast regions of the United States, forcing Americans to stay in place rather than seek opportunity elsewhere. Here, Yoni talks about the history of zoning and some very Riskgaming-esque incentive patterns in local communities.

One of the stories, one of the arcs I trace in the book, as you say, is that wherever you get a polity which gives a local community the power to police who gets to belong to that community, over time, it will tend toward exclusion, and there’s a very good reason for this. Even if it is populated by a band of angels, the politicians who are running that community, the members of that community will prioritize the interests of long-term residents over the interests of prospective residents.

3.

Democracy versus authoritarianism is a challenge we discussed repeatedly in 2025. To what degree in a democracy like America should we quote-unquote just build, at the expense of endless due process. One of the most popular books on the subject came from Dan Wang’s book Breakneck, which chronicled his six years in China through the Covid-19 pandemic. Here’s Dan on China versus the U.S.

As a meta point I would say, let’s think a little bit more about what it is that people really want. Maybe sometimes it is the case that they want to feel that the contract on how to build the subway and who is subcontracting, that process is fair, but maybe often what they care the most about is having that new subway at all.

4.

These debates about governance are high-minded, but we also live in a whole new world where terrorists have highly-scaled access to cheap drones and generative AI that’s revolutionizing their businesses. We talk defense the next few clips, first with Colin P. Clark, who is a Senior Research Fellow at The Soufan Center. He talks about how technology is radically lowering the barriers to global terrorism.

Technology is changing the way that we look at terrorism, tremendously. I think it’s just the tip of the iceberg. And so there’s often a lag effect to these things. And in general, terrorists tend to be early adopters, right? Because technology can be what we call a force multiplier. So if you are a small insurgent group based in the Sahel in West Africa, and you’re now tinkering around with generative AI that can set it and forget it in terms of your propaganda, you’ve now freed up a significant amount of manpower hours to go do what terrorists do, which is plan, plot, and conduct attacks.

5.

It’s not just non-state actors like terrorist networks that are taking advantage of the enabling power of technology. Even large nation-states like Russia are updating their models, taking advantage of cheap freelancer labor to revolutionize the clandestine war behind Europe’s front lines. Next up, we have Daniela Richterova, a Senior Lecturer in Intelligence Studies at the Department of War Studies, King’s College London, talking about the “gig economy of Russian Sabotage”.

We’re seeing a shift from these well-trained and tested agents executioners to basically amateurs who are being recruited in all walks of life from all sorts of countries. There was one from Latin America, a number of Eastern Europeans who were recruited, but also other nationals, some western nationals as well. And from what we saw, this recruitment happens online now, and that’s where the whole gig economy idea comes from. Google platform such as Telegram, these individuals are able to basically volunteer for a job.

They are told how much this job would cost, where this would take place. They’re not always told what the purpose is. They can bid for a job as if they were an Uber driver.

6.

Russia is adapting its espionage playbook by cutting costs and outsourcing its activities to third parties. Meanwhile, China is fully embracing the digital- and data-centric world that we find ourselves in. The country’s aggressive counterintelligence program has cost the CIA and its assets dearly. Next, we have Tim Weiner talking about his new book, “The Mission: The CIA in the 21st Century” and how the agency is struggling to adapt its tradecraft.

Espionage in an age of ubiquitous technical surveillance is a whole new thing, and the American intelligence community has been desperately trying to stay at the curve or ahead of the curve in AI and particularly in quantum computing.

7.

Sabotage and skullduggery are changing rapidly in 2025, with countries taking advantage of new forms of hybrid warfare that allow them to sap their enemy’s capacity without attribution. One new domain of concern is agricultural security, which can be easily and wrongly confused with food security. Here is our Riskgaming scenario design consultant Ian Curtiss talking to Alicia Ellis, professor at the Global Security program at Arizona State University, about why ag security is an emerging new threat we should all be worried about.

If you’re talking about agriculture, economic warfare or cyber information and narrative warfare, all of these things are not simple responses. Most of them don’t even traditionally fall under the purview of the Department of Defense, so who handles it? Especially when you’re talking about agencies that don’t typically think of it from a security perspective. So those hybrid threats are almost perfectly designed to not so much counter as slip past the way US systems are designed to respond to threats.

8.

Let’s pivot a bit from the world of espionage to the world of defense primes. War came to Europe when Russia invaded Ukraine in early 2022. The stasis that was the hallmark of Europe’s post-Cold War defense infrastructure has since flowed away, and a vital new energy has been restored on the Old Continent. Mirroring some of Marko Papic’s views on the dynamism of Europe, here is Eric Slesinger, founder and general partner of 201 Ventures, talking about European defense autonomy and its new wave of startups.

There are things that European founders are watching happen in the US and doing differently. As an example of that, a lot of the M&A heavy capital intensive plays in the US, I think European founders are asking if there’s enough ground support for that to actually work in Europe. Are there enough acquisition targets? Is there enough capital to be formed to be able to do that strategy? I think that’s going to be a little bit different.

9.

Eric was mostly talking about defense startups and the funding required to get them to scale in Europe. Over here in America though, we have a bigger problem: how to protect critical companies like Intel and its fabs. The U.S. government swooped in to buy 10% of Intel’s equity this year, a shocking change to the traditional libertarian views of the Republican Party. Was it the right choice? Dylan Patel, founder and CEO of Semianalysis, joined to talk about the hard choices for the U.S. on semiconductors.

I don’t like state capitalism either. I think that’s a terrible idea. We win by being an economy that’s competitive and an economy that is capitalistic. And there’s certain ways to do state investment into industries that actually improves that, right? Look no further than Chinese auto. It is the most cutthroat competitive market in the world. Yes, there are some SOEs, yes, there were a lot of subsidies, but actually it’s extremely competitive, which is why prices are falling so fast and which is why they’re out-innovating any of the traditional auto OEMs in the world.

10.

We finally made it to chips! Artificial intelligence, of course, was the major theme of 2025, a watershed year when AI diffused into more and more workflows and it became clear that everything is going to change in the years ahead. But how will humans co-exist with AI? That’s the question Jacob Ward explores in his work, including his book “The Loop.” Here we talk about the cognitive harms of AI and surveillance, and how those harms might be ameliorated.

Once upon a time, if you asked your average cigarette smoker in 1957, “do you like cigarettes? Do you want to keep going with the cigarettes?” They’d say, “Yeah, it’s a positive refresher. I love this cigarette. This is a fantastic experience.” The enthusiasm we will have as people for the experience of the product is not going to be, I think, the right measure of whether that product is okay to be selling to people.

11.

Privacy is a major challenge on the web, and it’s not getting any better with AI. But that doesn’t mean smart builders and hackers aren’t trying to improve the status quo. To close us out for 2025, here’s famed internet researcher Renée DiResta talking about the 2026 midterms and what new tech like generative AI will mean for the future of elections.

My friend Katie Harpeth has this phrase, panic responsibly. You don’t want to say, “Oh, the sky is falling and everything’s going to be terrible.” I think this was for the 2024 election. I think it was important to say, these are the risks. This is what could happen. Recognizing that this could happen, how do we think about mitigating it if it does, or making the public aware that there’s potential there? I think that’s responsible risk management. Let’s just educate as many people as we can about this.

Well, from the economics of Europe and China, to the security threats of the gig economy and digital authoritarianism, to the next generation of hybrid war and onwards to chips, AI and the future of the internet, we had a whirlwind of topics to discuss on the Riskgaming podcast in 2025. I am sure 2026 will be so quiet that we won’t know what to talk about. For now though, enjoy the holidays, and listen again soon. I’m Danny Crichton, and my f’ing god, we’re done with another season.

Well, there is, at least in theory. A coalition of software and tech luminaries, joined by hundreds of supporters, recently launched the Resonant Computing Manifesto. They want software that is private, dedicated, plural, adaptable and prosocial — the antipode of the offerings available to us today. It’s a fresh vision, one desperately needed as LLMs rapidly democratize software engineering to everyone.

The Orthogonal Bet host Sam Arbesman and I jointly host this special episode with Alex Komoroske, founder of Common Tools, which dubs itself a new fabric for computing.

The three of us talk about the manifesto, how LLMs are changing software design, the same-origin paradigm, fully homomorphic encryption, remote attestation, and whether it is possible for software to be good and also be profitable.

This interview has been edited for length and clarity. For the full episode, please visit our podcast.

Danny Crichton:

You have a new manifesto out on resonant computing. So for those who haven’t read it, I would love a primer on what resonant computing is and how the manifesto came together.

Alex Komoroske:

The core idea is that hollow things leave us feeling regret, and resonant things leave us feeling nourished. Superficially, they look very similar, but they’re fundamentally different. The tech industry is really good at providing and creating things that look good but are hollow inside — and large language models will take that an order of magnitude beyond what it’s ever been before. So it’s more important than ever before for us to make sure that the computing experiences we’re using are resonant. There are five specific qualities of resonant computing.

One is that they’re private, that you are the steward of your own data, and it’s used in ways that align with intentions and expectations.

Two, that it’s dedicated. It has no conflict of interest. It acts as an extension of your agency as opposed to somebody else’s agency.

Three, it’s plural. There isn’t one centralized power. It’s a lot of diffuse federated power structures.

Four, it’s adaptable. It’s something that lifts you up, it doesn’t box you in. It’s not like, oh, well, some product manager said I can only have these five options. No, it should be something that’s open-ended.

And five, it’s pro-social. It’s something that helps bring you in harmony with the world around you. It brings you into the world as opposed to allowing you to retreat from the world.

Danny Crichton:

When we think about where the technology industry’s been over the last 20 years, I can see a lot of platforms, experiences and apps that are in opposition to resonant computing. One of the challenges when I think about pro-social tech is, well, social media. People would say, “Well, social media was pro-social. It was designed to connect people.” It just didn’t really work out that way.

So how do you create a distinction between the past — at least the things people have been trying to build — and where you want the industry to go now?

Alex Komoroske:

I don’t know about you, but I joined the tech industry because of the hacker ethic. In the last decade or so, though, as the tech industry has consolidated, it’s gotten to this late-stage thing where all consumer minutes are sliced up between five or so aggregators. We’ve gotten to this overly optimized phase, and everything is hollowed out.

It’s not just about tech. I think it is also true in politics. And in business. We’re so focused on optimization that we’re inadvertently hollowing out the thing we’re optimizing. And so, to me, resonant computing feels less like a new thing and more of a reminder of when we did technology differently.

And now is the right time. Large language models undermine a lot of assumptions that are baked into the current tech industry. Much of the tech industry assumes that software is expensive to write and cheap to run. LLMs undermine both of those, it’s now possible to write shitty software for basically free but running it isn’t.

And that is very destabilizing for the tech industry. So, if we’re going to destabilize, what are the principles we want to come back to? What are the seedlings we want to grow?

Danny Crichton:

Tim Wu just launched his new book, The Age of Extraction, talking about how platforms are increasingly taking more of what we might dub in economic terms, “consumer surplus,” and trying to put it on the production side of the equation. There’s this very broad intellectual banner that says like, “Look, we’ve lost control over our computers. We’ve lost control of the customization.” We used to have HyperCard and you could do really fun, interesting things. You had this open sandbox where even as a nine-year-old, you could make little games and it was super exciting.

Now it’s locked down. I can’t even build my own computer, so to speak, as the parts aren’t even available. RAM prices are triple the cost. I’m fighting with Nvidia to get access to a chip. The forces arrayed against you are the richest, most powerful, most influential, most popular companies in the world. What do we do?

Alex Komoroske:

There’s a dozen or so of us who have been working on this, Sam here is one of them. But Danny, two things to your point. One, I’m so grateful we have large language models available as APIs, and we have multiple options. You could imagine a world where ChatGPT went big before OpenAI ever made one of their completion APIs public. You can imagine a world where they go, “Oh, shit, I’m not going to give away my primary thing.” And then, you can imagine Google and Anthropic being like, “Oh, it’s too dangerous to expose this to anybody else or whatever.” And that would be a very different world than we are today.

But we’ve got multiple providers, and they are all competing on cost and quality. There’s a number of amazing open-source models that keep on nipping at their heels, and at some point will surpass them. And that’s great, because it is destabilizing current power structures.

The second thing I’d say is this: I’m obsessed with the “same-origin paradigm,” which is the laws of physics that describe how our software, especially the web and apps, work. I think somewhat surprisingly, the same-origin paradigm, this fateful decision to slice up different origins into silos and then give the owner of that origin — the creator of that app, that software — full control over that data, was a very fateful decision in 1994.

It helped lead to significant centralization and aggregation. A lot of things we’ve seen in the tech industry are all downstream of that decision. To me, if we were able to transcend that model, add in new or alternate models that work for different use cases that otherwise aren’t supported, combined with LLMs — wow, that could really catalyze something big.

Danny Crichton:

One of the big things people talk about is no one wants to run their own server. No one wants to run their own VPN. Not just because it’s complicated, but also because of the knowledge required to be your own sysadmin.

Isn’t that why a lot of these centralized services start? You have an app, it is home-baked. It is literally running on ramen for the first week or two. There are eight users. And then, it scales up because it’s successful. It gets up to hundreds of millions and a billion people.

In his book Enshittification, Cory Doctorow really emphasizes this evolutionary challenge: platforms always start by trying to give as much of the users as possible. Then, there’s some sort of pivot point where they either want to get to profitability or they just want to extract more money.

Alex Komoroske:

Yeah, I wrote a note to myself that said the SaaS business model is downstream of convincing you to accumulate your data on someone else’s turf, and then they rent access back to you in perpetuity to your own data.

You could argue no one wants to be their own sysadmin, but we’ve combined not being a sysadmin with not owning your data. And that is what leads to this downside. Have we talked about confidential computing? I forget.

Danny Crichton:

Not on this podcast.

Alex Komoroske:

Nobody knows about confidential computing, and it is wild, because it can do things. It changes fundamental assumptions about where control can live — and who owns what. Confidential computing is hardware support basically baked into all the chips that have been deployed to public clouds in the last few years, including H100s. Now, if you run it in confidential compute mode, no one — even those with physical access to the machine — can peek inside, and that’s awesome.

The reason you haven’t heard about this before is because it’s really a tail need for the vast majority of things. If I’m running some dinky web service, and I’m running it in Google Cloud or whatever, I already assume that the Google SREs are not peeking into my thing. It’s against their terms of service, and also it’s actually quite difficult to do. The people who use confidential compute are primarily defense contractors running secret workloads or people in the finance industry doing highly sensitive calculations.

But, one of the things confidential compute can do as a party trick is actually the most interesting component: remote attestation. There’s functionality built in, where each chip has a private key that’s derived from Intel’s root key that’s burned into the silicon and impossible to remove unless you destroy the silicon. That key can then be used to sign an attestation that says the bearer of this attestation was indeed running in confidential compute mode. And here is the fingerprint, here’s the shot of the VM that was running.

You can pass this to somebody else across the network, and they can verify, “Oh, this is definitely signed by Intel. It’s definitely running in confidential compute mode. And it’s definitely running this software, bit-for-bit.” I’m simplifying, but this allows you to have a very high degree of confidence that your cloud provider is indeed running precisely the software they say they’re running.

This, in turn, allows you to construct a system you might call a “private cloud enclave” that allows you to have your turf in somebody else’s area. It’s like having an embassy. Technically, someone could break in, but that would be an act of war.

So this allows you to get the benefits of cloud computing with somebody else as sysadmin, but it’s still on your turf.

Danny Crichton:

I want to get to another piece of resonant computing: the qualitative factor. What makes something feel like I have expressive control over it, that I am the one who is the master of the piece of the software as opposed to the software disciplining me. How do you define what that would feel like? What kind of software do you think inspires you that way?

Alex Komoroske:

I think the key test is, as you look more closely, does it work? Does it impress you? Is it something you like more the closer you look, or do you like it less? A resonant thing is fractally aligned. Every layer, the more you peel back, the more you go, “Ooh, I didn’t even think to care about that, but I like that. “ Whereas hollow things, you pull back and you’re like … “Oh, wait, they can send arbitrary firmware updates, and it’s got a microphone and the terms of service say they can send it to any of their partners?”

Samuel Arbesman:

Danny, you were asking earlier about the pro-social nature of social media. Yeah, they were designed to be pro-social, but that’s very much on a single level. It’s like the first order of thinking. Once you get to the second, third order of thinking, you realize these things are not pro-social at all. They’re actively against empowering and improving social interactions. That goes along with this fractal thing.

Danny Crichton:

Apple is famously very expressive. But Apple would tell you, “Look, the only way to create a really expressive device that allows you to feel like its yours, it is private, et cetera, is we have to own it vertically integrated top to bottom. We own it all the way down to the chip at this point.”

Yet you have this value in the manifesto of pluralism, which sounds really great, but competition is in many ways what triggers some of the negative effects where people are trying to grab more of the surplus value of these products and platforms.

I’m curious: When you think about competition and vertical integration, is it possible to allow all these different layers to be plural, or does it need to be vertically integrated?

Alex Komoroske:

I think it has to be vertically integrated. You make a really good point about pluralism requiring competition in some form. I would argue that Apple fails the test precisely because they are not plural on these things. If you told people when the iPhone was announced that this was going to become the single most important computing device of all time, and it was going to be the vast majority of your computing, the idea that one company — which by the way, has consistently put a pretty heavy finger on what things they allow on that device — owns it would be insane. That’s an insane thing for us to allow as a society. iPhones are great. And also, the idea that one company that is very jealous gets to decide what goes on the iPhone is wild to me.

Samuel Arbesman:

Just taking a step back in terms of thinking about pluralism, we talk about this idea that a healthy ecosystem requires lots of different choices and things like that. Putting on my lapsed evolutionary biologist hat, healthy ecosystems are not just imposed or designed. They evolve over time into a thing that becomes healthy.

And I feel like that’s the argument being made: Yes, on the one hand, maybe some company can design a perfectly resonant experience in this vertically integrated way, but more likely than not, the resonant experiences we want in the tech world are going to evolve from the bottom up. And that’s why this plural principle is so important, because you need to almost cobble it together.

Danny Crichton:

Let’s project forward five to ten years. Your resonant computing manifesto has gone absolutely bonkers. Everyone has signed onto it. What does that world look like? What would companies be doing differently in terms of designing software, in terms of how they’re giving it to folks, in terms of the collaboration that’s built around it? What does it feel like?

Alex Komoroske:

My mental model is this: software today feels like a thing you go get at the big box store. You choose which of three basically similar options to buy, and they all suck. What if software felt instead like a thing that grew in your personal garden, something that was nourishing, that was specific to you. Your data comes alive in ways that help you accomplish the things you care about.

In that world, software doesn’t look like it looks today. Software does not look like having silos that do one particular task. Clay Shirky has this essay from, I think 2004, 2007 called “Situated Software.” Situated software is software that’s highly situated to a specific context.

To anybody else who looks at a piece of a spreadsheet you’ve modified, that’s situated software. It’s highly situated to your particular need. And if you show it to anybody else, they go, “That’s a piece of shit. It barely works. It’s ugly, it’s insecure.” But, to the person who made it, it’s perfect, that’s exactly what they need.

I think situated software, that’s what it will feel like in the future.

Samuel Arbesman:

Just adding on to what Alex was saying: there’s this blending of consumer and creator. There doesn’t need to be a single place where there are people producing software, and then people just consuming it. In some ways, it’s a return to the earlier days of computing, which unfortunately was only available for a very small set of people.

My family’s first computer was the Commodore VIC-20. And during that time, one of the ways you got software was you would go to a magazine, and find these type-in programs where you literally had just pages of text and typed them in. And you could see this very clear relationship between the text you were typing in and the result on the machine. Sometimes it worked, sometimes it didn’t.

But that blending and blurring of creator and consumer was only open to a very small subset of people. And hopefully, that kind of thing will now be available to everyone.

Danny Crichton:

I think this works within capitalism. The biggest challenge right now is people are worried they can’t make money. They have to own more. They have to control more.

We all know how there’s no sustainability to open-source software because there’s no money. But I do think there’s a world where you can align the value that’s created — that still makes plenty of money for all the companies. It’s very much in line with the resonant computing mission. You’re also seeing this with the fatigue a lot of folks have with the systems and apps they use every day right now. They’re not being nourished. They’re getting tired of it. You see how many people are quitting various social-media platforms.

And so, I do think that there is a world here. It requires imagination and vision from a lot of leaders, which you have brought quite a few around the table.

Alex Komoroske:

Yeah, I agree. To me, this is totally compatible with business goals — products that are resonant are ones people love using, and they evangelize them, and they feel good about them. When you take a long enough time horizon, and have a broad enough perspective, resonant products are simply better products. They’re better from a business perspective, for shareholders, and what have you.

I have another essay about the optimization ratchet. We didn’t necessarily choose to get so hyper-focused on optimization. It’s just that as you get more efficient, the benefits of optimizing are concrete, short-term and very visible. The downside of optimization is that you’re losing adaptive capacity, you’re losing resonance. But what you’re losing is not nearly as obvious. And so, with each of these micro decisions, the universe tilts towards optimization, and you do more of it.

But you have to know that there’s something on the other side of the balance scale.

Samuel Arbesman:

As people see all these things in modern society, especially in the tech world, they’re very unhappy, they’re burned out with everything in big tech. In the absence of realizing that there’s another path forward, you just see technology as the problem. But it turns out it’s not. There are just certain ways of building technological systems and software that are very problematic. But, if you can provide a name and a framework for this other path, it shows that it needn’t be the default.

Don’t miss

All of this is ably documented in Christian Davenport’s new book, Rocket Dreams: Musk, Bezos, and the Inside Story of the New, Trillion-Dollar Space Race. A follow up from his massively popular book The Space Barons, the book chronicles the last seven years of the new space race with alacrity and intimate details.

Christian and I talk about how he transitioned from covering Afghanistan and America’s home front in the War on Terror to covering the crazy billionaires at the heart of the space race. Then they talk about Musk’s fight over the United Launch Alliance, why Blue Origin has been taking on more risks in recent years, NASA and the government’s growing dependency on private companies for manned spaceflight, what the next decade of the space economy will look like, and finally, how China is entering the picture in a frenetic way.

This interview has been edited for length and clarity. For the full episode please visit our podcast.

Danny Crichton:

Christian, you just published a new book, Rocket Dreams, focused on Elon Musk, Jeff Bezos, and their billionaire race for the stars. Tell me how you got into the space world.

Christian Davenport:

I was at the Washington Post during 9/11, and at the time, I was a metro reporter covering a suburb of Washington DC. A lot of the people in the communities I was covering were being called up to go over and fight in Iraq and Afghanistan.

I like to say I was the first reporter to be embedded on the home front. If you’re a reserve component, you don’t go back to a huge military base where you’re surrounded with people like you, you take off your uniform and you go back to civilian life. I wrote a lot about that. But then, in 2014, I got assigned to cover this press conference at the National Press Club because some guy named Elon Musk was suing the Pentagon for the right to be able to launch national security payloads — national security satellites for the government — on his rocket. I don’t think I really knew who Elon was at the time.

Danny Crichton:

Yeah, he had a much smaller profile back then for sure.

Christian Davenport:

Right. The whole premise of the news conference was a bit insane. He was filing a lawsuit against the Pentagon even as he’s trying to get the Pentagon to be his customer. He wants to work for them, and the way he’s trying to do that is to file a lawsuit.

But at the time, a company called ULA, the United Launch Alliance, which is made up of Lockheed Martin and Boeing, had this monopoly on all national security launches and were just bringing in billions upon billions of dollars and charging the government a lot of money for that service. Elon’s point was, “Don’t give me the contract, but at least allow me or others to compete for those contracts.”

I left the press conference and I wrote my story, but then I started researching SpaceX and re-usable rockets and Elon and what they were doing and ultimately decided we needed to be paying attention to it.

Danny Crichton:

Your book centers on the competition between two big billionaires, and that’s the crux of the story. But there’s also this massive legacy of NASA going back to the 1960s and the Apollo program and everything else up to the present day. I’m curious how you balance between the government and NASA’s public mission versus that ego-driven ambition that’s coming from the private sector?

Christian Davenport:

That tension is at the heart of the book. And it’s growing and evolving. It’s a push-pull, because governments have had exclusive dominion over human space flight and space exploration for decades — since the dawn of the space age. Yeah, they’ve always used contractors, even going back to Mercury, Gemini and Apollo. But now, you’re seeing this erosion, where the private sector is going in and taking parts. In part, that’s the way NASA designed it. That’s the way the government designed it. They decided to cede some of these responsibilities to the private sector.

You have it starting from outsourcing missions like flying cargo and supplies to the International Space Station, which evolved into flying astronauts to the International Space Station, which has now evolved to the flagship program, Artemis. We’ve outsourced the spacecraft that’s going to land astronauts on the moon to the private space sector, outsourced the spacesuits. Those are going to be built by the private sector. The rovers, some of the technologies for power generation, all these sorts of things are being outsourced.

So what is the government’s role in this? It’s setting the direction, having oversight of these companies. But I think you’re seeing a huge effect, where so much of the talent and expertise has gone into the private sector. If you’re a young engineer today, are you going to go work at NASA or are you going to go work at SpaceX?

Danny Crichton:

Right. We see this in our own portfolio. We have a bunch of space tech companies and many of them are alums of SpaceX. Theoretically, in a different generation, they would have been at NASA. Instead, they’re building rockets, they’re optimizing the fuel systems on these vessels, and they’re very effective at it. On one hand, they’re very public-minded and they really do believe in these missions. On the other, the economics, as we all know, are really tough, especially when the government shuts down, NASA is adrift with an administrator coming back and forth.

Most of Rocket Dream is focused on two companies: SpaceX and Blue Origin. SpaceX always gets 97% of the coverage. Part of that is Elon Musk, who is just a larger-than-life figure. But we just had a big breakthrough with Blue Origin a few weeks ago, where it had the successful orbital launch and booster landing of its powerful New Glenn rocket, carrying NASA’s ESCAPADE Mars mission and it landed the reusable first stage of the rocket on a drone ship.

In your book, you describe SpaceX as very ambitious, very experimental. Blue Origin has been this middle ground. They’re not plotting, they’re trying to slowly build up capability, they’re much less flashy, a little bit more grounded. Did the news this last week from Blue Origin fit with that narrative?

Christian Davenport:

That’s a great question. I should be thankful to Blue Origin for landing the booster, because actually, I think the book sets it up perfectly, so it’s great timing.

You’re right, SpaceX gets 97% of the coverage, because they do 97% of the work. They’re launching a rocket every two days. They’re the ones winning the contracts, from the Pentagon and NASA, they’re really doing it.

Blue just has not been as successful. But in the last two years, you really have seen a change in their metabolism and, I think, risk tolerance. There were several inflection points. One came in 2016, when Jeff Bezos was saying to his team, “They’re pulling ahead of us, SpaceX. We need to catch up and we need to go after everything that they go after,” and he urges his team to do it. They were supposed to win this big moon-lander competition, but they ended up losing.

And that leads to the second inflection point. Remember, for Jeff Bezos, there are a few failures in his career at Amazon here and there, but nothing like this. No defeats like losing this contract to SpaceX. They just change, I think, as a company and they say, “Let’s forget NASA. We tried bidding and offering NASA the solution we thought they wanted, but now, we’re going to build what we want and we’re going to innovate and we’re going to come up with something completely different.”

The architecture for their moon-lander today uses all these innovative new technologies. It’s essentially a reusable spacecraft that stays in orbit around the moon and then a service spacecraft goes back in between the Earth and the moon to service it, to refuel it. The barrier to entry for that is so difficult. It’s so much more risky. But it’s also the SpaceX way.

Danny Crichton:

When you think about the venture capital industry, we generally try to invest in companies we’d want to exit in a decade — that’s the lifecycle of a fund. One of the interesting things here is that these are old companies from a Silicon Valley perspective. Blue Origin is actually older than SpaceX. It started in 2000. SpaceX, I think, was about 2002, 2003. They’re almost entering their third decade, and you have two CEOs who have many other things going on.

How much does Elon and Jeff’s attention drift onto and off of these companies, and how does that change how those companies function?

Christian Davenport:

I think you see it perhaps more at Blue Origin than at SpaceX. Famously, Jeff was working at Blue Origin on Wednesdays and one or two Saturdays a month. He has said that since he left Amazon, Blue Origin is his top priority. He moved to Florida, probably for tax reasons, but also I think he likes being closer to their Cape Canaveral facility. He’s focusing a lot more attention on Blue.

Elon is just Elon. He could be doing a million different things, but when he’s dialed in to SpaceX, he is dialed in. I think there was a lot of discussion, particularly around the time where he was embedded in the White House and doing DOGE, whether SpaceX was a little bit distracted.

I, frankly, heard both things. I heard that there wasn’t the focus, but on the other hand, I think a lot of the people were glad to have a little reprieve from Elon. The other thing I think about with those two is what happens to Blue Origin when Jeff goes away, retires? Or what happens to SpaceX? Can those companies exist and thrive without them?

Danny Crichton:

Jeff’s 60+ years, so still spry. But there’s this balance between their time and attention and skill and recruiting and the cachet they bring these companies, and then there’s the capital aspect: they’re putting their money in, they’re putting their dollars to work.

Christian Davenport:

You’re exactly right. Blue Origin has existed essentially like a nonprofit with one benefactor, and that’s Jeff Bezos. He’s said that, every year, he cashes out a billion dollars in Amazon stock and plows it into Blue Origin. Elon funded SpaceX to the tune of about a hundred million dollars early on, but it’s been able to exist and thrive with government contracts and leveraging that alongside private investment to build the Falcon 9, build Dragon Spacecraft, and use those vehicles to then go into commercial markets and build another line of business.

That’s a very good business; it could have survived for a long time doing that. But Elon disrupts himself. Enter Starlink. If SpaceX moves from being a rocket company to becoming an internet provider and a satellite provider, then the upside could be 10x, or maybe potentially even more. They’ve launched 10,000 satellites and have 8,000 operating, but now they’ve got a satellite production line, where you can change the guts of the satellite. It doesn’t have to be internet and communications. It can be who knows what? Remote sensing? Spy cameras?

We know the National Reconnaissance Office has a contract with SpaceX to use that satellite line, whereas Blue Origin is just beginning to do that. That launch the other day, that was a launch for NASA. They have a lot more lined up, but in terms of the upside that Elon and SpaceX have, there’s no comparison.

Danny Crichton:

Two years ago, Ronan Farrow had a piece in The New Yorker. Basically, the U.S. government used to have the unique capability to put astronauts in space. Now it’s just becoming the strategic setter of the country’s direction as opposed to actually owning the operations.

Up until recently, that has basically been a monopoly on the SpaceX side. They are the only ones with the capability. Increasingly, and this was the Farrow point, SpaceX has the ability to just say no. And if it says no, so to speak, the government doesn’t really have an alternative way of solving some of these sorts of problems.

Does Blue Origin catch up and say, “Look, now there’s an oligopoly. We can form at least a duopoly, which will put pressure on SpaceX to reengage”?

Christian Davenport:

I think NASA’s idea, from the very beginning, was that if space was going to go commercial, the United States would need a growing commercial space industry. And so, the question was: Do we have a commercial space industry or do we have SpaceX? For a long time, we’ve just had SpaceX.

But now, you’ve got so much of the national space enterprise in the hands of one company, essentially one person, it’s not healthy. You saw perhaps one of the most dramatic illustrations when Elon and President Trump had their little spat. Elon threatened to take the Dragon Spacecraft away from NASA, and that’s the only way NASA can fly its astronauts anywhere.

A lot of people thought it was reckless and dangerous, and it sent off alarm bells. I know people at NASA and the government were reaching out to companies like Blue Origin, Rocket Lab, Relativity, Stoke Space, and other providers to say, like, “You guys got to hurry up.” And I think the government will bend over backwards to bring those companies in. Even if they’re not the lowest bidder, just by virtue of not being SpaceX, they’ll be able to get a foothold.

Danny Crichton:

One of the big concerns in American politics and the economy is short-termism. We’re always on this quarterly earning cycle. Yet we’re talking about companies that have been around for several decades and two individuals in particular who seem to have extensive long-termism around solving these problems. Why so much long-termism here? What sustains them on this compared to so many others?

Christian Davenport:

For Jeff, it’s a very important mantra and very key to understanding how he thinks. He’s got a thousand-year clock that chimes like once every hundred years or whatever it is, which is a symbol for long-term thinking and solving problems. Space is just hard. The barriers to entry are extreme, but they both think it’s really important, and they’re thinking about it in terms of what their legacy for their life is going to be?

There’s also this tit-for-tat and the competition between them and the rivalry, but the fact of the matter is they’re actually very similar in their thinking about space, and their long-term thinking about space and lowering the cost and increasing access.

As Jeff said, when he founded Amazon, you could found an internet company in your dorm room, because you had the infrastructure. Phone companies had laid down the cables for the internet. The Post Office could deliver books. An invention called the credit card that could take people’s money. He could build a business on that infrastructure.

But if you want to build a business in space, the infrastructure isn’t there. It is starting to get there, but I think that’s why they talk about the long-term vision and building that infrastructure. And then, once you can access space frequently, what does that open up? What about data centers in space? What about space-based solar power? Commercial space stations? Mining celestial bodies?

Danny Crichton:

Even on our side, we’ve led a $25 million series A into a company called Reflect Orbital, which is putting space mirrors into orbit. Five or ten years ago, this would have been ridiculous. Now, with the advent of Starship, hopefully in the next cycle here, the economics of space — the economics of payloads — change. This is the bet we’re making in the company and the team. On the other hand, I’m still not a complete believer of data centers in space. Chips heat up, and it doesn’t seem like we’ve found any solution to that.

To your point though, in 2025, what I am seeing is that we’re pivoting from this world of how to get stuff up. That’s increasingly a solved problem. The next 10 years to me is like, “Well, what do we do with this?” You’ve done two of these books covering basically a 20-year period of space flight and commercial space flight history. Is there a third book coming in 10 years? What do you predict is going to happen?

Christian Davenport:

I’m thinking about this a lot, and I do think of it in terms of a series. The analogy that I use is, westward expansion and the pioneers. I think we’re in the covered wagon stages. We’re going, we’re accessing the territory, we can get there, but we’re moving into homesteading. As a Space Force general put it to me, we spent the first 50–60 years of the space age trying to get to space. Now, we focus on getting through space, moving through space, maneuvering, staying there, changing orbits, traveling, refueling, creating businesses up there and doing all that.

I think once we get to that point, there will be a book there, many books there, about the homesteading aspect of it. That’s how I see it.

I was talking about my thesis with someone who’s much smarter than me and they said, “No, no, no. Before you get to homesteading, you have to have the cavalry.” You need the rules of the road. I’ve been thinking a lot about national security. A lot of people think Space Force is a joke. It is not. The threats from China and Russia and others are real. They can hold our assets in space at risk. What we have up there — GPS, just for an example — is incredibly important, and not just for Uber Eats, DoorDash and Waze. The timing signal on a GPS satellite is used for every ATM transaction, every trade on the New York Stock Exchange.

Those satellites, the 31 of them, 34 of them… they are vulnerable. China, Russia, North Korea and others have shown they can take them out. That’s a real risk. Anyway, that’s what I’ve been thinking about.

Danny Crichton:

There’s also the cyber aspect. These are digital systems, they’re hackable. In many cases, they’re not easy to patch. And so, once a nation-state actor investigates and identifies a vulnerability, it’s not necessarily easy to close.

And also, with all these thousands of objects in orbit, there’s the issue of space debris. I think it was an MIT study that came out earlier this week that said there will be so much space debris coming back down to Earth, that one plane every decade will be hit and blown up, and we have no solution for this problem whatsoever.

So to your point, there’s a huge amount of stuff here that has come together all of a sudden that didn’t exist 5, 10 years ago.

Christian Davenport:

Yeah, and the regulations, as you pointed out, are just not there. When you talk about space debris, the U.S. government can track it and they can tell satellite operator A, like, “Hey, you may collide with satellite B,” but it can’t order them to maneuver. Some of them can’t even maneuver. We just saw a Chinese spacecraft apparently hit with a piece of debris. I joke that if the problem of space debris gets really, really bad, then it’ll be like a moat protecting Earth from any incoming aliens.

But in general, what China’s been able to do in space is remarkable. There is no differentiation between their civil space program and their national security space program. It’s not like in the United States. Really, China’s space program is run by the military. And what you are seeing now, by the way, is — and this is a Space Forces term — Chinese and American satellites “dogfighting.” Not over a 10- or 20-minute period, this is days, even weeks, where they are flying by at very close proximities, which in space is miles apart, but doing it in a way where they gain tactical advantage.

Danny Crichton:

You titled your book “Rocket Dreams,” and I want to emphasize the word “dreams.” NASA, the intelligence community and the Pentagon — one of the reasons these were all separated is they had different dreams. NASA was focused on scientific exploration, the Pentagon focused on strategic and tactical advantage. You have space races between countries, where they are trying to gain advantages over each other. You have billionaires who are competing on ego or from a commercial perspective. There are so many different types of dreams.

But to me, there’s this balance between dreams like “Hey, we’re going to have space stations in orbit” — those already exist, it’s just scaling up — versus that long-term dream of going to Mars. We may never achieve it. Which one do you think is motivating folks better?

Christian Davenport:

On the Mars side, that’s part of what Elon does, and he’s good at it. He’s trying to make space cool again and driving this vision. That’s why he cares about aesthetics. That’s why he cares about the space suit looking cool. The spaceship has to look cool because he wants public support.

But the fact of the matter is this, as someone said to me: Imagine the first human landing on Mars, and imagine it goes perfectly. Seven minutes of terror, entry, descent and landing — this moment of absolute triumph is also a moment of pure peril in its crisis mode from day one, because Mars is just so difficult.

Jeff says, “Oh, you want to know what Mars is like? Go live on the peak of Mount Everest.” It’s actually like living on the peak of Mount Everest in Chernobyl. It’s not a pleasant place to go. I think it would be astounding.

We romanticize it so much. But spaceflight is really, really hard. It’s incredibly dangerous. To think that we’re going to go to Mars and not kill a lot of people, I think it’s foolish. I think this vision of having big space stations in orbit, where you can come back to Earth, I think that’s a long way off, but it seems potentially more suitable for human beings.

Don’t miss

Jacob Ward — the former editor of Popular Science, long-time tech correspondent, podcast host of “Rip Current” and the author of the popular book The Loop — is skeptical. Via his own personal experiences and reporting, he sees AI’s addictive qualities and its lack of safety as a serious challenge for regulators and society as a whole. He analogizes this challenge with the cultures of software and hardware engineers, where software is about “if we ship, then we’re going to sort it out” and hardware is about how “scale compounds your problems.”

We talk about biases and decision-making, the connections between AI and casino gambling, why LLMs are like experimenting on people in the wild, how to think about regulating edge cases, ex-anti legal frameworks, Nita Farahany’s idea of cognitive liberty and why product enthusiasm is not a substitute for safety.

This conversation has been edited for length and clarity. For the full version, please subscribe to our podcast.

Danny Crichton:

You’ve described your work as the “Black Mirror” of Thinking, Fast and Slow. So for listeners who haven’t read your book, The Loop, or know your body of work, talk a little bit about your main thesis.

Jacob Ward:

Sure. I was in a documentary production PBS did. We got a couple million dollars from the National Science Foundation to create a kind of crash course on Daniel Kahneman’s work and the constellation of research that came out of it. That show was called “Hacking Your Mind.” And for me, it was a totally life-changing experience. Not only were we meeting all of these incredibly important thinkers in behavioral science, but I was also kind of a guinea pig.

The conceit of the show was here’s a guy who thinks he’s immune to all of this stuff, right? He makes his own choices and has no biases. And then they subjected me to test after test after test, and I did exactly as anybody else would have done. I was the prototypical person who had no idea how automatic my thinking was. I realized how much of it was out of my control and based on these heuristics that Kahneman and others had identified.

At the same time, I was also a technology correspondent for Al Jazeera and at NBC. In my day job as a tech correspondent, I was bumping into company after company after company using these pre-transformer models, ML and neural network, human-reinforced-learning systems to try to identify behavior in people and predict (and sometimes influence) their choices. Some companies were trying to help people lose weight or save money — there were some positive paternalistic ideas there. But in many cases, I was finding people who were working for big gambling companies and trying to shape people’s behavior that way.

As I learned more and more about transformer models — which then came along and made the current LLM language model possible — I suddenly realized, “Oh, we’re about to enter a world in which you have all of these behavioral science findings creating a kind of manual for how people make choices.” And then you also have a whole world of pattern recognition systems that can pick apart that data in ways that human researchers would never be able to and come to predictive conclusions about how we’re going to behave.

And if I knew anything from 20 years as a correspondent studying how businesses make choices, it was that people with the best of intentions wind up preying on human vulnerabilities if they start to run out of funding and need to come to a minimally viable product as quickly as they can.

Suddenly I was like, “Wow, this is an emergency.” Where a lot of people worry about the “Terminator” possibility, I worry about the “Idiocracy” possibility. I worry we are going to wind up amplifying our most ancient, primitive instincts and become, if we’re not careful, a more primitive version of ourselves.

Danny Crichton:

When I think about the last 15 years, I see two sides of the coin. So on one hand, you have a neuro-side focused on gamification and how you take insights from psychological behavior. For example, the idea of offering variable rewards to which humans respond very, very well. And then on the other side, about 15 years ago, Nicholas Carr was working on The Shallows. More recently, Shoshana Zuboff wrote about surveillance capitalism, and Karen Hao came out this year with Empire of AI.

So there’s two sides. One is like, “Here’s how to make money — a guide for how to do it effectively, efficiently.” And then there is an increasing number of books on the shattering of the intentional commons and the inability of folks to focus.

When you think about all this in 2025, three years after ChatGPT came out, how do you start to think about the impact that’s having?

Jacob Ward:

One of the things I’m most concerned about is that there’s some very dated assumptions about how problems with something like this get worked out. What I come back to all the time is the idea that people experimenting on other people in the wild is kind of the essence of a lot of these large language models. The companies behind them have the idea that we’re just going to get it out there. If we ship, then we’re going to sort it out. And that has to do with, I think, people running these companies who came from software. They are people who believe scale solves your problems — over time, scale will work out the bugs.

Hardware people instead will tell you that scale compounds your problems. I once interviewed the CEO of Midjourney and was asking him about some of the problems he’d been seeing on his platform, how much porn was being developed, how much violent porn was out there, all kinds of stuff. And he basically said, “Well, I take this attitude that people are generally good and we have to act on that assumption and we’ll sort of work it out over time.”

And then he said, just sort of thinking out loud, “If I’m somebody who makes muffins — I’ve made 10,000 muffins — and somebody gets food poisoning, am I supposed to stop making muffins?” And there was this awkward silence between us. I was like, “Yeah, dude. Yeah, you’re supposed to stop making muffins. That’s what the FDA would say. You’ve got to find out what’s wrong. An inspector is going to come to the facility. That’s the whole point.”

I mean, I’ve gone to big industrial bakeries. There’s one here in the Bay Area called Semifreddi’s. I went on a tour of the Semifreddi’s factory. I was there for an inflation story, but at the end of the assembly line, he showed me the metal detector all the products go through. And I was like, “Metal detector? Why do you put croissants and granola through a metal detector?” He’s like, “You’ve seen all of the machinery we have here. Well, a bolt is eventually going to come loose and fall into the dough. And if even one person cracks a molar on a bolt in a croissant, then that’s the end of my business.”

So you’ve got people who are in the business of putting stuff out to 800 million people thinking edge cases are such a small percentage of the total, it doesn’t matter. But meanwhile, you’ve got people who make croissants for a living saying, “The edge case is my responsibility. It’s my fundamental responsibility to make sure that never happens.” And that disconnect for me is where I get really worried.

On Monday this week as we record this episode, OpenAI released a big report on the mental health effects we’re seeing, and they said that 0.15% of their users are developing an outsized emotional attachment to the chatbots. Similar numbers of people are developing or showing signs of psychosis and mania in their interactions with ChatGPT. And 0.07% of people are talking openly about suicide with this stuff.

You could look at that and be like, “Oh, well, that’s a tiny fraction of a percent.” But with a product that reaches 800 million weekly users, that’s 500,000 people a week exhibiting suicidal ideation. That’s more than a million people a week exhibiting signs of psychosis. We have to be thinking about the circuitry of human decision-making as if it is hardware, or at the very least we’ve got to meet the standards of food, because this stuff is really, really important. It’s going to change how we think.

Danny Crichton:

At the same time, we’re also dealing with the fact that, in these cases, we have a single piece of software with a very specific ethical, moral — we’ll even call it religious — point of view. It is Western, it’s built in Silicon Valley. It is generally imbued with the norms of the builders of these platforms, but it’s now deployed everywhere.

I wrote about this a couple of weeks ago — why doesn’t your chatbot recommend religion? It’s not going to do that because it’s been trained not to, even though maybe someone should be counseled that way. So I do think there’s this open question of ethical quandaries as you scale up.

Jacob Ward:

I think you wind up in a place where you sort of say what a lot of the social media companies say, which is, “Well, at this point, it’s gotten to be so big and unwieldy. We can’t truly take responsibility for it anymore.” The new fashion for the heads of these companies is to say, “Well, now we’re just a reflection of humanity. And humanity has a bunch of problems. We can’t be expected to solve those problems.”

But now that we’re seeing real effects, I’d like to think that attitude could change. And it’s going to be incumbent for it to change; we haven’t even begun to see how powerfully this stuff is going to seal people off socially, seal people off psychologically. It’s going to invite some new assessment of what “harm” legally means.

In this country, we only like to think about financial harm and physical harm. We regulate against money losses and death pretty well, but we don’t like to regulate anything else, and we certainly don’t like to imagine that people are anything short of entirely responsible for their choices. I’m a former drinker who, when I see “drink responsibly” on the bottom of alcohol ads, it makes me crazy. There’s no such thing. That’s why I had to quit drinking.

One thing — and I’ll be curious to hear your perspective on this Danny — is how ready these companies seem to be to own the short-term horrors that are going to be visited on people on our way to get to some utopia with AGI that’s curing cancer for us and taking us to Mars and the rest.

It’s weird to hear companies say, “Oh, there’s going to be enormous job loss, and tons of scams are going to come out of this.” It’s this weird ex-ante public relations tactic. That promise of utopia is the new excuse that, I think, people are using now to avoid responsibility for the short-term stuff. They just say, “It’s going to be worth it when we get there.” I’m curious what your perspective is on that.

Danny Crichton:

What’s interesting is that AI is the first time I’ve seen all the alignment teams and the super alignment teams with many, many people focused on this. It goes beyond PR, at least in my view. There are too many people out there who are very deeply passionate about this subject. Now, I don’t know if they’re always empowered to do what they need to do.

And unfortunately or fortunately, there’s an immense amount of competition in the AI modeling space. It feels like you’ve just got to keep running to maintain your edge. You can’t take this step back and ask, “how many jobs are going to be affected?”

I bring this up in a lot of the policy work that I do. People can transition, they can retrain, they can do a lot of stuff. We did that to some degree in manufacturing, but it was not fully complete. And that has led to populism and a lot of today’s challenges. That was a 30-year transition. This one is going to happen at a speed we’ve never seen before. 30 years will be three years. There is no world in which people can retrain that fast, and that’s what worries me — just the speed.

We’re just on the threshold of figuring out all the implications for these technologies. But I am curious about this alignment piece. It doesn’t feel like window dressing, but also doesn’t seem like anyone’s really willing to shut things down at this time.

Jacob Ward:

That’s right. I mean, I have a new book project I’m working on. The tentative title is Great Ideas We Should Not Pursue. It is a tongue-in-cheek look at points in human history in which we have had a great idea and said, “let’s not do that.” So I’ve been looking for examples of places where we’ve held back or in some way slowed down the commercialization of a thing because we thought maybe it wasn’t a good idea. It’s really hard to find examples of that!

Jennifer Doudna and the CRISPR revolution — they signed a one-year moratorium on experimentation on human germ lines. That was one example, but that was also one where no one was beholden to financial interest. That was an innovation that came from inside public universities.

Meanwhile, I think about the disconnect between how powerfully this stuff is going to work on our civic lives and how little civics the people making it seem to have studied.

I was at a meeting once — this was pre-transformer models, or right when they were coming out — and I can’t name names, but early founders of some of the big companies were at this meeting. One of them was presenting the idea of a human-reinforcement learning program, where they were having human workers complete some sort of fill-in-the-blank sentences. And he said, “There’ll be sentences like this: ‘I would never blank with a coworker because that would be unethical,’ that kind of thing.” And after they have done enough of those, he then said, “we will arrive at a set of universal human values, and now I’ll take your questions.” And every hand in the room goes up. Some of them are trembling, they’re so outraged.

The first person was this political scientist. She says, “Okay, I actually have three questions. What’s universal? What is human? And what are values?” And the whole meeting implodes. And that was it, because he’d presented this freshman’s idea of how political science works and how morality and philosophy work with the keys to the car.

I think everyone who makes this stuff has good intentions. To my mind, there are really good long-term intentions that are easily distracted by the need to pay off vast capital expenditures. And as a result, I don’t think these companies and good intentions are going to save the day here.

Danny Crichton:

As we’re recording this today, Character.ai said that they will start to do identity checks for anyone under the age of 18. So there’s a sense that at least children should not be part of this story. I’m curious: How do you think about precautionary principles around harm?

Jacob Ward:

I’ve mostly just been thinking about the big dumb hammer of litigation — and that discovery in litigation will reveal some of the techniques being deployed and create a little bit of coalesced legislation around those. Now, I just have to always remind everybody, there’s no federal data privacy regulations in this nation. We have no rules about that stuff.

When you look at a regime like Australia’s, which is a real ex-anti legal framework, we’re going to get out in front of these harms rather than sort them out afterward. But it is so powerful and important to remember that the models are being built by that guy who was in front of that room of social scientists. This is a guy who’s trying to build into the system an ex-anti regime of values that are going to govern this stuff. Anthropic has gone on to create this constitutional AI. They have a similar notion. There’s this obvious idea that they’re going to have to pre-build some rules into this stuff.

Well, if the technology is going to be pre-building all this decision-making, then we have to pre-build some decision-making around the legal liability. That’s going to require real civic debate; what do we agree or disagree on about what’s right and what’s wrong? There are some really smart thinkers about this. I feel like I mention her name in every public appearance I make, but Nita Farahany is a Duke law professor who wrote a book called The Battle for Your Brain. She has this whole idea about cognitive liberty and how we need to enshrine it as a civil right. It sounds very, very gray, but she’s a sharp legal thinker who’s created a real framework for assessing this stuff. She has a whole framework for figuring out the difference between harm and manipulation — and where they overlap.

And I think we’re going to have to get into that world in the same way we eventually had to deal with cigarettes. If you asked your average cigarette smoker in 1957, “do you like cigarettes? Do you want to keep going with the cigarettes?” They’d say, “Yeah, they’re a positive refresher. This is a fantastic experience.” The enthusiasm we will have as people for the experience of the product is not going to be the right measure of whether that product is okay to be selling in the first place.

Something similar is going to have to happen here. We’re going to look at people and be like, “Oh, it’s working on your circuitry.” The loop that this has set you on is something beyond your control. It feels good to you, but that doesn’t matter because it’s causing you to make all these terrible decisions. And here are the ways in which these companies have clearly made these choices internally to pursue this kind of behavior.

Once we get to that point, you’re going to start to see big financial losses. And that’s where I think the rubber is going to meet the road in terms of people changing how they conduct business with this stuff.

Danny Crichton:

I like this idea of the lack of humanistic thought in tech circles. I’ve been in the tech industry arguably since my first startup experience in 2007, so almost 20 years.

I do think people had a much more well-rounded education in an earlier generation. One argument would be, particularly in AI, people had to go deep into their degrees. They had to get a master’s degree.

When did you have a chance to walk over to the English department and read literature and understand the human condition? So I wonder how you’re thinking about bringing that back into the equation.

Jacob Ward:

Man, I wish I knew the answer to this question. I’ve been in these weird conversations with people, where I sort of want to say, “Hey, bring me on as an advisor. I can help you.” I’ve had a couple of companies come to me and say, “Hey, I want to do this the right way. I’ve read your book and other books like yours, and I’d like to try and be on the right side of history.” I’ve had a couple conversations like that, but I really feel like that hydrophobic sand that you put it in water and it can’t get wet. I feel like that around here.

But I like to assume, as the people of the generation you’re describing have kids, they’ll sort of look at the world that their creation makes possible in a new way. I mean, all great breakthrough thinking about how humans evolved, for instance, has to do with long-term generational thinking. So we have to get into long-term generational thinking.

Unfortunately, that’s not the way we are incentivized right now. It’s not even how our circuitry is supposed to work. I got to spend some time with a tribe in Tanzania. They’re one of the last remaining tribes on earth that lives the way we all did 60,000 years ago. They have no last names, they have no property, there’s no marriage, and they are nomadic. They don’t raise crops or anything.

And one of the most amazing things about them is that they have no word for any number beyond five, because why would you need that? In the way that our brains are built, you don’t need to be thinking more than five moons in the future. You don’t need to count the number of people in a group. Once it’s more than five, you can just say, “It’s lots of people.” There’s three pieces of meat or there’s enough for everybody.

And this is the whole thesis of my book — we’re not built to think beyond the circle of our campfire. And that’s where some of our higher functioning needs to come into play. But it doesn’t feel good, and it doesn’t make you money in the short term. And so that’s the essential problem I think we’re going to be facing in this generation and the next.

Don’t miss

That’s why I am excited to bring my good friend Uri Bram on the podcast this week. He’s written a book on Bayes’ theory, has been a publisher of a very successful online newsletter, has hosted olfactory gallery parties, and he just published his first party game called Person Do Thing inspired by trying to order vegan food at a restaurant in Thailand. In short, he’s constantly experimenting with new forms of media and ways to bring people together.

In this episode, we talk about Person Do Thing; communications and the curse of knowledge problems; Amazon recommendations as a social networking tool; and his recent viral blog post, “21 Facts About Throwing Good Parties.”

This interview has been edited for length and clarity. For the full conversation, please subscribe to our podcast.

Danny Crichton:

Uri you are the Dos Equis man of events. You’ve been hosting meetups all around town, you’ve run a publication, you wrote a book, and you’ve joined a lot of our Riskgaming experiences. And now you just launched a new game called Person Do Thing. So when did you start thinking about the game?

Uri Bram:

Person Do Thing started out when I was in Thailand 10 years ago. I was at a cafe and trying to explain that I’m vegetarian despite not speaking any Thai. I was there with a friend who spoke a very small amount of Thai. So we were kind of talking to this waitress, and we’re saying, “No beef, no chicken, no pork.” And there was this moment where you could see in her face she understood. She was like, oh, these people are trying to tell me they want food with no meat in it.

And so we started this conversation about what’s the smallest set of words you could use to describe every other concept? The game is essentially that you have 34 very simple words you’re allowed to use, like “person” and “do” and “thing.” You are given a card, a word you’re trying to describe, and then you try and describe it using only these simple words. So you sound very stupid, but it does, I think, have quite a lot of philosophical depth.

Danny Crichton:

And what do you think that depth is?

Uri Bram:

The game embodies the curse of knowledge problems. You often see people playing badly because they know what the answer is, and the description they’re giving does fit the answer, but the description also fits many other things, and they just can’t understand why the guessers are not on the right track. I think that’s very interesting.

The best way to get an idea into someone else’s head is often not trying to directly and abstractly describe it, but to tell them a story, to narrativize it and kind of draw them along a path. I’m obviously talking my own book here, but I think this game can help you communicate better and think better about other people in a way that’s incredibly interesting and just very fun.

Danny Crichton:

That’s a theme across a lot of your work.

Uri Bram:

Yeah. You know what people say about good in software and good in hardware? Some people are born with natural empathy and the ability to read other people’s minds. Anyone who went to high school with me would say I lacked those abilities as a teenager. And so I spent a lot of my life trying to figure out how to communicate, and thinking about it and practicing it. And I feel like a lot of my projects, you’re right, do tend to bring that together.

Danny Crichton:

Your book is literally called Thinking Statistically.

Uri Bram:

Yes. I wrote this book just after college. I had a really beloved professor in undergrad who I asked one day, “How many of the undergrads here do you think know about Bayes’ theorem?” And he said, “A hundred percent. Everyone knows about Bayes theorem.” And then he kind of stopped and reflected and said, “Maybe not the music majors. So let’s just say 99%.” And I thought, you’re very ill-calibrated here.

I loved this professor, but I thought the interaction explained why so many of the stats classes — the intro stats classes — will just tell you how to calculate a standard deviation. I actually think there’s this separate set of conceptual tools that all statisticians and econometricians have, but it seemed to me that no one was trying to explain them directly. So I wrote this very short book, which tried in a relatively light, fun way to explain those concepts.

Danny Crichton:

You keep going back to this pattern — you’re in Thailand 10 years ago, you wrote this as a post-undergrad. It’s all about trying to upgrade people’s thinking and cognitive capabilities.

Uri Bram:

Interesting. I was not expecting Riskgaming to be a therapy session, but this is working very well.

Danny Crichton:

You also hosted this event, Museum of Scent. Explain that a little bit.

Uri Bram:

Yeah, there’s an amazing gallery here in Manhattan called Olfactory Art Keller, which basically stakes the claim that most art is visual, so why can’t we have art for the other senses? So this gallery is a series of bottles you sniff, and I organized a number of events getting people to come. We had a little quiz, where he would give you a bottle that represented a country or a musician or something and you had to guess what it was. I found that very fascinating.

I also came to an amazing scent event that you hosted, right? With Osmo.

Danny Crichton:

With Osmo, yeah.

They’re inventing new scents all the time now, and scents that no one’s ever smelled before, which is the fun part. Obviously, we always had existing ingredients that you mix, but now that we have the 3D modeling for smell, you’re actually able to go in the inverse direction and say, okay, what is something that we predict you will perceive as a certain scent, but no one in the world has ever smelled before? And so they now have multiple of these.

Uri Bram:

I mean, this is incredible. It’s hard to explain in audio, but this feeling of being on the frontier — I think it just opens up space. Most people, once they are doing something that has been done many times previously, will naturally fall into the same patterns, the same rules. I often feel this with nonfiction books, there’s a format that most nonfiction books fall into. But if you’re in some new open land that has never been explored, you can suddenly make all the decisions yourself.

Danny Crichton:

One of the challenges I have is that, when you have a new thesis, it takes time to persuade. It’s not enough to just expose people to a new idea because, oftentimes, they’ll reject it immediately. So you have to do it slowly. It is the same with games. The different scenes open your mind to a concept that you don’t understand. And then you have these moments at the end, these epiphanies. But epiphanies are really, really hard to come by.

Uri Bram:

Something I’ve been thinking about a lot recently is agreeableness and disagreeableness, and the fact that most of the progress in the world comes from disagreeable people. You kind of have to buck consensus and be willing to say something and have people disagree with you.

Danny Crichton:

I agree with you. Look at Brian Eno‘s idea of a scenius, genius plus scene coming together in one place. Why are there certain communities, like the Vienna Circle or the Bloomsbury Group, that were so productive? It’s not that everyone was collegial and nice and pleasant, and in some cases they were very much not.

But you end up figuring out that arguing triggers folks. If you’re an intellectual and someone tells you you’re wrong, it galvanizes you, it goads you. I mean, one of the biggest challenges I think with a lot of internet writing is that people are passive readers. So you get a lot of page views, but no one ever responds because very few people actually write or respond to it. How many people read something you do and take it seriously and write a riposte? I want more! I want people to just trash the hell out of it. It’s great. It’s fun.

Uri Bram:

One of the things I wish people understood is that you can just email people who have written things on the internet and that except for a few very, very successful people, you’re probably way overestimating how much mail people get and how delighted they would be to hear your opinion about whatever they just put out.

Most of the things I write get no comments, even though people have read them. And sometimes, years later, I have found out that someone read it and found it meaningful, but I never heard anything about this. And then, yeah, you have some highly disagreeable people who spend a lot of time online telling everyone else that everything they do is trash.

Danny Crichton:

I guess this is where I was getting triggered by “disagreeableness,” but having high quality critics is very motivating. It’s one thing to send a flame, but what you really want as an intellectual is someone who takes you seriously.

In your role, you bring a lot of communities together, and that’s why the disagreeable thing is particularly funny because you’re one of the most agreeable people I know. You show up, everyone’s having a great time. In fact, you just wrote a piece on rules for hosting parties that went pretty viral. I saw it all over the place.

Uri Bram:

I will say first of all that it was a sad lesson to me that people truly do love definitive statements. There are no rules for hosting parties, but it was called “21 Facts About Throwing Good Parties.” And I think if it had been 21 opinions, it would’ve done worse, and if it had been 21 incontrovertible truths, it would’ve done even better.

Danny Crichton:

But what were some of the rules?

Uri Bram:

One of my most concrete tips is to always start parties at a quarter to the hour because if you start something at 8:00 people will arrive at 8:30. If you start it at 7:45, people will arrive at 8:00. This is an empirical thing I’ve noticed.

And I think you should do everything you can to get people to circulate. My least favorite kind of dinner is when you’re seated at a big round table, so you can only talk to the two people next to you. I say no chairs, and put food and drink in different parts of the room so people have lots of excuses to go, “I’m just going to go over there and grab a drink. Just going to go over there and grab a snack.”

Tell people who else is coming. I think people rarely go to an event if they know fewer than three people, at least in my social circles.

Danny Crichton:

I also believe in the no chairs rule. At one of our original Lux summits, we had a huge crisis in which there were no chairs for four hours, and the negative feedback on that was very, very high. I have never seen so many ones and twos out of 10 on the anti-chair. But no, I think I’m with you. People tend to get locked in. Circulation is hard. Some people do it more gracefully than others.

In some ways, Riskgaming is very good at circulation because you need to negotiate and you need to run it the right way.

Uri Bram:

And you have a shared context. You have a thing you’re doing together, a thing to start conversations. And yeah, I think shared context is a catalyst for being able to talk about whatever else you might want to think about.

Danny Crichton:

Before taping, we were just talking about the shattering of the attentional commons, which is a really fancy, intellectual, obnoxious way of saying that everyone’s on their phones all the time and are distracted. It’s very hard to get people to focus.

Uri Bram:

For sure. I find it very strange that I’m unable to stop myself from doing things I don’t really enjoy. So I think we’ve all had this feeling when you’re scrolling Twitter or you’re watching Netflix, or whatever, and you’re like, I’m not even really enjoying this anymore, and yet I am compelled to continue. I think there’s a certain property of compulsiveness that is orthogonal to quality and enjoyment.

So yeah, this is a huge problem. If someone else can help you break out of that loop, you’re just clearly better off.

But I don’t know that there’s an individual-level solution. Obviously thousands of people are spending their time figuring out how to optimize the TikTok algorithm to addict me, and as one person, am I able to defend against that? There are supposedly apps that will postpone the dopamine high from various other apps, and yet I don’t use them.

Lee Child, the thriller writer, has this New York Times piece called, I think, “How to Create Suspense.” And it’s one of these amazing pieces where he says, before a certain year, televisions didn’t have a particular feature and therefore they didn’t have to worry about people coming back after the break. Once they did have to worry about people coming back after the break, they started doing something differently. They started asking questions before the break and then giving you the answer after the break.

And what is the thing that changed everything? Well, it’s the invention of the remote control. And because of the remote control, people would switch channels during ad breaks. I’m not sure that I’m doing this justice, but he basically says, anytime there’s an open loop, there’s a question that you don’t know the answer to, your brain will get stuck on it until you hear the answer. And as you’re reading this article, you realize that he has done this to you while telling you about it. He has exactly created this open loop. I basically worry that all of the apps and shows and contents that addict me are in one way or another pulling off this trick.

Danny Crichton:

Well, that’s back to the attention shattering thing, and it’s true of readers, viewers, listeners — we are all very distracted and we want to read 500 things simultaneously. And so we read a couple paragraphs of one thing, move on, come back, all over the place. It’s also true of the writers and the content producers themselves. They’re distracted as well. We are supposed to write a book, a newsletter, a podcast, host events. You get distracted and you get bored.

But I know we’re coming toward the end of the time here. We’re getting close to the end of the year as well. 2026 is around the corner. What are the new ideas? What are the new projects? What are you thinking about these days?

Uri Bram:

Oh my gosh. I am thinking a lot about games. I have a game I’m excited to show you, kind of like the next Cold War.

Danny Crichton:

The next Cold War?

Uri Bram:

Yeah, yeah. I’m not saying war will happen. But the game is basically about technology, space races. It’s about all-pay auctions. So this idea that everyone spends, but only the winner gets anything out of it. I think this is an important dynamic in the real world that games don’t historically model very well. I have a game around it I’m very excited about. I am also excited to spend time with you and go to whatever Riskgames may or may not come out in 2026.

And I’m excited about small groups. We were talking about this just before. I think the optimal social media is a four- to six-person WhatsApp group, and I’m excited to have more of those in 2026 and spend more of my time directly engaging with a small group of other people rather than indirectly engaging with a very large group.

Danny Crichton:

Yes, we were talking about small group chats because there are so many large ones. I’m sure you’re part of these massive chains with hundreds of people, and I mute all of them. I do think there is this model of — if you look at Dunbar’s number, 150 is sort of the number of people we can keep track of. That’s the comfortable number of people we can do. So you can see this in a lot of contexts. A lot of companies in the startup world change culture once you hit 150, because you’re moving from a culture of knowing everyone to a culture of abstractness. I vaguely know everyone works here, but I don’t even know your name anymore.

And so I do think that there’s this movement in social media where we’re going from broadcast to these very, very tight interlinked networks with little nodes. You yourself are a node of a bunch of networks, others are similarly inter-networked, and so things will flow through those networks, but it’s much more filtered than it is on the worldwide web today.

Don’t miss

Kirby is a filmmaker, and he’s most well-known for his work around remix culture in a documentary series titled “Everything is a Remix.” Ale is the co-founder and Chief Design Officer of Runway, a Lux portfolio company that builds a generative AI studio for anyone producing films.

Kirby, Ale, Laurence, and I talk about the current economics of the film industry and the recession underway in Los Angeles, why filmmaking would have changed even without the rise of AI, how to think about remixing through software versus remixing through culture, the coming convergence of narrative forms like films and video games, and what the future of artists will look like with co-intelligences.

This conversation has been edited for length and clarity. For more of the discussion, please subscribe to our podcast.

Danny Crichton:

About a week ago, The Wall Street Journal had an article on the future of Los Angeles. There was a massive dive in the filmmaking industry. Obviously it’s in recession. Jobs are down 25%–30%. A lot of this has to do with the strike two years ago, but there’s also an AI component, which is that filmmaking is entirely changing in the next five to 10 years. The solutions we used in the past seem to not be what we’re going to project forward. So from both of your perspectives, what is filmmaking going to look like in the next couple of years?

Kirby Ferguson:

What a nice softball question to open with. Holy crap!

I’m skeptical about how much AI has to do with it, really. I’m not sure it has much of a role. It seems like it’s more a systemic economic thing. What do you think?

Ale Matamala Ortiz:

Yeah, I agree. When I talk to studios, to teams, one of the things that comes up often is that films have become really expensive to create. That’s one of the main reasons some of the teams are not making more films.

But I do think AI is going to transform the industry. On the positive side, we see that it is going to enable us to create new things that were harder to create before. We’re already seeing it. Somehow, some teams are getting the green light for projects that were not getting green-lit before, because they were either too expensive or too ambitious. So we are seeing it from a very positive side.

Kirby Ferguson:

I think there was just some oversupply going on for a while there, too. Do we need this much Star Wars or whatever?

Laurence Pevsner:

As I think about oversupply, one of the pros and cons of AI in terms of this field is that it’ll empower so many more people to become creators. We already have vibe coding, and we’re starting to get into vibe filmmaking. I’m curious what you think is going to happen as more people get access to these tools?

Ale Matamala Ortiz:

That’s been one of the key things since we started: democratization of who can actually tell the stories. I come from Chile, I dropped out from film school, but I wanted to be a filmmaker.

The idea of creating a film back in Chile was even more crazy than perhaps creating an AI lab here in the United States. Creating a film is hard, it’s complicated even in the traditional places, and we do believe that with these tools, with this technology, we’re going to start seeing more stories being told.

Generative AI is going to be one of the things that will unlock different creators around the world. We ran an AI film festival and we received 3,000 films from teams all over the world. People created films that ranged from five to seven minutes, and those were local stories that people wanted to tell but they otherwise couldn’t.

Laurence Pevsner:

Any favorites from the film festival?

Ale Matamala Ortiz:

The winner is a must-watch. It’s really, really good. It’s a story of how many images you can create within one image. So if you take the math of the pixels that create an image, you can create so many more possible images. It’s a very well-narrated film.

Kirby Ferguson:

I think regular people will probably make the most exciting stuff with these tools. If you look at the history of any genre, any art, it’s poor people, it’s amateurs — like blues music, like jazz, whatever — it’s poor people who are outside the system who look at these things in new ways and create new stuff. So I think whole new genres are going to emerge out of this. New styles of filmmaking and new voices. That’s probably the most exciting part of it to me, rather than just professionals who are trying to get cheaper or make their stuff higher quality or whatever.

Danny Crichton:

I’m reminded of careers like Chris Nolan’s. His original film [Following], even before Memento, was a $100,000 budget film he did in school. And then Memento was also a very, very cheap film. He’s progressively grown into these massive Hollywood budgets. But to your point, democratization means that there will be many more Chris Nolans.

Kirby, you’ve been a filmmaker a very long time, but one of the themes that comes up frequently in your work is remixing, this idea of recombining previous material. Remix culture is something we see in certain fields, certain subcultures. But in others, it’s verboten, it’s considered offensive, you’re not an original artist. Talk about AI and remix. What do you see as the future of the culture there?

Kirby Ferguson:

It’s super fascinating, because AI is a remix, of course. It’s deep learning, it’s based on training on existing stuff and then creating new stuff out of the old stuff. I think a version of that is what we do for all sorts of creativity. So it mirrors us, in a way. There’s obviously a lot of stuff that is not in the AI mix — mysterious aspects of the human imagination are not there. But there’s this interesting parallel between the two, between of how it creates things and how we create things.

So it’s interesting just learning about how AI works and using it, because I do think it is a mirror of how we work. We don’t create out of nowhere, we learn music, we read books, we watch movies, and we start copying what other people are doing, and then we start putting our spin on it and combining it with other things we get from other places.

Laurence Pevsner:

As we think about AI in film, what do you think even defines it? We’ve had CGI forever, and that’s a form of using computers to advance film technology. Do you think there’s a distinction, or are we just looking at a curve of different kinds of advancing technology that we’re going to keep using as filmmakers adapt and use new tools?

Ale Matamala Ortiz:

At the AI film festival, we have a preface that, “some AI was being used to create this idea.” But I don’t think we’re going to be talking about AI films in the near future.

I do believe that as this technology progresses with new capabilities, such as real-time interactive outputs and multi-shot sequences, we can aim to have different kinds of stories — different ways of telling stories — that are perhaps going to blend what we know today as films and what we know as video games. I think then we’re going to start talking about new mediums rather than whether something is made with AI or not.

Danny Crichton:

As a filmmaker, when you think about authorial vision, we have this focus on how the director is in charge of the film, all the shots, all the framings. When you think about AI, does it empower the authorial vision as a collaborative agent for you?

Kirby Ferguson:

I really like Ethan Mollick’s book Co-Intelligence on this subject. I like the idea of AI being a quirky, brilliant, weird collaborator. It’s brilliantly good at some things and incredibly inept and stupid at other things. I don’t think of it as collaborating with a machine, though, I’m collaborating with the entire history of human expression. Everything that was on the internet that it got trained on — I’m collaborating with that. So to me, it’s this social thing, where you’re collaborating with the human imagination itself.

Ale Matamala Ortiz:

Yeah. I see different layers. One is, as a creator, the way we design our tools and our software and our models is to make them collaborators, to empower artists to do the things they want to do and help them achieve their ideas. We’re not creating or designing systems that are one-pronged, one shot, give me a final output. But I do realize there are some use cases that might be very powerful on that end.

But even within the creative workflow, the collaborator can also expand to be something that can be more autonomous, one that gives you some results that perhaps you were not thinking of. Or, maybe it’s too complicated to do, or maybe you just need someone else who can run and do that thing. So on a feature film, a creator might have some ideas — this is the place, here is one scene — and they want to have fine control over that scene. But perhaps within that same feature film, there is a sequence of a car going through Manhattan that is one-minute long. Maybe we can use systems that are more autonomous, that can create multi-shot scenes, that can deliver a longer take with different scene cuts, that can also work for that creative.

On the other side, more on the consumer side of things, I think there is a future where we will be generating videos to consume on-demand for different use cases — for learning, for support, for entertainment in some cases. These are perhaps going to be a little bit more disposable, but they’re going to be run on their own by a machine.

Danny Crichton:

With films now, they are definitive — specific. They are the way the creators want them to be watched. With AI, you have this capability to say, not only can a film be in different formats, it can be in different languages, we can auto-translate, it can be dubbed, including with the right lip movements. And you could potentially change plot lines, it could be adaptive, the story itself could change, the film could change.

Does this worry you? Does it change the definition of filmmaking? Do films go from a singular vision to something where the film is more of a world, an environment — something you can explore?

Kirby Ferguson:

I think the scenario you outlined there, where AI can translate and convincingly convey it to cultures that don’t speak your language, that seems like a total win for me. Anything that can extend the reach of a filmmaker is definitely a huge win, so I love that aspect of it.

Ale Matamala Ortiz:

To me, this is one of the most interesting opportunities, to be honest. We shouldn’t be designing this type of technology to continue doing everything that we have done in the same way as in the past. We should push for new ways of working and new ideas and new ways of telling a story.

Today we have films that end, but why do they have to end? Why can’t they keep going? Why can’t a viewer take action and continue the story, or maybe see the story from a different angle? Why does a TV show have to end? Why do we have only one season finale? Why can’t we have the same season finale from a different character, from a different point of view?

Those are the opportunities that we can start considering now. I’m not saying that we are there yet, but these are the opportunities that we should be looking at.

Don’t miss

For more than a decade, Halldór Berg Harðarson lived and worked in China as part of EURAXESS, an initiative of the European Union to connect European and Chinese scientists together to accelerate frontier research. From the heady and optimistic early 2010s to the serious challenges of Covid-19, Hardarson saw it all live from Beijing — a far cry from his home fishing village in Iceland. Today, he works at a biotech unicorn in Iceland called Kerecis, which uses fish skin for tissue regeneration.

Alongside me and Riskgaming scenario consultant Ian Curtiss, we talk about what Iceland can learn from China and vice versa despite the massive population gap, the transformation in the European Union’s relationship with China, and we throw in some optimistic notes at the end for a nice aftertaste.

This conversation has been edited for length and clarity. For the full interview, please subscribe to our podcast.

Danny Crichton:

You two are both long-term China hands. You met in 2011, I believe. How did you meet? What was the environment like back then?

Ian Curtiss:

One of my first memories of Halldór is, I think it was in Pan Wei’s class and we were debating human rights and individual liberties and how the state serves individual liberties in different ways. Halldór tells the professor something, and the professor responds with, “Your country has 300,000 people. That is the population of one district in Beijing. I’m not going to listen to you.” I remember just everybody in the room was like, oh, whoa, this is a different perspective.

Halldór Harðarson:

I enjoyed those classes quite a bit, actually. It’s very uncommon to get the chance to debate someone with such a different frame of reference for almost everything. Really, really deep thinker, but almost everything, all the assumptions, were very different from anything I would ever assume about the world.

It was a totally different world back then. People seemed a lot more optimistic about China. It was almost a given that there was a transition happening, things were slowly developing for the better. When it came to the leadership transition that was around the corner, people assumed it would be a peaceful thing.

Ian Curtiss:

I remember in our classes, one of the professors was referencing local county-level elections as an example of progress and change in China and the building momentum toward democracy. It was very much a discussion at that point.

I joined the American Chamber of Commerce after our degree, though, and it was fascinating watching how dramatically the tone shifted from 2013 on. I’m curious to hear your take, Halldór, from the EU perspective, but when I started, it was overwhelmingly, keep the engines going, everybody’s making money, everybody’s really happy. But when I left, it was the business community that confronted the Obama administration. If anyone remembers the U.S.-China Bilateral Investment Treaty that was being negotiated, it was the business community that confronted the U.S. Trade Representative and said, “Actually, I don’t know if this is a good thing because so much weird stuff happens once firms enter the market.”

Halldór Harðarson:

Right after university, I started working on projects for the European Union. I ended up leading one for more than a decade. It was on innovation, research, and science. I remember really distinctly in 2015, at the early Horizon 2020 research framework program from the European Union, we had our Director-General of research and innovation coming to China. He comes and we have this big event sort of welcoming China into the research framework program. The main slogan of Horizon 2020 was, “Open to the world,” and the idea back in those days was that it would be a great benefit for the whole global system of science to include China, which arguably still is.

But it was undoubtable in those days that Chinese science was not very significant. It was growing, but there were not a lot of significant contributions yet. They were spending only a fraction of what Europe and the United States were spending on science. They were also not getting super good results.

The whole idea was to reach out, in what we then assumed would be a non-competitive arena, like basic science, to sort of build up the basic knowledge for the good of all humanity. Some of the biggest challenges on the planet were in China, and you were not going to solve things like climate change without working with the Chinese. That was the tone in 2015.

Danny Crichton:

Coming from Iceland, a country of 300,000, into a country of 1.3 billion, how was that transition for you? I feel like you’re a unique entity here, because you stayed all the way through 2024 — through Zero-COVID and all kinds of different dynamics. How was that experience? Was this a burning building but it’s slow and you sort of realized like, oh, the smoke is getting into more rooms, but it’s okay in this one?

Halldór Harðarson:

COVID was definitely a canary in the coal mine. The experience really revealed the true colors of the system, for better and for worse. When the Chinese want to stand together to do a certain thing, they have an extremely effective culture for doing so. They have institutions within their political system that can be activated at a moment’s notice to mobilize huge initiatives.

I know people in the West were really relatively unhappy with their governments, but in China, it went way beyond anything you could imagine. There were a couple of times where I woke up and my door was locked from the outside because I had been quarantined without being told what was going on. I was just locked in my apartment.

But because of my topic of expertise, the experience was also super revealing. One of the things I’d been working on was this network of European researchers in China. The idea was to support researchers’ mobility. So, researchers are coming to China and Chinese researchers are coming to Europe and so on. But COVID was a bit of a slap in the face.

Just before the closures, I had just finished doing a big survey on all of my researchers, so I had this really lovely dataset about the community. Then in March 2020, we found out like 75% of the European scientists in China were locked out of China. This is clearly bad for China, you would think. China’s trying to attract these researchers, right? It took decades to build up the institutions to attract these researchers.

I did another survey in August that year. I was like, okay, now COVID is over. It wasn’t, but I was also wondering, is this all about COVID? What’s going on here? In the end, I surveyed the researchers again in 2022, 2023, and this year. My main hypothesis now is that COVID accelerated a development that was already happening behind the scenes. When China finally opened again in 2023, there were fewer than half of the European researchers there used to be. This year, I also found that there’s been no renewal.

So, what happened? If you ask the European scientists in China, they would say COVID happened. I don’t know about you guys, but COVID is ancient history by now and people not coming has very little to do with what happened four or five years ago. There’s something else going on, and that’s really remarkable in my opinion.

Danny Crichton:

It really does feel like the world has changed. There was this idea that there’s sort of a flat scientific community with no walls and no barriers, that all the data can flow all around the world. Now we seem to have moved on to a world of walls and barriers and borders. How are you seeing it from the European Union perspective? How do you see the EU balancing between the two sides?

Halldór Harðarson:

The liberal world order goes really deep in the DNA of the European Union, but that has been changing.

I don’t think China is the main reason, though. The big thing that happened was the invasion of Ukraine in 2022. And that really is changing European attitudes toward what it means to be European and a European Union.

But even before that, and even in my job working with China on science and innovation policy, attitudes were already getting cynical. One big thing was that we should probably not be working with them on things where we are directly competing. 5G was the big thing back then. It was hard because some countries were having really useful collaborations and other countries were very much against it.

There were also countries saying, “China is becoming really good at certain things. We should maybe stop having this colonial attitude where we are over here handing over infrastructure and helping them out or whatever, and maybe it is better to think about them more like a peer. We can teach you certain things, but what can we learn from you? What are things we can take out of this relationship?” That was really new.

And then down the line, especially after 2022, the attitude started to really, really center on the idea of risk. That’s the word Europeans love to use. The Americans used the word decoupling, and the Europeans talked about de-risking, which indicates that you still want to work together but you want to be responsible about it.

The practical effect was that in Horizon Europe, which is the framework program after Horizon 2020, we basically weren’t targeting Chinese participation in any area anymore, except for in agriculture (mostly biology) and climate change. I believe they are still working on those.

Ian Curtiss:

This is such a great case study, returning to the civilizational discussions of our degree. It’s a great case study of realism versus liberalism, and liberalism can only exist in a system where the rules are completely clear and enforced equally. Only then can everybody agree to get along.

But when you think about the Chinese perspective, what they would always say to us was, “Well, the liberal world order is just industrial policy as written by the United States government,” meaning open borders allow the better U.S. products to get into everybody’s economies. Also meaning that IP protection is really just the wealthy country’s way to keep their advantages.

COVID was such a critical moment that opened people’s eyes, changed their perspectives — however you want to describe it. They realized that maybe politics is involved in so much more than we thought.

Halldór Harðarson:

Yeah. I could go very philosophical on this. A concept that really helps for me is that you have traditional land powers and you have mercantile naval powers. And so you have the British Empire and then the American world order on and so on. And these powers, they thrive on trade. But countries like China and Russia, they don’t feel safe relying on that alone. I’m also from a small island with a huge moat where we have extremely high GDP purely from trade. And so we think, why don’t we all get along and just do that?

I’m really, really sad about how things are going these days. I feel like we were really heading toward a cliff at the beginning of the 20th century with the two world wars. The international order was extremely precarious, and we were not really sure how to go about fixing things. In the end, we ended up with very wise people, who had experienced a lot of things, building institutions that worked out really well. They definitely worked out for the United States and the West. But they also worked out really well for China. When they decided to be open to the system, it worked really, really well for them. So, it’s sad when people start tearing that down.

The big hesitation I have toward China is that their system is almost designed to tear down these institutions, no matter what they’re saying. That being said, the U.S. administration is also not doing its best to keep those institutions alive and maintain them. I have regrets on both sides of that aisle here.

Danny Crichton:

Look, a lot of institutions are buckling. Many were made in the post-war era, and a generation of the people who built those institutions — maintained them for a long time — are passing away.

Today, you look at the United Nations, and just in the last couple of weeks because of UNGA, there’s a spate of articles that the UN has never been in a worse position on its 80th anniversary. Its budget is a mess partly because of U.S. underfunding. But it also feels like it’s an institution that just isn’t reflective of the world today.

I think we’re seeing that kind of pressure across the board, and it comes from two places. In some cases, the institutions are out of sync. I wouldn’t say out of date, but out of sync. Power has shifted economically, militarily, whatever the case may be, and so they’re just not reflective of the new balance of power. And then in other cases, I think there is a new generation of folks who are coming up who just don’t have experience with World War I, World War II, or the Cultural Revolution.

I don’t know if that changes. The war between Russia and Ukraine, the war with Israel, tensions over Taiwan, the Philippines and China. Are these experiences going to create a new generation of folks who say, “God, we’re doing it all over again. We’re getting closer to the brink and we won’t get as lucky as we did last time.” My hope is we will learn from experience — from close calls — and we won’t go all the way over the cliff.

Ian Curtiss:

To tie this back to really current events, there are all these ongoing negotiations and big promises that administrations keep making to their constituents. My prediction is that a lot of people are going to walk away disappointed because there are things going on in the world that, to your point, no single administration can change. The American president might be the most powerful person on the planet, but they’re still only representing 350 million-ish people out of eight billion. In terms of global economic growth, we’re just not the most powerful engine, relatively speaking. The lever just isn’t there, there’s just not the ability to control outcomes like American constituents expect.

So, dealing with China and dealing with issues like tech controls, export controls, and the future of U.S.-Chinese competition and so forth, I think one thing that’s going to be really interesting is people understanding that we can’t win everything anymore. So therefore, maybe we do want more global systems to limit the outcomes when somebody else wins. That’s the traditional game theory approach to political systems: you want to mitigate the downsides. For a long time, I think the Americans have had less downside risk, and as a result, there’s a lack of appreciation for the system that existed.

Danny Crichton:

Look, I mean, the other part of game theory is when you have two superpowers, everyone else is in the middle. I mean, just yesterday as we’re recording this, Canada is saying it might allow Chinese electric vehicles into the Canadian market, which would be the first time Chinese cars come through North America. I think we’re starting to see that in the European Union as well, where they’re looking at these anti-coercion measures.

But let’s end on a positive note. You started in China back in 2011. It was a pretty optimistic, positive time. That feels like distant history. What are we happy for? Ian, you just hit a huge success milestone outside of your Riskgaming work here.

Ian Curtiss:

Yeah, thanks, Danny. So part of a passion project of mine is getting some B2C games out there in the spirit of getting people to simulate and experience systems outside of the ones they’re used to. And so, we just got a board game funded on Kickstarter. It’s called Iberia: Kings and Emirs, and it is all about Medieval Spain and the complexities of religious identities at that time. So, it is a good, light topic, but the game really shows how politics is always so nuanced on the ground and the stories we tell about historical periods often don’t reflect the human experience of what actually happened at the time. So, it ends up being quite a lighthearted experience.

Danny Crichton:

Well, congratulations. And Halldór, you’ve switched over. So you had this extraordinarily long career in China, but now you work at one of Iceland’s most successful startups. What are you working on today?

Halldór Harðarson:

Yeah, I work for a company that is doing tissue regeneration and tissue repair in the medical field, it’s called Kerecis. It’s really, really cool, and people should look into it. I think my work there goes to one of my fundamental beliefs about the world. Ever since I was very young, I just really, really thought technology would save us all.

I grew up in a small fishing village here in Iceland. I would’ve never thought that the most valuable part of the fish we catch would not be the fish itself, the meat of it, but the skin. The company I work for takes that skin and uses it to regenerate tissue, and it sells it for thousands and thousands of dollars, these little pieces. We’re now entering an era where incredible things are going to be possible — amazing things.

So, what I want to leave this conversation with is this: We talk about collaborating with China in science and research, or not collaborating with them. But I guess what we should ask ourselves is, can we put a little more responsibility on them? Ask them, “If you want to collaborate with us on the future, on building this future, could you please contribute here?”

Instead of talking about tit-for-tat and all of that, if we are serious about the global challenges we want to tackle together, we should not be afraid to ask a lot of each other. And if people are not able to contribute as much as they take, then we need to be a little bit selfish while we rebalance.

Don’t miss

Even so, there remains prodigious work on the research frontiers of artificial intelligence to identify ways of improving model performance, merging models together, and ensuring that training and inference costs are as efficient as possible. To that end, we brought together two stars of the science world to talk more about the future of AI.

Kyunghyun Cho is a computer science professor at New York University and executive director of frontier research at the Prescient Design team within Genentech Research & Early Development (gRED). Shirley Ho is Group Leader of Cosmology X Data Science at the Flatiron Institute of the Simons Foundation as well as Research Professor in Physics at New York University.

Together with me and Laurence Pevsner, we talk about the state of the art in AI today, how scientific discovery can potentially be automated with AI, whether PhDs are a thing of the past, and what the future of universities is in a time of funding cuts and endowment taxes.

This transcript has been edited for length and clarity. For a full version of the discussion, please visit our podcast.

Listen now

Danny Crichton:

Thank you so much for joining us here at the Lux AI Summit today. Your panel was about the technology frontiers of artificial intelligence. Give us a little summary of what you just discussed on stage.

Shirley Ho:

There was a huge range of topics, but most importantly was how do we get from where we are now — which is using a lot of specialized tools and LLMs — to true scientific intelligence. What are the roadblocks? How do we make it systematic? How do we actually create the data set we need?

Kyunghyun Cho:

Yeah, absolutely. We talked about needing to distinguish between solving the problems we know and just need to solve more efficiently versus solving the problem of discovery. Scientific discovery is very unique in the sense that it’s all about finding something that has not been found before. How do we make AI internalize this whole process?

Laurence Pevsner:

Right now, if you tell an AI, “I want to get better at discovery,” what does it do? Where does it fail? How is it helpful?

Kyunghyun Cho:

At the moment, a lot of focus has been on how these systems are trained. What they are often encouraged to do is answer questions based on the known knowledge. Sometimes we’re going to let AI run small experiments, in particular in the coding agents. But these are extremely narrow. What we actually want, though, is for the AI systems to know the process of discovery and make suggestions on what we need to do in order to provide AI systems with more data so that the information gain is maximized and then, eventually, future success is maximized as well.

Danny Crichton:

We talk about AI for science in this general format — as if there’s some sort of generalized “scientist.” Yet when we look at human society, we see the complete opposite. People get narrow PhDs, postdocs get narrower and narrower and narrower. So where do you stand on general compute versus more specialized LLMs models and data sets?

Shirley Ho:

There are multiple answers to that. The first answer is that I think we can get AI to become more generalist. Humans are doing all these specialized things. AI could learn everything, but it has to learn not just from the literature, but also from real data. I think that’s the piece that’s missing right now: AI has all the literature, but that’s not how you get discovery. If you talk to any scientist, it’s not that you just read all the papers from before. You actually need to go create new experiments, get more data, analyze it and come back.

A lot of our research has been around creating a polymathic model, which can go across many different areas — from physics to chemistry to biology and astrophysics to everything all together. And I think that actually helps the system become more generalist in a way that can bring new ideas from physics to chemistry, from chemistry to biology and maybe to something else, like health. So that’s my hope in the near term.

Further down the line, I will want to see whether AI can come up with new ideas without a human’s help. But that’s definitely more futuristic.

Kyunghyun Cho:

When we think about AI or AGI, we are very much constrained by what we think we know how to do. When in fact, of course, AI is built in a very different way from us. So we do have to specialize individuals and then form a society in order to cover different aspects of the very difficult problems people face. But AI doesn’t really have that particular constraint. In fact, AI systems are often worse at a lot of things we know well, but can solve problems we just don’t even know how to approach.

One example I can give you is in drug discovery. In that field, there are people who are specialized in trying to figure out human physiology. There are people specialized in a particular molecular modality who work on designing drugs and developing them and optimizing them. There are people who are working on the commercial side as well, and then there are people who are working on the clinical trials. All those people are extremely specialized, and it’s really difficult to see what happens across these stages and make connections. This lack of connection is one of the reasons the success rate is so low, end-to-end.

But AI doesn’t have those constraints. It can, in fact, look at all the data coming out of all the different stages and identify a tiny bit of correlation across them. And then that correlation becomes something we can use in order to improve the success rate. So what that means is that the distinction of the generalist versus specialist applies to us, but not necessarily to the large-scale models we are building now and will build in the future.

Laurence Pevsner:

It’s like that famous metaphor about the blind men trying to understand what an elephant is. They’re all feeling different pieces. Hopefully a generalist polymathic AI can actually see the whole elephant at once. I’m wondering: Is polymathic AI used right now? Is it helping in your research currently?

Shirley Ho:

We recently built a model. It’s a fluid dynamics model that captures everything from blood flowing through an artificial heart all the way to oceanography and astrophysics. So fluids across all the different scales, from smallest to the largest.

And you’re like, “Why did you build such a model?” Well, you can take this model and fine-tune it — give it a little extra data on something it’s never seen before. In this case, something that explodes: an exploding star. There are only five simulations of exploding stars in the entire world. It’s very expensive. But now I can make predictions with just one example, which was impossible before. Before you needed 10,000 examples, a million examples. Now I need just one, and I can already make predictions.

That is a huge step, and this is a great thing about foundation models where data from nearby disciplines can actually boost performance or make the impossible now possible.

Danny Crichton:

In the world of science, it seems like we have a bunch of stuff happening all at once. On the one hand, we have terrifying cuts to federal funding and challenging finances from the endowment side because of endowment taxes. On the other hand, we have a rebuilding of the pipeline for scientists. The idea that you would have AI biologists didn’t really exist 10–15 years ago. Now, it seems like AI is the first step to becoming a frontier biologist or astrophysicist.

How would you start to rethink the future of research careers in terms of skill building, in terms of what you’d be focusing on? Does it change radically?

Kyunghyun Cho:

I think it should stay the same. The existing system has a huge amount of merit. We want to be able to choose as a society to do work that is not blindsided — or narrow-sided — by profit motives. That’s how we originally decided taxpayers would essentially outsource this selection and execution to the federal government. And the federal government outsources this research to the universities and other non-profit research organizations.

I think this is probably the only way in which we can actually continue to invest in the future of research, not the research that’s going to be developed into a product today or tomorrow or next year. Those things are going to be done by industry.

The issue here is not that this system is wrong or outdated, but unfortunately this whole system started to be somewhat overshadowed by shiny new tools in a shiny new industry with a shiny new amount of money.

If you look at some of the National Science Foundation solicitations in the computer science field over the past few years, it’s really difficult to tell whether the solicitation is for research or for developing products. And that’s a mistake by the federal government, because the federal government was supposed to take the taxpayer money and try to invest in the things that companies would have overlooked so that we can protect our future.

What I think is really important is making sure the old system works by ensuring that the execution and choice of the topics is least influenced by what is being productized today or in the next few years.

Shirley Ho:

Yeah, this is interesting. There’s this idea that you should fund frontier research, and the government should not decide where the frontier is. But I think you were also asking about another issue, which is how you prepare future generations to make best use of all these shiny new tools. Should they just be, like, just forget it, don’t learn software engineering because there’s no more software engineering jobs.

I would say that people should do what they love, because then you have the most motivation and creativity pouring in that direction. But on top of that, having some idea of what tools are available and what tools might be upcoming will be super useful.

Kyunghyun Cho:

When it comes to programming, even though we have amazing coding assistants and whatnot, I think we have to teach programming to everyone as early as possible. In my view, programming is not only amazing because it gets you better products and improves your productivity. But it is also a good way to test the logic of your thoughts. You learn how to think about things, you learn mathematics, history, social science, all those courses, but you never actually get to execute. You never get to use your logical thinking and then see the consequence. Programming is the only place where you can.

Laurence Pevsner:

Yeah, when we were talking about polymathic AI, I was wondering if one of the benefits will be that we humans, too, get to evolve. In a future in which an AI can really go deep, we can also become more generalist, more interdisciplinary. We’ll all have a bit of programming. We’ll all have a bit of astrophysics, we’ll all have a bit of biology. That way, we can actually work with the AI models too and be able to think this way.

Shirley Ho:

I do love it as an educational tool, but I want to be slightly controversial. I’m just thinking that programming is not the only place you can test your logic. In a lot of lab experiments — physics, chemistry — you can actually hypothesize and then go test it out. But programming is probably the fastest and easiest.

Danny Crichton:

We’re coming towards the end of 2025 and one of the most important AI conferences is coming up with NeurIPS. I heard you have quite a few papers accepted, so congratulations. I have two questions: One is, what were some of the developments over 2025 that you think were overlooked? And two is, when you look forward either to NeurIPS in a few weeks or into 2026, where does the AI research agenda go next?

Shirley Ho:

I’m not going to talk about our papers! I want to talk about some other people’s papers! There is a huge amount of literature on how to merge models, how to understand models, and how to steer models. All that is usually classified as “interpretability.” And as a scientist, I think it’s great if we can understand the models a little bit better, because if you tell someone, “I found a new fundamental rule about the universe, I have no idea how it came about,” no one’s going to believe you. So can you somewhat interpret the model? I think it’s super useful.

Kyunghyun Cho:

What I’ve seen, not necessarily this year, but over the past few years, is the trend of acknowledging that we have built this amazing correlation machine. You feed in as much data as you want and it’s going to capture every single statistical correlation that exists within this data.

And now we are actually going to the next level up. So if you think about learning or inference, what is the next step? After correlation is causation. And causation is really important. It connects to what Shirley just said because, if we want to be able to control or steer any of these systems, we have to have a causal understanding of the system. Otherwise, we’re going to make all those random mistakes.

So how we get to large-scale, high-dimensional causal analysis is a big trend. It’s all about framing the problem in a way that maximally benefits from this amazing correlation machine. We’re going more into meta-learning or meta-inference, and that’s what I’m interested in. It naturally connects, of course, to discovery. Without causation, you cannot actually discover things very easily.

Don’t miss

This week, I talk with the editor Rwizi Rweizooba Ainomugisha. He’s a wargaming fanatic based in Uganda, where he first started learning the gaming world in high school when he built an underground casino. He then discovered gamification while working in B2B marketing before eventually entering the wargaming world as a player and now as a designer of two micro-games. He’s also a co-founder of the gamified fintech startup, Lupiiya Books.

We talk about Rwizi’s background, how gamification is infiltrating all kinds of different fields, the rise of wargaming and the artificial delineations that still plague the field, the ebb and flow of strategic thinking, and finally, why solitairy games offer a unique on-ramp into this world.

This interview has been edited for length and clarity. For the full conversation, please subscribe to our podcast.

Danny Crichton:

So, Rwizi, you’ve launched a publication called Wargaming Weekly on Substack, covering this extremely niche part of defense policy planning and strategizing. I’m curious, how did you get into this field? It’s not something you sort of stumble upon easily.

Rwizi Rweizooba:

It’s especially peculiar given my location. I’m literally in Uganda, and here I am having a lot of opinions about what the U.S. Department of Defense should do.

My story of wargaming goes all the way back to my high school days. I had the blessing and the curse of attending a Catholic high school — arguably, the best Catholic high school in my country, St. Mary’s College Kisubi.

Typically, with Catholic high schools, you develop a double character. You have the character you present to the school administration and your parents, and then there’s who you are with the students. So for me, my naughtiest adventure as a student was when I started and ran an underground casino.

Danny Crichton:

That’ll get you rapped over the knuckles by the sisters.

Rwizi Rweizooba:

For us, it was the brothers! Looking back, it was a really risky adventure because, if I had been caught, it would’ve meant automatic expulsion. But the reason I did it was this: In our leisure time, we had a bunch of guys who used to play cards. I was never good at cards, and I wanted to have something where everyone was on a level playing field. I figured if we could start playing poker, which none of us understood, we’d all start from zero. So I went to the school library, picked up a few books, bungled my way through the rules, got a bunch of my friends to buy into joining me on the weekend. I was the dealer, they were the gamblers. I would basically guide them through my understanding of the rules.

Fast-forward a few years later, and I had dropped out of university and was working as a B2B marketing writer on Upwork. This was just a few years after the 2008 crash. Remote work was picking up, and I was making money, good money, as a B2B writer. Fast-forward about 10 again, and I met my most interesting client to date, the Octalysis Group. They’re basically one of the top, if not the top, gamification consultancies in the world, led by Yu-kai Chou, who’s the face of the company.

They hired me to do some SEO content writing, and I was hooked. My contact at Octalysis would give me assignments like content on gamification in healthcare or gamification in finance, and I’d research those. Eventually I asked him, “Why don’t we do something about gamification in government?” And he was like, “We’re not really interested in that because we are not interested in government clients. Their procurement process is lengthy and what have you.” But for me, it stayed in the back of my mind.

One day on my own time, I ended up typing in something like gamification in government or gamification in policy. And that led me to discover RAND’s wargaming. It blew my mind on so many levels because in my mind I was like, “These people were playing games about the end of the world, the most serious thing in the world, and they were figuring it out through games.” And that led me into the Georgetown University Wargaming Society’s YouTube channel, which is the most comprehensive online database of professional wargaming knowledge.

Last year, I started having a bit more time on my hands. AI has kind of killed the content market I’ve been working in, and I had reached a stage of my knowledge consumption where I felt like I needed to give back. So I started the newsletter. My initial thought was, “Let me just curate. Let me just create a weekly breakdown of buckets of wargaming information.”

But from there, I was like, “Okay, you’re not a wargamer if you’re just reading and writing about it. You have to play some games.” And then eventually I was like, “Okay, you’re not really a wargamer if all you’re doing is writing, reading and playing. You have to design.” So I started working on a wargame about Uganda’s heads of state. I put that out this year. And then I also put out a second game called African Election. So for me, this journey is all about my imposter syndrome and trying to prove to myself that I’m a real wargamer — but I keep moving the bar.

Wargaming Weekly

A weekly newsletter that explores the latest news, views and innovations in the world of wargaming

Danny Crichton:

It is interesting to hear that this started with poker. To me, poker is so static; the rules are fixed and you’re just trying to optimize your own play. The power of wargaming (and your own evolution) is the breadth. There are so many different domains, so many different tactics. I just love the diversity that comes out of this category beyond what you get in the poker world.

Rwizi Rweizooba:

I have to say, I quickly realized that I was not that good at poker. I was just as mad at it as I was at cards. But I love being a dealer.

What you said about wargaming is true. When I initially entered it, I thought, “Okay, here’s a nice small niche I’m going to be able to understand and cover within at most six months.” I quickly discovered that it’s huge. There are so many branches to wargaming, and I am going to have to find my own nice corner, get comfortable there, and give up on trying to eat the entire wargaming world because I can’t.

For example, there’s commercial wargaming and professional wargaming. I’ve kind of veered toward covering professional wargaming. But even within professional wargaming, there’s military wargaming and there’s business wargaming. Within military wargaming, there’s all these branches of educational wargaming, analytical wargaming. For me, it’s just a never-ending curiosity chronicle.

Danny Crichton:

I’ve been to a number of wargaming conferences over the last couple of years, and one of the things I will say is that there is this opaqueness, this lack of transparency, that makes the field seem much more complicated and much more cloistered than it is. There’s no need for it.

When I think of these games — the commercial wargaming market, the professional market, stuff that the Pentagon does that’s classified — they’re all part of the same spectrum even with different goals and different motivations. But whether you’re targeting business, military, diplomacy, whatever the case may be, at the same time the goal is always to try to help people make better decisions under uncertainty.

Rwizi Rweizooba:

My mind is blown by the fact that games, which are the silliest thing — literal child’s play — are being used to investigate the most serious topics of our time. It’s really coming full circle, like the most silly thing meeting the most serious thing in this synthetic environment. It all makes perfect sense at the end of the day, but when you step back a bit, it’s hilarious.

Danny Crichton:

Yeah, one of the biggest challenges is pushback that you can’t have fun because it’s a serious topic. I have a completely different view, which is for adults, the reason we play is it allows us to explore experiences that otherwise we have no creativity to bear on. It opens up the aperture for us to say, look, I’m a CEO, I’m a general, I’m whatever I want to be. What decisions are possible?

I still have to defend gaming. I think the skepticism is built in — that there’s absolutely no way to have fun when you’re talking about war — but it’s also not to say you should be flippant about the subject. Obviously, there are real consequences to some of the decisions that come out of this. But if you want those decisions to be the best, the ones that improve people’s lives the most, having the creative flexibility of mind is critical.

Rwizi Rweizooba:

I’ve noticed that a lot of wargaming articles these days about trying to convince people that, hey, it’s okay to play a game to investigate these things. I look forward to a time when more wargaming articles are about the technicalities of wargaming, but it feels like we still have to work through this initial stage where it’s just about convincing people that we need to be wargaming again.

For me, that’s interesting because I thought it would be a given.

Danny Crichton:

The other challenge is that there are just so many threats around the world, so much operational intensity. People are overwhelmed with the number of decisions they already have to make.

If you look at the history of wargaming since its genesis in the 1800s, the reality is it seems to go up and down based on the number of crises in the world. The last surge for the Pentagon was around the late Obama years. Robert O. Work was a big patron of wargaming across the defense community here in the United States, and when he moved on, there was no one to pick up the torch. So now it’s in recess. And a lot of institutions that would normally fund this are kind of in reverse. Someone else will come up, though, and then suddenly strategic thought will come back into the fore and people will need games again.

Rwizi Rweizooba:

It’s so counterproductive to keep going through this building and destruction of knowledge, awareness, and practice. And that’s one of the things that I talk about in my newsletter, the need for the cultural shift for wargaming to go really mainstream. People should be going on wargaming dates, people should be watching wargaming TV shows. That’s how proliferated it should be.

Danny Crichton:

With the rise of artificial intelligence, we’ve had a lot of engagement recently with universities thinking about the future of in-classroom education. So in a world in which you have LLM — you can basically do a lecture through ChatGPT; you can learn; you can write — what’s the purpose of a classroom experience?

We’ve talked to a couple of different universities, but people are increasingly seeing gaming as an experiential part of learning that could really transform the classroom from this sort of dry rote lecture, an education model that frankly dates back to the Greeks, to one where you can say, “Look, imagine a 20-person wargaming scenario played live in a classroom over three to four hours that helps you understand the complexities of logistics in the 21st century.” And now you’re getting beyond just reading some books. You’re getting into the actual dynamics in the real world.

Rwizi Rweizooba:

I’ve noticed some people even doing it for high school students. Clint Warren-Davey teaches at a Catholic high school in Australia. I’ve seen him teach wargame design to his high school students. I can’t tell you how jealous I am!

But like you said, with AI, now more than ever, we need analytical thinkers. And I don’t think we’re going to get those without games.

Danny Crichton:

I’m curious: One of the things you’ve been doing with Wargaming Weekly is reviewing a lot of games. I want to hear about patterns you’re seeing.

Rwizi Rweizooba:

I’ve mainly played micro wargames, so that’s what I can confidently talk about. These are really a game changer because they compress the entry point for wargaming. They make it easier for someone who’s new to it to experiment. A lot of them are solitaire games, and because they’re designed to be completed within 20 to 30 minutes, it doesn’t feel like a huge commitment compared to the typical game.

I think they’re going to be great for getting more people into the pipeline. Sebastian Bae has about three collections of micro games. Fight Club has a collection. Even the Georgetown University Wargaming Society has a collection of its own on its website. The Women’s Wargaming Network has a micro game design that they’re planning to release later this month.

In the militaries, one of the big trends is an effort to get wargames down into the lower echelons, because traditionally wargaming has been for generals and up. So there’s now this drive to get wargaming down to the platoon level, to squad level. And I think micro wargames are going to be perfect for this. They can fit in soldiers’ pockets and can be played in very compact spaces in very short spans of time.

Danny Crichton:

We’ve only built interactive games, mostly because for us, Riskgaming is an opportunity to bring people together. I do think that if you want to learn strategy, it might be best to have opponents because human opponents help you sharpen your thinking.

But at the same time, if you’re learning something new for the first time, there’s absolutely no reason you need to start with an interactive game. A solitaire game that allows you to get into it at whatever speed you’re comfortable with is a unique experience. It is also true that the biggest challenge for us is always convening that group of people together. If people were actually willing to play and do it in a solitary fashion, then you potentially have the ability to enhance the thinking of millions of people.

Rwizi Rweizooba:

On the business side, I also think there’s an opportunity for business micro wargames to make a splash. You can say, “Okay, you don’t have a lot of time, but at the beginning of this one-hour meeting, can we play a quick 20-minute micro wargame about our supply chains?” And then that can guide the conversation at the meeting.

Another design trend I have noticed is commercial PME wargames, so professional military education wargames. Traditionally, it’s the military that gets commercial wargames and tries to use them for their officer training. But now that going in reverse, whereby games originally designed for professional military education are finding commercial success. The biggest example of this, of course, is Sebastian Bae’s Littoral Commander: Indo-Pacific.

So that trend is also very interesting to watch because it’s really bridging the two worlds, the commercial world and the professional world. Now there are games that are both fun and can deliver professional military education.

Danny Crichton:

Yes, I said it earlier: I think there’s been an artificial distinction here, and the idea that something is better by being hard to play is actually just a sign of a bad product. You can have tremendous intellectual depth with a game that is very easy to onboard. And honestly, it’s publications like yours that are popularizing gaming and pointing out that there’s a whole spectrum. You can play games solitary, as groups, or otherwise. You can play them in the military, outside of it. They’re applicable to all these domains.

And that leads to my last question: What’s next for you?

Rwizi Rweizooba:

Like I told you earlier, for me, Wargaming Weekly is a continuous series of fighting my imposter syndrome. To feel like a real gamer, next, I want to work in wargaming. I want to do some kind of consulting work in wargaming, whether it’s business wargaming or military wargaming. But even if that doesn’t pan out, as long as I am playing games every week, as long as I am watching wargaming webinars every week and writing about wargaming every week, I’ll be happy.

Don’t miss

Those questions and more are at the center of Joel Burke’s book, Rebooting a Nation: The Incredible Rise of Estonia, E-Government and the Startup Revolution. The nation has outperformed across the board, and Joel takes a full look at the unique institutions and cultures that led to such success.

Joel, Laurence, and I talk about the early years of Estonia’s existence, why eGovernment can actually be even more private than our existing data systems in the United States, and finally, why Estonia’s government has so deeply embraced the private sector.

This conversation has been edited for length and clarity. For the full conversation, please check out our podcast.

Danny Crichton:

Joel, not only are you the author of a book called Rebooting a Nation, you were also the rebooter of sorts of a nation. So let’s dive in. We’re talking about Estonia, a country of a million people on the Baltic Sea. It generally comes into the news in the context of U.S.-Russia-Ukraine relations as this beacon of hope. It is a democracy in a very tough neighborhood. So give us a little context. Why Estonia? Why were you there?

Joel Burke:

It’s a great question, but a little bit long to answer. So I’ll try to give the Cliff Notes version. I was early at a YC / Andreessen–backed startup, and I had always wanted to live and work in Europe. I ended up moving to Berlin to run a company for Rocket Internet, which is a venture builder/private equity.

Once there, I quickly decided German entrepreneurship was very different from Silicon Valley entrepreneurship, and I wanted to do something more impactful and mission driven. I came across a job ad for a startup within a government, and I never heard of such a thing. The job was working in Estonia for the e-Residency project, which is a flagship digital initiative.

They ended up hiring me, I assume, because I was the only person foolish enough to move to Estonia in mid-February when it’s dark for something like 22 hours a day. But I had a front-row seat to the country’s change ... Well, I shouldn’t say it changed. By the time I got there, it was already changed. Their digital government was already set up and going, but I got there at the time when the rest of the world was learning about it. And I was just astounded by what I saw and thought that there were things that we could learn.

Fast forward a couple of years, I returned to the United States and did this program called TechCongress, which takes tech people and puts them in Congress. And I was frankly shocked by the lack of tech policy knowledge and the lack of digital state capacity. So I decided to write this book about Estonia with the hopes I could take some of Estonia’s tangible lessons and share them in the United States.

Danny Crichton:

So Estonia was an independent nation, kind of survives the Bolshevik Revolution, does not survive Stalin, gets annexed, regains its independence in 1991. The title of your new book is Rebooting a Nation, but the nation was already rebooted in ‘91. So this is a 3.0 situation. Why the sudden change? Were the fresh, new institutions just not cutting it?

Joel Burke:

It was a perfect storm.

For one, after five decades of Soviet occupation, you can imagine people were ready for something radically different. So you had folks like Prime Minister Mart Laar, who was one of the earliest PMs. He was very much in the Thatcherite vein and was ready to embrace the opposite of anything the communists had done or how they thought about economic policy.

He basically put the country through economic shock therapy. I mean, removing subsidies, helping foreign investors take over ailing state-owned businesses, agricultural reform to remove collectivization.

And then it happens that independence, in 1991, is coming at the dawn of the modern web. So former President Toomas Hendrik Ilves, is Estonian American, educated at Columbia University and grew up in New Jersey. In his education, he had to learn programming. Very early in the country’s independence, he encountered Mosaic, the browser, and just identified it as a place where Estonia could be competitive. There was this new technology, a new frontier with a more level playing field.

And then the last thing I’d mention is the mentality. As you can imagine, there’s the Lara-esque approach, there’s rejection of communist ideals, but looking at Estonia on a map, it’s quite close to Finland. And Finland was always held up as this benchmark for the country. Estonia considered itself more culturally Nordic than Baltic.

Throughout occupation, because Tallinn was relatively close to Helsinki, they could catch radio waves and rebroadcast American TV shows. So Estonia had these repeated glimpses of what life was like in the West. Once they finally gained re-independence, there’s this feeling of, “Wow, there’s such a huge gap here that we needed to jump over and we’re not going to be able to do it by following the traditional development playbook.” So they were willing to do pretty bold and big things. They had a massive chip on their shoulder, and they were willing to take risks to try to get there.

Danny Crichton:

Josh Wolfe often says that chips on shoulders put chips in pockets. That clearly worked for Estonia.

One thing you show in your book is that this wasn’t just one decision. There wasn’t just one moment where they say, “Oh, we’re going to modernize,” and it happens with this particular law. You have everything from the Tiger Leap Initiative and then you are trying to get ID card rollouts more than a decade later. And then you have to deal with the Russian cyberattacks. I mean, this is a long journey. I’m wondering if you could talk about how, at each one of those moments, the country was able to keep going and double down.

Joel Burke:

This is a really important point. There was sustained political will across different administrations and varying political parties. They decided this is the direction that we’re going to go, and they did. I do think this is critical.

First, there’s the Tiger Leap campaign, which, for context, is basically the idea of connecting all the schools in the country to the web, and then there was an additional educational campaign that went along with it. So one, you give kids access to the web so they can learn digital skills.

And then two, you also provide supplemental digital skills while at the same time, you’re training the teachers. All this helps filter out digital skills across the rest of society. That was critically important for basically building a workforce that was ready for the future.

Just a couple weeks ago, Estonia signed a deal with OpenAI to roll out OpenAI services across its schools again. So they’re doing this repeatedly for new technology. So there has been basically a long consensus to continue to invest in digital in recent years. But I think that was built up through elite-driven pushes to get this done. I mean, in retrospect, the Tiger Leap campaign sounds obvious. Of course, we should connect school to the web and teach kids digital literacy, but at the time, you’re in the late ‘90s and you might be taking that money away from paying teachers’ pensions. I mean, the country was economically pretty desperate until semi-recently.

Certainly, there was fighting. I do think part of the reason Estonia has seemed so stable and so willing to push forward is because many of the fights happen in local media in the Estonian language, which many outsiders don’t really see or pay attention to. That’s one of the perks of being a small country. I mean, every single thing that the United States does is litigated and talked about across the world.

And so I think there were many more debates than were on the surface, but indeed it has been quite something to see, first, elite consensus and then general population consensus to invest in digitalization.

Danny Crichton:

From Estonia’s own history under Soviet occupation, there’s a lot of concern around privacy, individual protections. You had secret police, you had secret intelligence, informants all across the world. Why did it go in such a different direction where people are willing to say, “Look, I trust my government to have a digital ID. I can use it for the bus.” Clearly, this data could be used against me, so to speak.

Joel Burke:

It’s a fascinating question.

When I was working for the e-Residency program in Estonia, I was the guy they’d roll out to talk to American and German delegations. And indeed, that was almost always one question like, “Well, Estonia’s interesting, but we have different cultural values and think differently about privacy.”

But I guess I would reframe the question and push back a little bit. So let’s take me as an example. The government has my information at the IRS, it probably has some healthcare information because I worked in Congress, I’m sure it has some security information and background checks information on me. I don’t actually know what all the government has. It’s probably spread out on paper files across a bunch of different agencies. I don’t know when someone goes and accesses it. So, yeah, maybe we have security through obscurity, but I actually think any motivated actor within the U.S. government could get that data if they wanted to, and I would have no idea whether they did or what information they had access to. Not to mention the fact that they can just go on private markets and go to a data broker and probably buy it anyway.

In Estonia, they built up basically a legal system that puts the citizen or the resident in the driver’s seat. So when I go to my dashboard on the e-Estonia site and look at all the digital services, I can see in a unified view the information that the government has, so I know what information they have. I can also see who has viewed my data and why.

How they got there, I think, was in part because of the push for independence. I think there was initially so much trust in the independence leaders who helped shepherd the country out of Soviet occupation and into re-independence that there was a willingness from the people to trust them. But they also built in these safeguards legislatively and on the tech side.

Even in the United States, I think that’s something that we can learn from. I don’t think we can and should adopt Estonia solutions wholesale. I think it much more likely, if we ever do something like digital identity, we’ll partner with Apple and Google because they’ve already got biometric scans of my face. I trust Apple to verify my identity far more than I do the DMV, right?

And so I think we’ll do it a different way, but this mentality of doing these hard things, having sustained investment, and building in these safeguards is really important.

Laurence Pevsner:

So when you went to Estonia you knew in advance that it would be a technologically advanced country. What still surprised you anyway when you arrived? And how did the country continue to change while you were there?

Joel Burke:

One thing was just how far the government went to try and make itself accessible and change proactively. So they have this program, Accelerate Estonia. The tagline is, like, make illegal things legal. And the government actively works to support the private sector and create innovations.

Even the e-Residency program I work for was initially pitched at an ideathon — a hackathon style thing. And then what happened is that everyone got enthusiastic about it. The parliament basically voted to create the legal changes that would allow the program to exist, and then the program was able to get going in the matter of a couple of months. I’m sitting at Stanford in California, I can’t imagine Governor Newsom saying, “Yeah, we’re going to make illegal things legal. We’re going to proactively work with the private sector to make it easier for you to build houses, let alone defense or something.” This wholesale mindset shift of trying to work proactively with the private sector is really important.

Laurence Pevsner:

You write in your introduction about how America is facing what you call our Macy’s moment. I was wondering if you could say a little bit about that.

Joel Burke:

I think we’re at the dawn of another technological revolution and a change in the way the world operates. When I think about the United States and how our government was architected, it was architected for a pre-digital age. Maybe we’ve digitized some things — you fax things or you email forms instead of mailing them. But by and large, the way our systems actually operate are still made for the industrial era before digital.

So the Macy’s moment: It was at the start of eCommerce, and there was this question about whether big box stores still had a role. Macy’s is still around, but it’s not the heyday of the Macy’s parade and being the iconic American brand.

What I want people to take away from Estonia is the mindset, the culture that we need to build together. If we keep having this situation where, every four years, someone comes and kicks down what the last guy built, we’re never really going to build the systems and the kind of society and efficient government that we want.

And so we need to come together at a consensus point for society and think of how we want to face the next generation. And I think it’s going to be AI. We need to have a conscious strategy about how we’re going to adopt that. And I am very excited. I think AI and digitalization more broadly open up huge avenues.

Danny Crichton:

On this front, let me ask you: We recently had an episode with Kelvin Yu, who is a fellow at the Foundation for American Innovation and published a bunch of proposals to improve industrial policy — less governance and digital IDs and all the stuff that we’re talking about. But one of the key points he made was a focus on details, a focus on small things. Instead of trying to do a big rebooting of a nation, the idea is to make tiny patches to the code base all over the place. If you do hundreds of those over a period of time, you aggregate to something that is much larger.

Joel Burke:

Kelvin’s great. I have the playbook open in a different tab, and I haven’t actually dug into it yet, so I’m not going to get over my skis here. But I think both are true in a way. This is a bit of a cop out.

When I was working in Congress, there was a talk of secret Congress, which sounds much more nefarious and cool than it is, but it’s basically this idea of doing the things that aren’t sexy, that don’t appear on primetime television. And that is where things are most tractable, because if you can get a small bill done and it doesn’t have to get litigated on CNN and Fox and MSNBC and become highly politicized, it’s much easier. There is something true here.

However, I guess where I feel we’re at as a society is we need to find consensus again. Maybe I’m romanticizing the days of the Cold War, but I think about our foreign policy, for instance, and it feels like we were very aligned over several decades.

The wholesale challenge from technology is going to be so important. It is going to affect society on such a massive scale. No doubt we need to fix the potholes in the road, too, and we need to make these thousand small changes, but we also just need to get our stuff together as a society and figure out what the pathway is for America.

Don’t miss

In 2022, he joined me and our own Peter Hébert to talk about the time he walked out of a position just two weeks on the job, how the serendipity of the Valley led to Apple, how to build human connections, why he thinks many innovations in the tech world today are a waste of time, and the importance of storytelling and product marketing when building a product.

His interview stands as a master class to aspiring entrepreneurs and investors alike.

This conversation has been edited for length and clarity. For the full discussion, please visit the Riskgaming podcast.

Danny Crichton:

I read your book Build, and one thing that struck me is the panorama of your experiences. You’re most associated with Apple and then with Nest, but you’ve actually been attached to a number of other startups in Silicon Valley. I wanted to start with a very short experience you had back at the height of the dot com bubble, RealNetworks, a company that you joined for literally, I guess, two weeks.

Tony Fadell:

So RealNetworks, how did I get there? Well, I was at Philips at the time — Philips Electronics from the Netherlands. I started a team there. I started a whole business group called The Phillips Mobile Computing group. We were doing handhelds for mobile-based computing.

I was learning all about digital media because we were the first devices to have Audible back in the day. I was thinking all the time, what if this was music? I was a DJ, and I was sick and tired of carrying all these heavy CDs — you know, a thousand CDs — to each gig. I thought, “wow, if these devices are doing audiobooks, one day, they could be doing music.”

So I switched to Philips’ strategy and ventures arm, and it was the dot-com craziness. I was watching all of these companies getting started and funded, and I didn’t know anything about that world because I was always deep inside Philips or General Magic before.

I decided I wanted to go learn about it because I saw all my friends doing it. So I started learning about venture, and at that time we were investing in companies like TiVo and Audible.

I’m watching this digital transition happen, and I come up with this idea: Phillips makes all this hardware and RealNetworks was making this software called RealPlayer. And what RealPlayer was all about was getting audio. So it was audio streaming but they also had music files — MP3s were coming out — and so you could listen to music.

I was like, “oh my God, this is perfect.” Phillips could make the hardware and Real could supply the software. Let’s put these two things together and create a whole line of consumer electronics for music. So I got the CEO of Phillips together with the CEO of RealNetworks.

What happened was the CEO of Phillips was about three hours late to the meeting. So I sat there with Rob Glaser for three hours, telling about the vision I had for this thing. He was getting all anxious and excited.

So the CEO came and they had a little thing, but it was very short. Literally 24 hours later, I get an email from him saying that he wants me to join RealNetworks and build what I want to build. They’d let me build a team in Silicon Valley, build the hardware and put Real software on it. And we’d license the devices out to people.

I thought it sounded really cool, so I joined up. When I got there, the CEO goes “Well, we have a change of heart… Well not change of heart. We just changed our strategy. You’re going to need to move to Seattle now.” And I’m like, what? What do you mean? We had already agreed on Silicon Valley.

He’s like, well, I changed my mind.

Peter Hébert:

This is day one.

Tony Fadell:

Yeah, day one or two. Strategy change! And so I go up to Seattle to look at it, and I’m sitting there. I just left my whole community that I was building in in Silicon Valley from General Magic. So over those two weeks, I was like this is not going to work for me. I’m done.

From there, I created my own company called Fuse to then take some of those ideas and change them up a bit.

At the time, everyone was all about home theaters. But home theaters — you had to assemble them with different brands of speakers and TV and home audio and video components to get this digital experience. So I was like, okay, we’ll make a one-box product. It was a software-driven box where you could put CDs in, it would rip them to MP3 files, and you could have this digital jukebox of all your music.

And then, of course, the internet crunch happened, and people didn’t want to fund hardware companies.

Danny Crichton:

So you joined the fruit company.

Tony Fadell:

Right. I was trying to raise capital for Fuse, and this was over the latter half of 2000. I went on 80 pitches. I got 80 “nos,” and it was really, very difficult.

I took a week off in the beginning of January. I went to go find myself, I guess. I went to lunch with a guy. His name was Ali Alasti. We worked together at General Magic, and I told him about the disaster that was going on. The next day, he had lunch with a guy from Apple who he had worked with at Next. His name is John Rubenstein. And what John said was “Who do you know who does digital devices and might know some things?”

And Ali said he should talk to me. I got a call days later, and that’s the beginning of my contracting assignment with Apple, which was to create the Apple version of an MP3 player.

And so over six weeks, I envision the product, put together the different options, the bill of materials, what the architecture would look like… all these things. I presented it to Steve Jobs in March of 2001.

Danny Crichton:

There’s a number of really interesting storylines throughout your book. One thing that’s almost implicit is the serendipity of Silicon Valley.

You don’t know if it’s today that something’s going to hit or it’s going to be months or years later. But at some point, all that serendipity will come together and you can actually do something amazing.

Tony Fadell:

There’s a chapter in the book called “Death of a Salesman,” but it’s not about transactions. It’s about relationships. And the serendipity of Silicon Valley – that’s how it happened going to Apple. It was about keeping up a relationship. You would work together with somebody and you would build a relationship. It wasn’t about the company you’re at, or the deal you’re trying to get done. You met on a human level.

We still had relationships even as the companies changed or the technology changed. Maybe in some cases, we were competitors. But that spirit kept going and that relationship kept going. If you invested in it, other people would invest as well.

That’s a really key piece of what Silicon Valley was and is. And now, those same kinds of networks are growing up all around the world. You don’t know if it’s today that something’s going to hit or it’s going to be months or years later. But at some point, all that serendipity will come together and you can actually do something amazing.

Danny Crichton:

You talk in the book about the importance, even at a large company, of not just building relationships internally, but to get out of those walls. It’s very easy to basically only know folks inside the building and never have the opportunity to leave.

Tony Fadell:

You have to get outside. You have to understand what’s going on in the markets. You have to understand what’s going on with your customers or consumers.

You do have to build the relationships inside the organization, but you have to maintain those relationships separate from the companies you are at, because you never know what’s going to happen. Maybe the companies can be partnered together at some point, or maybe they might recruit you, you might recruit them.

Peter Hébert:

You’re very overt about your appreciation of the old-school approach: keeping it gritty, working incredibly hard.

Having been in many ways a product of the 2000 crash, what do you think about the current environment? Do you think some what we’ve seen in the last decade — where these companies have become large and bloated — will change and we will get those grittier, more focused entrepreneurs?

Tony Fadell:

It’s the best time to invest. It’s the best time to get great teams when the markets are down. Those people who know how to do a lot with little, those are the people you want on your teams. Not the ones who have to be surrounded by more and more teams and people.

You’re going to see those cultures that got fat, dumb and lazy — even if they didn’t have revenue, but because they had too much free capital coming. You’re going to see massive layoffs. You’re going to see the culture change. All the employees who are like, “Oh, you mean I have to work now?” they are in trouble.

It’s better to always stay tight, stay gritty. Because it eventually turns. It always turns. This is a sinusoid. It isn’t a straight line. So if you keep that grittiness, when the downturn comes, you don’t have to worry about changing your culture or mindset. Just remember to stay grounded, because you can get through those tough times. You know, all those perks and everything? They will go away, eventually. They have to go away because that’s not reality.

Peter Hébert:

Would you have been the same entrepreneur — have achieved the same level of success — had you not had some of those early failures?

Tony Fadell:

There’s a chapter called “Fuck Massages” in the book. You know where I learned that? I learned it from my grandfather, who grew up in the Depression. He would save every nut and bolt. If a piece of equipment broke, he would tear it down to the component pieces and put them in little cigar boxes all around the garage. So if he ever needed it again, it would be there as a resource to him.

That’s how I was raised, with a Depression-era grandfather who said all of this stuff is precious. Use it and save it because you never know when you need it.

Danny Crichton:

That was one of the surprises for me in your book. I identify you with these iconic products — iPod, iPhone. I didn’t know that you quit Apple three times. That you actually left, came back, left, came back.

I’m curious about some of the stories there. When did you know that you needed to go? And when did you know that you needed to come back?

Tony Fadell:

So first, when I say I quit… I said “I quit,” but the process didn’t complete. So two of the three times, I said, “I quit,” I was walking out the door. In one case, I had my stuff loaded in a box and I was having a thank you party for the team that night. And then, in the subsequent four hours, it got fixed. And I told the team, “Tony staying party!”

So that time, it was really about necessary conditions that weren’t satisfied when I first joined the company. That was the first time. The second time was really about a fundamental decision that had no reason for happening except it was political. And I said, this is bullshit. The third time was really personal. It was a real personal thing about our family.

So each of those times was for real reasons, and it came after I tried all kinds of positive and constructive ways of trying to fix the problem. But sooner or later, if you really believe it and you want that to change, you have to walk out the door. People aren’t going to protect you. They’re not defending you. You have to defend yourself.

And when you do, then you really know what your value is.

Peter Hébert:

You are now a successful investor. When you look at entrepreneurs, when you are meeting with a founder and are captivated, what is it? And as you’re evaluating opportunities to invest in, are you giving points for extra difficulty? Does that create greater affinity for you?

Tony Fadell:

Absolutely, but we’re not investors. We’re mentors with money. The reason we do what we do is because I tried to be retired and it really sucked. I thought, “Well, can I work with teams that are doing really hard things?” I want to work with them and I’m also curious about them. I think they’re doing something that’s going to fix the planet, fix society, fix health.

That’s what we really go for. We’re looking for disruptive deep technology. We care about when you’re doing something hard, something really challenging. You need to have a great set of people around you. That’s why, writing the book, I had mentors around me.

We want to try to help those companies we go in, because they are trying to do something that is so against the grain, something that most investors won’t invest in till it becomes the fashion. We do it with our own time and money, and we don’t care about anybody else’s things because we believe in their trajectory.

Danny Crichton:

You have this quote in the book that if you’re not solving a real problem, you can’t start a revolution. And one of the things that is interesting across both your own work as well as the companies you’ve been investing in, is a lot of them solve real problems. That has to be intentional.

Tony Fadell:

Look, if you’re not solving for pain then, what the hell are you doing?

General Magic wasn’t solving for pain. We were trying to impress the geek next to us. We were making the iPhone 15 years too soon. The technology wasn’t ready. The market wasn’t ready. Society didn’t have the problems we were trying to solve, which were like mobile email. Most people didn’t even have email. E-ticketing for travel. Why? This was before Wi-Fi, before data networks and mobile phones.

You could say, well, General Magic had to exist for the iPhone to exist. Sure. But at the same time, we could have used a lot of that talent to solve other problems along the route to get to the iPhone in a much smarter way. I look at things like the metaverse. I’m saying what pain are you solving? I’m all for AR and VR and XR technology, but only in service of a pain point to create a new tool.

There’s so many things we can fix where there are problems, why the hell are we putting all this capital to work for things that aren’t problems?

Danny Crichton:

One other question I have is around the profession of product and design. Obviously, 30 years ago, there were not a lot of product designers out there. Now, it is a huge profession.

I’m curious when you look forward, what’s going to change for the folks who are working in product, in design, in the coming years.

Tony Fadell:

I think we’re finding more and more that learning by doing is it. That’s the way you do it. And so, you’re going to see more, apprenticing and more internships — getting in and learning with the teams in the environment that they’re in.

You know, the best school that we ever hired from had one semester in school and the other semester in a workplace, and that was Waterloo. That’s what we need — a semipermeable membrane.

More and more universities are starting to focus on learning by doing, or more project-based learning. The more you can share with the corporate world or the startup world, the better when they’re younger. Because that’s how I learned. I learned by building my own companies and going back and forth between work and school. I had my own startup in high school, had three startups in college, and I would go back and forth from school to my own job.

I think that’s a much more powerful workforce and a much powerful way of educating people.

Peter Hébert:

I will give you the last question. There’s a volume of guidance in your book. But if there’s one thing that some aspiring entrepreneur, the Tony Fadell of today, should take away, what is it?

Tony Fadell:

Learn how to tell a non-fictional story. Tell a story that matters; that is solving pain and giving a superpower to somebody. Make sure it’s an incredible story, and that you can actually deliver it. It’s not Theranos. Have a great story and a way to show that you can implement it, and it’s meaningful. It’s a large enough market, all this stuff.

But most people talk about the what, what, what? Not the why, why, why? Why is the most important from an entrepreneurial perspective.

Don’t miss

We don’t have a guest this week on the podcast, so instead, Laurence Pevsner and I recorded a special outdoor episode from Carmel Valley during our Lux Leaders Retreat to reflect on our recent Riskgaming launches and runthroughs and what happens next.

We discuss the emergent gameplay in Gray Matter that surprised us — from spontaneous character-driven debates to unexpected trading outcomes — and we explore the game’s core mechanics around information markets, attention economies, and strategic publication decisions, examining how players navigate the tension between keeping information proprietary versus making it public for advantage.

We then shift the conversation to Southwest Silicon, our full-immersion simulation focused on Arizona’s semiconductor industry and water security challenges. We discuss how this immersive format differs from our larger salon experiences, highlighting the intricacies of water rights in the American Southwest and how players grappled with resource allocation decisions that mirror real-world policy dilemmas.

Looking ahead, we preview upcoming games exploring urban automation, critical mineral supply chains, university research funding, civil infrastructure security, and the future of international development institutions. Finally, we also discuss our philosophy of using game design to model emerging technologies and reimagine institutional frameworks for an AI-driven future.

This conversation has been edited for length and clarity. For a full version of the conversation, please subscribe to our podcast.

Danny Crichton:

Laurence, we just launched your new game, Gray Matter, at two locations — in New York and here in Carmel for our Lux Leaders Retreat. How did it go?

Laurence Pevsner:

It was a smashing success, I think, but I’m a little biased.

There’s nothing quite like having all of these ideas in your head. You’re designing the game. You’re thinking about how players are going to interact, what they’re going to do, and then what they actually do is always different. It’s always incredible.

That’s the beauty of these games — the emergent quality. People do things you don’t expect. In New York, for example, at one point one of the players took a champagne glass and rapped their spoon against it as if they were making an announcement. The whole crowd hushed because they thought it was one of us making an announcement about the game.

And instead this player went on a rant about how all of the NeuroLifts [the three drugs at the center of the game] were terrible — it was all part of their character. And then all of a sudden other people in the crowd started shouting back with their opinions based on their characters.

And it really felt like, okay, this is happening. These people are really embodying these roles. And we didn’t tell the players to do anything like that.

Danny Crichton:

Let’s take a step back here and say how this game works.

Laurence Pevsner:

So Gray Matter is about the intersection of informational markets, attentional markets and money markets. Who has information? Who’s putting that information out there and making it go viral? And who’s using that information to make money?

One of the main aspects of this game is that people can take private information and make it public to everybody — all the other players, the FDA (which is monitoring decisions), and fact-checkers (who are looking at the truth environment of the game). There are some players who are heavily incentivized to publish private information. And some players are heavily incentivized to keep information private.

Logically, it’s the attention agents, the people who are podcasters or Substackers, who want to put out the information they get. That’s their whole stock-in-trade. And then you’ve got the people who have money. Sometimes people who have money want to put information out there; think about the investor who wants to pump their product and make sure that their company is getting the eyeballs they need. But sometimes if you’re an investor, you want to keep information private.

Danny Crichton:

Right, so we have information markets, we have attention markets, and we have money markets. You also have a mechanism, and I don’t know if this happened in either of the games, where if you can keep one of your clues completely unpublished so that it doesn’t become available to the public — it’s basically information held only by you. There’s a small bonus for doing this with one of your clues and a massive shoot-the-moon bonus if you can get both. Did anyone actually succeed?

Laurence Pevsner:

Yes. Not only that, but the two players who won here in Carmel Valley both succeeded on that front.

Danny Crichton:

Amazing.

Laurence Pevsner:

Right? And actually it was amazing to watch. I was watching one of the players, our own Michelle Fang. There was a player who was going to publish information, and she literally dove in front of the public board and was like, “Wait! I need that not to be published. What can we do to make that happen?”

She gave away quite a lot from what I could tell. It was a little bit of carrot, a little bit of stick, and eventually they got a deal done, and she was able to maintain that high quality, highly credible clue that didn’t get published, and it ultimately allowed her to be tied for first at the end of the game.

Danny Crichton:

Unbelievable. This game is a lot of fun.

Beyond Gray Matter, we also just launched Southwest Silicon with our designer, Ian Curtiss. We did an episode with him about the game two weeks ago, and he just subbed in for me as a guest host in last week’s episode.

Let’s talk about Southwest Silicon, because one of the things we’re really just now learning as Riskgaming entrepreneurs is the model of having two types of games. We have full-immersion simulations like Southwest Silicon, which is played by smaller groups of people, six to eight or nine, multi-hour. We really go in-depth on a specific topic. And then your game Gray Matter, and two others — including DeepFaked and DeepSixed, which is on our website — are these sort of salon experiences. They are often designed with partners who are hosting a conference and have an hour or so when they want to do something more interesting than sitting at cocktail tables regaling about their LinkedIn.

Laurence Pevsner:

If you think about a game like Gray Matter, we are covering a lot of different subjects. We’re looking at all kinds of things in the field of biotech, nootropics, which can be seen as a metaphor for AI. We’re thinking about Ritalin and psychedelics and all these crazy topics. And it’s a big broad-based view. I had people come up to me to talk about the lessons they learned.

But it’s fundamentally different from a game like Southwest Silicon, through which you go really deep on one particular topic. And in this case, it is a topic most people don’t really understand. Of course, everyone’s talking about chips these days. When we played the game in London, a lot of those folks were very familiar with the topic. They knew about the importance of chips to the future of tech. But almost none of them knew about the water security element of this. Every chip fab has to deal with water security, it’s just that the problem is particularly acute and complicated in Arizona, where the game is based.

I was struck in that game by a couple of things. One, by how much the background really helped the players. So we gave a really in-depth pre-brief so that, when the players were negotiating, they had a really deep sense of the policy landscape. I think that helped them play the game appropriately.

The second thing I was surprised by, and I’m curious if you expected this as well, is that the game was a slow burn. Trades were being made, deals were happening, people were making moves, but it wasn’t frenetic with energy until the very end. And then all of a sudden there was this amazing auction that took place, where the person deciding where the factory was going to go was literally wheeling and dealing between the various cities, and other players were tripping over themselves to get the deal and then seeing what they could tolerate.

To me, that was demonstrative of how these things often work in real life, where there’s a lot of quiet conversations that happen behind the scenes. Maybe it’s boring paperwork. People are making their moves, but then it comes to a head. There’s a moment when it explodes onto the scene, and the big conflict happens, and the news reports on that. But to get to that moment, there was a lot of setup that had to take place. I’m curious if you feel like you’ve seen that in other Riskgaming scenarios we’ve run before?

Danny Crichton:

I think it depends on the runthrough and the group of people. Some people don’t want to make decisions upfront. We played a beta test of this game in Tokyo, and that was a slow burn. In that case, it was a very cultural dynamic, where people were just trying to learn the rules and the experience. At some point, toward the end of the game, it’s like, no, no, it’s time to make decisions!

In other cases, groups don’t compete or they’re trying to figure out whether they should compete or cooperate, and exactly when to do so, and so there’s just a miscalibration. In some other cases, we have business execs or folks who are in much more intense industries and are comfortable making decisions instantaneously. Often they make the wrong decision, and sometimes they’re basically irrevocably going to lose after the first round because they’ve made a terrible mistake early on. And so it really just depends.

To me, the thing that’s interesting about our contribution to the debate is that everyone’s focused on data centers, water, security, and how much water is being used. And there is a big debate on how much this matters versus a golf course, which is a huge issue. Or an Arizona alfalfa farm.

What they’re not focused on, though, is the complexity around water rights in the American Southwest and West in general, where over the last 150 years, in order to get water to a place, you either had to channel it, dam it, go through an irrigation district, et cetera. And so there’s just thousands of these water districts. And in order to get water to your chip fab in the center of Phoenix, you have to cobble together a path through all these individual water rights from Point A to Point B.

It’s a little bit like trying to cobble together property to build a building. You don’t want people to know you’re doing it. So there’s all this skullduggery going on. You have all of these corps with anonymous owners. They’re trading these rights. It’s very, very opaque. And yet water is the elixir of life. It’s the most important resource, particularly in the desert, particularly in the American Southwest, which is going through a drought with the Colorado River at historic lows.

We try to offer a little bit of a different perspective of saying, we should really focus on what these rights are, how they work. When we played in London, they didn’t understand any of this. It just doesn’t work that way, obviously, in London. So on top of the value of understanding the chip fabs and water and security and economic growth, we also have this another layer of imparting more understanding of the American political system.

Laurence Pevsner:

One of the important players in the game is a Native American tribe. In real life, tribes own a good amount of water rights. And I remember the players in London asked, “Can you give us a metaphor for a UK equivalent?” It’s hard to do.

Danny Crichton:

And look, these issues are going on in real time. The Gila River Indian Community, which is in central Arizona, was named for a river that has water, but that water was stolen by the federal government a century ago. Recent Supreme Court cases have basically returned access to that water to the Indian nation. And so all of a sudden, you had this bounty of a resource everyone needs. You want to protect it for yourself and legacy and heritage, but you also have this economic resource that’s valuable.

In the last couple of years, we’ve seen this massive shift around those rights led by Neil Gorsuch and others on the Supreme Court who have made that an issue for the first time in almost 100 years. And so these water rights are not just a couple sheets of paper in the game. They’re literally being traded at the highest levels of American jurisprudence, and they can massively affect communities throughout Arizona and the American Southwest.

Laurence Pevsner:

One other thing I noticed about the game as we were doing what’s called “the teach” — when we try to explain to players how to play the game — one of the more complicated aspects of the game is the water rights. There are different kinds of water, and there are different levels of rights you can have. The players are like, “Wow, this is complicated.” And what’s funny about that is we have already massively oversimplified the different ways water rights work to make the game function. And yet, that was still the most complicated part of the game.

Danny Crichton:

Yes, yes. So that game is online. You can download it. It’s the fourth one we’ve published. We have a total of seven games that have been designed. We have three others that, someday, you and I will get around to actually publishing on the internet. So what’s next? What are we doing?

Laurence Pevsner:

I am really excited about the games we have coming up. We have a number of new games and new partners we’re going to be working with to develop those games and to put them out. I’m particularly excited about a game exploring automation in cities. Players will actually get to play different kinds of cities and different places. So think resort towns versus college towns versus manufacturing towns. Everyone’s thinking about AI. And of course, we were here at the Lux Leaders Retreat, and all of the CEOs who work in AI are like, “Yeah, there’s a really good point about potential job loss here.”

Some are obviously arguing that AI is going to be a boon for jobs. Some are saying, “Yeah, if I’m being honest, it’s going to be really tough, especially for white collar workers and especially for workers coming out of college who aren’t already at the top of their field.” So that game is going to be particularly applicable. It’s only going to become more and more relevant as time goes on.

What games are you excited about?

Danny Crichton:

Ian is working on his next game, which is focused on critical minerals. We finally gave in to the constant demand for a critical mineral supply chain game! Although I think we have quite a twist in it that will make it way more fun than people expect. So I’m looking forward to that in the next couple of months. We’ll hopefully publish it early to mid next year.

And then we have a lot of conceptual stuff going on. We have stuff around Taiwan. We have stuff around space, specifically the balance between space commerce and space security. And we’re seeing more and more nation-state actors as well as just everyday hackers going into these systems.

These days, I’m particularly interested in civil security. So looking at civil infrastructure — power plants, sewage, water systems. These are increasingly being digitized. They’re on the web. In California, where we’re sitting now, PG&E has connected its grid to the internet to be able to monitor it, to be able to check for wildfires. But that means they are now vulnerable to hacking, and they need to be secured. And in many cases, there’s no security at all in these systems. It’s a huge topic we’ve even had senators mention.

And then the last is going back to sci-fi. We have a lot of science-fictional companies, companies in xenotransplantation, orbital mirrors for sunlight. So many interesting new businesses that are going to come in the next five to seven years and are going to transform humanity. I feel like we have a lot of present day security problems, and the bias is always what’s happening today. But we really want to look around the corner and say, “What happens if all these technologies come together?” A world in which photovoltaic farms are fed 24/7, 365. Does that transform the energy grid? What does that do? Suddenly solar doesn’t turn off at 8:00 PM.

We have these huge changes and shifts that are underway in the next couple of years from science and technology, and I’m really excited to try to model those in real games.

Laurence Pevsner:

I’ve been thinking along the same lines. I’ve been toying around with the idea of cutting-edge science research and how that happens at the university. I didn’t go to a big university. I went to a small liberal arts college. And so I found universities totally bizarre once I started looking into them and all the different functions.

How cutting-edge science gets developed there has become especially relevant. A lot of our companies are spun out of labs at these schools, and it’s become a big problem because all of their science research funding is getting cut at the federal level. We have, of course, the Lux Science Helpline, which you can still call in if you need help. We are here to help. Since we’re getting tons of outreach through that, it started to make me think like, well, what is it about the university that relies on federal funding in this way? And how would they do without it? How would that work?

Danny Crichton:

The last thing for me is thinking about the next generation of institutions like USAID and others that have been demolished. In many cases, they had good missions although there’s definitely a diversity of views on their effectiveness. Now that they have been dismantled, what does 2.0 look like? There’s a real opportunity from the Riskgaming perspective to simulate new types of institutions, whether it’s science or universities, whether it’s research funding, whether it’s USAID and global development, et cetera. There’s a whole new set of institutions that can come up.

Laurence Pevsner:

Yeah. And I think the exact same thing is true with the universities. It’s true with some of these big companies that are being disrupted. It’s true even within the game we have with the chip fabs. Okay, yes, we had the CHIPS Act. What’s next, and how do we actually make this work? That’s what we’re always trying to search for: how to build something better out of the rubble.

Don’t miss

That’s part of the work that Pijus Krūminas is focused on. He’s a professor at Lithuania’s ISM University of Management and Economics and is the head of its Wargaming Lab. His research converges the social sciences into modeling and simulations, tapping into fields like political economy, game theory, management and more to create a new synthesis.

Subbing for me this week is Ian Curtiss, our independent Riskgaming scenario designer, who recently launched Southwest Silicon as Lux’s latest offering.

Ian and Pijus talk about simulating the polycrisis, how academic research flows into wargaming design, how students are learning through experiences, why political economy is overlooked compared to traditional military wargames, and a variety of recommended games for newcomers to the field.

This interview has been edited for length and clarity. For the full episode, please subscribe to our podcast.

Ian Curtiss:

How did you land on wargaming? I think you’re at a really interesting place, career wise, to be so heavily involved with the wargaming community while also working at a school of business and economics.

Pijus Krūminas:

Yeah, indeed, I’ve been teaching at a private business school since 2012, and I started more on the political science side. And one of the thoughts I had is how to make the seminar activities more interesting. At the same time, I was teaching comparative politics through models. So, it’s a little bit of mathematics, and so on. And sometimes those models don’t necessarily come alive. They make sense once you solve them, you see how they work, but internalizing them is challenging.

About the same time, I got Phil Saban’s book Simulating War, where he discusses how he used wargaming to simulate battles. And that gave me the idea that maybe I could turn some of those models we discussed in class into a game we could play. So I would present the model, and how it looked mathematically. And then in the seminar, these groups of students would play with some adjustment of the same model.

And then a couple of years later, I talked with the head of the undergraduate program in political economy and I said maybe we could form a course focused specifically on wargaming as a methodology. Basically the same thing we did in class, but in the course they would become creators of the interactive models. Since then, we’ve been having this class every year with new students coming in.

Last September, we launched a wargaming lab at the university, also kind of a project activity that would give us a little bit more room to explore the potential of wargaming, for example, for research.

Ian Curtiss:

From your student’s perspective, how do you encourage them to model political economy as a game? Because when we — meaning the gaming community — talk about these difficult-to-measure concepts, people get so jittery about gamifying them, putting hard metrics on these things.

Pijus Krūminas:

One thing I do is basically I give them freedom in terms of the topic as long as it broadly fits within the political economy. So, they can go into historical examples, or they can model something more present. And I think giving them freedom to select their theme also makes them a bit more open. Not everybody would necessarily like to work on a kinetic war game, because of the topic, but then an election, or something like that, they would. Or we had a game on privatization processes in the 1990s in Lithuania. After the Soviet Union collapsed, the state had a lot of stuff it didn’t need. So, privatization takes place, and how it turns out, and to who are the winners, the losers, et cetera.

So, there’s different types of topics, and I think students go with what they know they can work with, and what they want to work with.

Ian Curtiss:

This brings me to the original thing that drew me toward your work: your presentation at Connections UK a year ago. I remember you defined wargaming as the combination of game theory, agency theory and relationship theory. And I think that’s just a beautiful way to describe it.

Pijus Krūminas:

Thanks. So, my basic idea is that wargaming stands at the crossroads of different social science methodologies, divided into two large parts. One group is specifically focused on determining the relationship between different factors, agents or whomever. And that can be purely empirical models, such as regression models, econometric models or system dynamics models.

And the other part is where you talk about the model’s decision-making capacities. That’s less about defining the relationship, and the trend, and more making assumptions about the actual behavior — so behavior modeling such as game theory, where you would make assumptions about rationality, or agent-based models where you would give rules to individual agents and then run a computer simulation for thousands of times to find out whether you have emerging patterns.

Wargaming can benefit from those different areas of research — for example, about emerging macro patterns from agent-based modeling. If you need to set up specific incentives in a war game, for example, you can look at the mechanism design literature, and then you structure your policy based on that specific goal. And the other part is we need to know how these things actually work, and that’s where all of these relationship-focused models can contribute.

Wargaming can borrow from all of those. And at the same time, wargaming can give back to each of these different fields, in that if you have a game theoretic model, a war game can allow you to test the assumptions or behavior. You can see what actual behaviors people do.

Ian Curtiss:

This is the beauty I find in this whole methodology. But the frustration is that it’s really difficult to get hard data from games in a practical manner without putting them into a computer and just running them a thousand times. Like you said, nobody has the time to play a game a thousand times. So, how do you leverage the games to achieve learning outcomes with your students in the political economy sphere?

Pijus Krūminas:

I heard a comment some time ago at the end of the course when students said, “No, I hate board games, but these make sense.” So, I think the important part comes from demonstrating that a game can be more than just a fun way to spend time. It can still be a fun way to spend time, but at the same time it is teaching something. And this helps to connect the subject to the learning outcomes of the program in general, because it contributes directly to their modeling skills.

It also contributes to critical thinking, and it contributes to independent thinking and independent learning through the game’s decision-making process and then arguing why you made one or another decision. And in the end, when they have to provide their own projects, it’s not only the rule set that they have to provide, they also have to provide designer notes where they explain why they made a choice for the game mechanics.

Ian Curtiss:

The powerful part of game design is, as you said, the rules and the game notes. And so not only are you as a designer having to think about what you are modeling, you have to create an argument, and a thesis behind your argument. You’re also building strategic empathy, as I call it, for the decisions people have to make in these scenarios.

And that’s a totally different outcome, and a totally different service you provide to a player instead of a reader. I’m curious, do your students see it that way?

Pijus Krūminas:

One of the things we discuss, actually, from the very start is who you are representing in the game. When they start working on their own games, that’s one of the first questions they have to answer. So, who exactly are the players? Because knowing that question shows you what decision space you can have, what is outside of the player’s control, and what is under the player’s control. If you don’t answer that question — who are you as a player, or as the decision maker — you can’t have a game.

We had an interesting talk today at the university, and I think one thing that came through quite strongly was the expected impact games could have on society in terms of education, and making decision-making better, and also working with different sectors to help them make better decisions, either through participating in games, or giving them skills they need to learn the game. So, I think there’s also an identification of these activities with longer-term outcomes, outcomes that are outside of the university with broader spillovers in the society.

Ian Curtiss:

What are the biggest surprises you’ve seen from your own students and the games they’ve designed?

Pijus Krūminas:

That’s a great question. One thing that stands out is the variety of topics. I don’t think we’ve ever had a repeat topic. So, certainly, you see the different interests come out of the designs, and other times it can be just integration of the games they know well, and transforming those mechanics into something else. For example, the game on privatization, they used some of the mechanics from Ticket to Ride.

From the games we play, what we see is that the ones we start with usually have the largest impact on students’ own designs. And so you play something, and then the mechanics carry over, et cetera. So, it’s interesting how these early games can have an impact on the final models in a student’s game.

Ian Curtiss:

That’s beautiful. What a great example. In general, we humans are limited by the models we are given.

So, I always talk about the games I make, and the research I do, mostly on gray zones — where the boundaries are moving in society. But “gray zones” are also talked about in war activities. My expertise is in China, but Russia of course does a lot of gray zone activities, and you’re in Lithuania very close to Ukraine.

I’m curious from a wargaming perspective, how is this all playing out for you in your field, and how do you see the field being applied there in Lithuania?

Pijus Krūminas:

That’s a very good question, and I think that’s the key benefit of wargaming, or riskgaming. It lets you sit at the crossroads with a lot of uncertainty and with a broad range of scenarios that could happen. I mean, if you have a kinetic war, it’s something you could reasonably analyze using other methods because you have a lot of data. The problem with the gray zone situations is that you don’t have the data.

And I think that’s where Riskgaming, or wargaming allows you to test out these different strategies in a safe and relatively cheap environment. For example, this spring, we had a “polycrisis simulation,” and we focused exactly on non-kinetic effects. We had an event that was… first it was tariff related. Then it was an increased presence of Russian and Belorussian troops along the Kaliningrad and Belarus border. It was not spilling over to a hot conflict, but it affected the investment climate. So, we didn’t go into this kinetic war phase, we just looked at what could happen otherwise that would affect the economy.

Coming back to the second part of your question on the regional awareness of wargaming, I think it’s increasing. I’m not necessarily talking specifically about Lithuania, but talking about around the Baltic Sea. The Polish are organizing their first wargaming conference in October, and want to get people from the region from Central Eastern Europe to come and join.

Ian Curtiss:

To wrap things up, two questions. What five or so games would you recommend for a wargaming class, and then what are your favorite economic, or political economy games?

Pijus Krūminas:

Okay, so the first question I can answer with what I always start with: The Shores of Tripoli by Kevin Bertram at Fort Circle Games, because it looks beautiful, and for a first engagement that’s very important. The game also structures the narrative well. It’s a topic that my students don’t — well, I’m not sure if anybody know much about it, including me before I got the game. It is about the Tripoli pirates, and Tripolitania, and how the United States got involved there. The narrative flows so easily through the game, and it’s really easy to engage students, plus it has a lot of dice rolls, which helps build excitement.

What I then like to play is John Company by Cole Wehrle, because it allows players to pose interesting questions. It’s sort of an economic game, and you are basically shareholders running a company, in this case the East India Company, but it also has all these questions hiding behind the players’ actions. So, what are you actually doing in India? Your actions can have very negative consequences. The game award system incentivizes you to prioritize your own personal gain over the company’s. And I think this leads to interesting questions that can then be discussed.

My students also always use Maria from Histogames. It’s a light war game, but it does well with the point-to-point movement, battle mechanics, and how you can use cards as a first introduction.

I also use The British Way, one of the COIN games from Stephen Rangazas, and I think it works very well because we play four scenarios. The model is very easy to see, what are the different relationships, and so on. And because we run different scenarios, we can then compare. So, if the core system is similar, you have some modifications, and they can help you tell a different story, whether it’s in Cyprus, or Malaya, or one of the other scenarios.

And this year we also played Churchill, which I think added a good diplomatic element to the games we explored because the war is very abstracted there. You advance toward Berlin, or you advance toward Japan. But a lot is going through the cards where you send guys to deal and negotiate over different issues.

When I have a larger group and can’t run a physical game, I like Take That Hill by Phil Sabin. I made a PowerPoint version of it, so I can easily move stuff and I ask players to vote using Slido. We run the game once or twice for them to internalize how it goes and then I show how it would look if you map out all of the relationships. So, basically map the model out, and it’s so confusing. Suddenly something you learned easily within one, or two turns, is confusing. So that is a good illustration.

And also this year with my students, I played Daniel Bullock’s The Gods Will Have Blood, it’s about the French Revolution, and it’s a solo game. So all the students acted as one and sort of voted for what decisions they would make. You are the judge or magistrate during the French Reign of Terror and during the French Revolution. And so you have to either condemn somebody or not, and we made it a bit ceremonial. It worked well to show how a game can also be emotional, and how it can tell a story of you as an individual can get pushed by the system.

Ian Curtiss:

There’s nothing like the French Revolution to educate people on how systems can push you towards terrible ends.

Pijus Krūminas:

Yeah.

Ian Curtiss:

Well, Pijus, this is fantastic. I so appreciate this, and love the conversation, and you dove so deeply into so many different aspects of this. Thank you so much for joining us.

Don’t miss

Yes, it’s the alfalfa. Yes, it’s the golf courses and swimming pools that dot the bone-dry desert. Yes, it’s the new chip fabs. Yes, it’s the massive population growth in one of the most inhospitable environments to humans in the continental United States. And yet, like the metaphorical beast, Phoenix is rising from the sands of the geologic past to become one of the most important global regions of the 21st century. How the city and Arizona more broadly manage that crucial three-atom elixir of life will determine the destiny for themselves — and all of us.

Southwest Silicon is a brand-new, full-immersion Riskgaming scenario for 8-9 players that simulates the tough economic, social and political tradeoffs of the American Southwest. We’re launching it today, and all of the materials that you need to play are now available for download.

The game was created by Ian Curtiss, our independent Riskgaming scenario designer who lives in Phoenix and has watched as the hulking concrete structures of chip fabs and office parks have emerged from the wilds.

The gameplay of Southwest Silicon centers on water rights and the frenetic competition — and requisite tradeoffs — over these all-important legal contracts that can be bought and traded across dozens of stakeholders with their own interests and biases. Ian modeled Arizona’s water market after extensive research and real-world reporting, capturing both the obvious economic implications of these trades but also their direct and indirect effects on families and communities.

Three players are the mayors of Phoenix, Chandler and Peoria, all trying to attract brand-new industries to their cities in spite of residential growth and increasingly scarce water supplies. Droughts are an omni-present threat, potentially triggering water shortages that can lead to immediate drops in polling and a political crisis. Should you gamble the city’s future through over-expansion that might never be a problem in the end? Or should you play it safe, but potentially lose a crucial economic development project to another regional city?

One player takes on the role of the governor of the Gila River Indian Community in central Arizona. The nation has bountiful water resources, but has its own challenging tradeoffs to balance like protecting its Native American legacy and influence while also taking part in Arizona’s economic success.

Other players take on the role of politicians trying to pass legislation that changes the game’s rules, consultants who are brokering deals for profit, and the local executive of the chip fab, who wants the best deal possible to drive the company’s stock price to its highest levels yet.

Southwest Silicon is a local game with global implications, one that we have played in the United States as well as in Europe and Asia. The challenge of growing under constraints is a commonplace, and the liminal space between positive-sum growth and doom-loop collapse can be astonishingly narrow.

In short: get a group together, play the game, and let us know what you think!

Get Southwest Silicon

For today’s Riskgaming podcast, I talked with Ian about how he approached designing the game, what he learned in the process, and what the experience can teach players.

This interview has been edited for length and clarity. To hear more of the conversation, please subscribe to our Riskgaming podcast.

Subscribe now

Danny Crichton:

Ian, your game centers on all these new chip fabs coming into Arizona. The context is obviously the Biden administration’s passage of the Chips and Science Act, which provided $55 billion of subsidization for the chips industry. Those loans and grants were sent out to TSMC, Intel, Samsung and others.

We go to DC, and a lot of people are interested in the growth, jobs, and economic development these new fabs could bring. But one angle that has been undercovered is water security. Arizona is a desert. Chip fabs are some of the greatest water guzzlers anywhere in the world. The Colorado River has had historically low water levels.

You created this game essentially to capture what you were seeing in your own hometown of Phoenix. Tell us a little bit about your story here and what you brought from your own personal experience into the game.

Ian Curtiss:

Yeah, I’m calling in from Phoenix, Arizona in the middle of the desert. When I moved here as a kid, I was like, “Nothing should be here. This is a miserable place. Why does anybody live here?” And yet there’s a miracle to this place because of the water brought in from the Colorado River. It’s this incredible feat of humanity that we can move water across the world in this way.

As you said, the water guzzlers — chip fabs — are rapidly expanding in the Phoenix Valley. So there’s this ongoing conversation around water. Where is water going to come from in the future? It’s a fascinating blend of angst and endless hope and faith. There’s angst about the water, but then there’s this endless belief in humanity’s ability to find more.

There’s all sorts of wonderful books and stories about this stuff. Cadillac Desert was written, I think, 40 years ago now. And it’s a wonderful book that goes into the history of the American West. Essentially the Bureau of Land Reclamation — it all started in the FDR days — said, “let’s start building dams and aqueducts to get water to the American West,” because it was underpopulated compared to the east. And one of those projects, of course, was the Hoover Dam on the Colorado River. And all this led to water being able to create Phoenix, Arizona.

When you’re in Phoenix today, there’s such rapid growth of suburbia. It’s urban sprawl on steroids times ten. And now all the chip fabs are coming into town. TSMC is making the largest investment in American history building its fabs out in Arizona. The most recent update was to promise that 30% of its three-nanometer (3nm) technology would be produced in Arizona at some point in the future. We’re reaching this interesting point where: What happens? Does this boil over, or rather, dry up? And where do we go from here? And so that was really why I wanted to make this game.

Danny Crichton:

You’ve had farming in Arizona for a long time, in some cases over a hundred years. And then you have these two new green shoots for the city. One is residential growth, as you were saying. You have hundreds of thousands of people moving into the region — all across Arizona, New Mexico, Southern California — who are looking for warmer temperatures trying to escape the New England winter. And then there are these semiconductor plants, each of which use millions of gallons of water potentially each day.

And so you have this unique tension between old industry, new industry, old population, new population, which makes for a great Riskgaming scenario. Maybe you could talk a little bit about how this plays out with the stakeholders within the game.

Ian Curtiss:

With industrial policy, it’s so local. But the core of any policy is that you end up creating winners and losers. And so when you start throwing government money around, and when you start getting such huge investments into an area like chip fabs, it can make or break a community. It literally changes the daily lives of people to have this kind of investment.

And so the stakeholders, they are a series of NPCs I put into the game. They’re a way for the eight to nine people who sit around the table playing the game to watch Arizona change based on the decisions they’re making. Some are farmers, a good many represent everyday businesses. Some are existing chip fabs.

And so me, as a player, as I’m playing the game, I am getting data feedback at the end of every scene about what my stakeholders want. I’m trying to calculate not only the opinions of the people around the table as I negotiate with them about what I should do in the game, but I’m also having to keep in mind these NPCs that still have an opinion on the matter at hand — much like real political life. They might not be at the table, but they still have an impact on the outcome.

Danny Crichton:

Arizona is one of the most contested battleground states in the country. You have a split legislature, you have a split in state-level offices and in the congressional delegation. One of the things we tried to include in this game is the factional dynamic between farmers who are Republican, farmers who are Democrat, new business owners on both sides. It’s not always easy when you have several dozen stakeholders to build up a coalition of people who all support you. You take one action, their responses don’t necessarily line up perfectly with each of those adjectives and characterizations. That creates a much more dynamic and I think rich game in Southwest Silicon compared to a lot of our others. So I’m really excited for people to try the stakeholder model.

Obviously, the best way to learn from the game is to play it. But what do you think are some of the key takeaways you want people to experience?

Ian Curtiss:

Well, first and foremost: water as a topic. The factoid I love to drop is that the most recent time one U.S. state deployed troops against another was in the 20th century over the Colorado Dam. Arizona sent troops to the Colorado River to try to stop California from building a dam that Arizona had not yet agreed to. So water is this incredibly poignant, critical asset that people need to survive. People are literally willing to deploy forces to protect it.

Beyond that, though, is the winners/losers dynamic. When you see news about drought, you see how this person is going to win from water, this person’s going to lose from water, there’s plenty of water for data centers, there’s not enough water for data centers, chip fabs are using all the water. You have to get more detailed and granular about the local realities of water and understand where the water comes from and understand who’s going to get the water and who’s not.

Water is a commodity. It can be bought and sold, so economics is inevitably going to be the decider. When somebody says there’s not enough water, the implication is that there’s not enough water for everybody to have everything that they want, but there is enough water for certain things to be chosen over others. I will pay more for water than you will, therefore, I get to water my plants more than you do.

But society is going to be choosing not only based on economics — taxing and so forth — but also on the infrastructure required to get water where it needs to go. And this gets back to my earlier point about the BLR, the Bureau of Land Reclamation. There’s all these little water districts across the American West with localized water prices, and sometimes they’re significantly cheaper in one district than they are in another, even just a mile away, because 40 or 60 or 70 years ago when the dams and aqueducts were built, one of the dams or the aqueducts was far more expensive to build than the other.

So here we are now: these water districts are not very economically productive ways to use this water, and a chip fab wants to come into Arizona. So where’s that water going to come from? This game works through how people will wrestle with these decisions: Do we want farmers to have water? Do we want chip fabs to have water? Do we want houses to have water? And if it’s not the farmers in Arizona, then where’s American going to produce its lettuce?

Danny Crichton:

When I think about the takeaway, we’re trying to complicate the picture a little bit. We’re not trying to complicate it so much that people are lost, but we’re trying to take something that has been very simple, very one-dimensional, usually around economic growth. We’re trying to add in a lot of other factors that people aren’t necessarily aware of. We’re taking advantage of your local knowledge in Phoenix and we include a lens on U.S. politics, specifically the balance between local affairs versus federal affairs.

There’s a little bit of a plot twist in the game that also highlights the individual versus collective goals of different groups of people and how sometimes, even if you’re a politician or a CEO, your personal motivations can become incompatible with your leadership goals.

I think Southwest Silicon is a great example of Riskgaming at its finest. It brings together a lot of different trade-offs that people are not necessarily considering. So I think it’s very mind-expanding, and I’m looking forward to everyone playing it.

We just played this game in London with some British bureaucrats and locals. Were there any surprises from that group? I think some people knew where Arizona was, but most did not.

Ian Curtiss:

Well, first off, I have to defend my home state of Arizona. They did know where Arizona was, they just didn’t know where all of our cities were.

But the game in London was great fun because, again, it’s so local. Brits expect water to fall from the sky everywhere they go, and that leads to many assumptions about how your society would function and about the role of infrastructure and localized economic policies and decision points. When we did the game pre-briefing, we quickly realized the players weren’t used to mayors making these sorts of decisions. They weren’t used to it being an issue that a mayor could buy or sell water with local businesses.

What was really fun was watching, as with any culture when we go around the world and play these games, how different people play the game differently. For one, our British friends had a blast stepping into American roles, and some of them were like, “Oh, well, if I’m going to be an American, I’ll be American.” And they leaned into the role-playing aspect.

The last thing, though, was their perspective on the chip development process. And again, as an American, we just take for granted that chips are an American-designed thing. For the players in London, so much of the content wasn’t as naturally a part of daily life. So that was the other interesting thing — catching them up on some of these assumptions about the chip fab investments, what they mean, the laws, the process for the CHIPS Act to be expanded. Getting our friends from across the pond to wrestle with these things was a blast.

Danny Crichton:

I want to go broader. We beta tested this game with a lot of folks. We’ve been designing and producing it. We’ve got booklets here now, so it’s all coming together. But there’s water issues going on all over the place.

One big story this summer came from China, a place you lived in for a very long time. Last week, Premier Li Qiang announced the construction of a new dam in the Tibetan Plateau at a current cost of $170 billion. This is the largest dam ever built in the world, projected to potentially produce as much as 3–4% of China’s GDP. It would produce enough electricity to power all of Britain. I talked about it in a newsletter back in July.

What was your take on this? You spent six years in Beijing, and you’re still doing a lot of work there. Were you surprised by this announcement and the scale?

Ian Curtiss:

What was so fascinating to me was, first, well, it’s China being China, right? Incredible projects, incredible mobilization of capital. I mean, there are theories that centralized projects to protect people from drought and flooding along the Yangtze and Yellow Rivers are core to Chinese political identity. And so now you see this play out in Tibet.

The second thing that came up for me was a memory of a hugely impactful moment in my time in China. We had a course on non-traditional security, and water security was one of the issues we talked about. There was also ag security, there was biosecurity. At the time, the foreign students thought all these topics were just Chinese communist paranoia. Like, “the market will take care of this if you would just get out of the way.” Well, here we are, and what is the U.S. government talking about every day? All these non-traditional security topics. The Chinese government has been thinking about them for at least a decade longer and in far more depth and nuance.

But regarding water specifically, there’s another thing that’s interesting about this game and the development in Tibet. Throughout the PRC’s expansion into Tibet and Xinjiang, there was open talk about applying U.S. government methodologies to Native American tribes in the Chinese West and South.

Southwest Silicon, interestingly enough, has a Native American tribe in it. That’s one of the roles people can play. And one of the things playing out in the United States today is essentially a series of legal debates and court cases (as well as now a federal law) to allow water to be transferred back to the tribes who were forced to move to all these arid regions. To oversimplify, the American government owes them a lot of water. If you put that in terms of Chinese perspectives of water security, they already figured this out and now they’re building the dam.

Danny Crichton:

I wrote about rivers in the Riskgaming newsletter this summer responding to In Praise of Rivers by James Scott. He writes about the Yangtze and the Yellow Rivers in China. They move dramatically throughout the years, in some cases, by as much as hundreds of miles as they go towards the ocean. And so there’s a real need all the way back to control the flows. I think it’s like the sixth century when the Chinese built this massive canal that connected the two rivers together. I mean, everyone’s heard of the Great Wall, but we’ve never heard of the Great Canal. These massive water projects become really crucial.

And then you come over to the United States. The character we have in Southwest Silicon is part of the Gila River Indian Community, and there’s recently been an update from the Ninth Circuit Appeals Court re-judicating some of the legal rights over water between farmers and the Native American tribe. This is a live issue in the United States and it’s not determined by engineers, it’s not determined by needs, it’s not determined by economics, it is literally, in some cases, determined by decades-long legal precedents going up and down the courts.

And so it’s one of those open questions: Is this a good allocation mechanism? Is this fair? Is this what we’re trying to do? The game is trying to explore that.

Ian Curtiss:

I could go on about this for far too long, but one thing that happens in the game is the tribe ends up having a shocking amount of power because they have rights to a lot of water. This very much mirrors reality today. There is a river called the Gila River that goes through Arizona, and a hundred years ago, give or take, American white settlers came into the region, dammed the river, set up farms, and cut off all the water that was flowing to the Gila River Indian Community. The tribe’s livelihood was pretty much taken away.

Now, years later, they are getting back all the water rights that they should have had all along. And so here they are with this incredible negotiation power. What a generational change.

Don’t miss

These are huge questions which will ricochet throughout the tech economy. Thankfully, we have a veteran hand this week to go over it with us in the form of Dylan Patel, founder, CEO, and chief analyst of Semianalysis. He’s the guru everyone reads (and listens to), covering the intricacies of chips and compute for a readership of hundreds of thousands of subscribers.

Joining me and Lux’s Shahin Farshchi and Michelle Fang, we discuss the questions above plus how Mark Zuckerberg is transitioning Meta Reality Labs, the hopes and dreams of new chip startups, the future of AI workloads, and finally, Intel after the U.S. government’s purchase of 10% of its shares.

This interview has been edited for length and clarity. To hear more of the conversation, please subscribe to our Riskgaming podcast.

Share

Danny Crichton:

Your beat has to be one of the most exciting in the world right now. There’s so much going on, most prominently the U.S. government’s investment in Intel. But that’s not all, right? Everyone from the hyperscalers, to materials companies, to U.S.–China, everything is coming to a head. What are the top priorities you are seeing right now? What are people not paying attention to?

Dylan Patel:

As revenue ramps for these labs and they continue to complete these massive rounds — whether OpenAI’s $30 billion round or Anthropic’s most recent round, which is well north of $10 billion — it tells you what their spend is going to be for compute. By the end of 2026, OpenAI could be spending anywhere from $50–70 billion on compute. Anthropic could be spending $30 billion, which is a humongous step-up.

Each of those companies will have more infrastructure deployed through their cloud partners than the clouds themselves had deployed as a whole when ChatGPT was released. So the level of spend here, the level of infrastructure build out, it really is mind-warping.

You get a lot of folks saying, “Oh, there’s no ROI here.” But actually, the ROI is there in terms of the revenue curve. Whatever dollars they’re making from services and inference are actually being spent on compute, of course, but a big portion of that is being round tripped to training.

Danny Crichton:

A few weeks ago we got GPT-5. And one of the most important facets of GPT-5 is it’s really not just one model. It’s a router across multiple models, ones that are very compute-efficient, some that are sort of in the middle of the performance curve, and then Deep Research and others that are extremely compute intensive.

To what degree do you think chip efficiency drives the cost curve down for OpenAI, Anthropic, and the other LLM models versus algorithmic improvements, not just to the inference model, but to these routers?

Shahin Farshchi:

I can take a stab at that. Ultimately, it’s hardware and software that go together to build these kinds of advantages. You made a comment, Danny, on chip efficiency. I would take that a step further and focus on system-level efficiency. You can make the chips super efficient, but if they’re not being utilized, then you’re not getting a return on your investment.

A lot of the more recent innovations are not only around chip performance but also memory performance, system-level performance and ultimately, chip utilization or investment utilization. I think there’s still a lot of opportunity there. Nvidia has done a great job with Blackwell, but I think there are greater improvements to be made. And there are a lot of great startups, a few of which that we’ve backed.

Dylan Patel:

The other aspect of this is that it’s really hard for these companies to balance their infrastructure today because of how bursty the traffic is, whether it be to Deep Research, whether it be Claude code usage being so high during certain hours. Volumes of all things on the internet undulate a lot throughout the day, but AI is even more spiky. And then you layer the constraints of service-style plans — say $200-a-month or $20-a-month type plans — or even free plans, and you have a double whammy there.

How do you deal with that? For OpenAI, it was adding the router. For Anthropic, if you’re using Claude code at night, they upgrade you to Opus, but if you’re using it during the day, they make you use Sonnet. And they have these hourly rate limits for that specific reason. Another way of doing this is what OpenAI does with “batch query,” where you can get half price off the API call if you batch it.

The nice thing about AI is that a lot of it is not latency sensitive, right? It’s not like search. If I use a thinking model, or if I’m using these sorts of applications that aren’t immediately responsive, there’s actually some time to ping a server on the other side of the nation or all the way to another country. So that’s another aspect of this infrastructure build out that’s really noteworthy.

Danny Crichton:

I don’t know if it’s apocryphal or not, but the argument for Amazon Web Services was essentially that Amazon had this massive surge of traffic during the winter holidays, when they had to double or triple their infrastructure. And then what do you do with all this hardware that you’re not using for the other 11 months of the year? The argument was, well, we can make that a platform, we can sell it off.

But how do you balance between the kinds of AI algorithms that you can batch, you can do overnight — “Hey, go through my data, I want a report the next morning” — and those that need an immediate response — “I’m on a call with a customer support agent, I need an answer in the next five to 10 seconds”?

Dylan Patel:

That’s a critical question, and it is hard to answer. OpenAI had to do so much work to get their voice mode to work. If you look at other voice modes out there, one way they solved it is they make the beginning section of the query go to a dumber model. They can respond faster, so you immediately get a response.

A lot of times, even humans throw “ums” and “yeahs” in the beginning of a conversation so you get time to think about the answer. If it’s a voice agent, it’s the same thing.

If it’s a robot, maybe the high intelligence is doing the route planning, but the on-device model for the robot is actually doing the fine motor feedback.

Danny Crichton:

Let me pivot the conversation, because I think GPT-5 is opening a new era of development for these companies. OpenAI is almost 10 years old. They have a product that’s used by hundreds of millions of people, arguably a billion if you include a lot of very soft users. That’s the fastest growth of any consumer app in history.

And so there’s a huge question of, as these companies are optimizing their models for performance and efficiency, how do they start to adapt to consumer use? Do you see an evolution of the business models?

Dylan Patel:

You can already see certain things changing. With Anthropic, it initially it seemed like $200 a month was easy. Then people used it way, way, way more. There were people on Reddit showing $20,000 worth of usage in a single month versus the $200 for the plan. And so they had to introduce rate limits, and they’ve had multiple iterations.

Pay-per-usage is really, really the hope. You pay per usage, you get some margin on that. But at the same time, the subscription model is the one that customers understand more.

The monetization method is where there’s a lot of work to be done, especially in the value-capture side. So I think it’s truly an open question as far as $20 a month, two hours a month, they’re going to release a $2,000 a month thing — and I’m going to buy it for every single person who works for me. I just know that’s going to happen. But it’s tough to say exactly where it goes from here.

Michelle Fang:

To double-click on the business model, Dylan, do you see a world where AI interfaces like GPT-5 become superapps that end up both serving ads but also becoming the platform that facilitates purchases?

Dylan Patel:

A week and a half ago, two weeks ago, we had a post that said GPT-5 sets the stage for ad monetization and superapps. So many people saw that title and angrily DM’d me like, “They’re not going to do ads, we’re not going to do ads.” It’s like, no, but that’s how you clickbait people into reading it. And then you explain the superapp. Their current CEO of Applications built AI monetization via purchasing on Shopify.

And what is a superapp if not something that can do everything for you? So you open up WeChat and it actually lets you book an Uber and it lets you book a nail salon appointment and it lets you get a haircut and it lets you pay people and it lets you buy things and it lets you chat with people. Obviously, it lets you do everything, and you’re going to do everything through AI. That’s the hope. That’s what Meta wants to do. That’s what OpenAI wants to do.

I think if you gave an AI app the option of delivering something to your home — like, “oh, hey, I bought you a pizza” — the user would love it, and then you take a “take rate” on it. I think the take-rate model has to be the easiest way to monetize the user. You give them your credit card information, you authorize them to take X percent of any purchase, and then that’s it. Just like Uber does, just like Airbnb does.

Danny Crichton:

We’ve placed a huge amount of focus on the application layer, and we’re talking about the infrastructure layer. But there’s also an innovation layer that goes across all of them, which is all the new chip companies. So we’re in a world in which Nvidia is one of the most valuable companies that’s ever been created in the history of humanity. But at the same time, you see this massive expansion of investments into a whole slew of additional companies trying to specialize in other parts of the stack, trying to compete directly within Nvidia, or just find their little niche.

When you look at the startup landscape right now, what are you seeing?

Dylan Patel:

People want to do everything and capture the entire value of, say, an Nvidia, and compete directly with them. But at the same time, if the infrastructure spend as a whole is skyrocketing, right, the majority of that goes to chips. Over time, the hope is that a lot of this moves away from Nvidia to other vendors, whether it be Google’s own chips, Amazon’s own chips, Meta’s own chips, et cetera. You do start to get more proliferation, OpenAI has a chip team. And then you could be a vendor for those firms as well.

Danny Crichton:

So you do a Shopify to Amazon. Amazon is a one-stop shop, and then you create a platform that empowers a whole ecosystem of new shops, or something like that. You take a different model, a different approach.

Dylan Patel:

Correct. And Shopify was very successful, so why don’t you do this with chips, right? There are many customers. Nvidia can be your customer, AMD can be your customer, as can all these other hyperscalers.

Danny Crichton:

Shahin, when you think about the startup ecosystem, how are customers picking from a wider ecosystem of companies?

Shahin Farshchi:

Customers are spreading the net very, very wide.

But one point to keep in mind is that a lot of startups like to talk about unseating Nvidia, but you don’t necessarily have to unseat Nvidia to build a massive company. Go back 10 years ago, AMD went from a single-digit billion-dollar market cap company to a double-digit billion-dollar market cap company just by taking single-digit percentage market share away from Intel.

You can very much build a massive business today just by taking some market share away from Nvidia. And the way you do that is by understanding where there could be gaps or becoming a viable second source to Nvidia for customers that don’t want to be locked in.

But to the points Dylan made, there are many ways to play this game. You can be a vendor into Nvidia or into any of these companies that are building accelerators and still be an interesting business. In fact, packaging and testing are areas people haven’t been paying attention to that are now becoming extremely important. In fact, if you look at the AI chip scarcity, it’s driven by packaging and integration more so than the semiconductor itself.

Danny Crichton:

Dylan, we overlap on a lot of the technical and business subjects, but you and I have a massive disagreement, which is the U.S. government’s purchase of 10% of Intel. I think it’s a catastrophe. I have never seen state capitalism in the United States prove successful. You are more optimistic. Tell us why.

Dylan Patel:

I don’t like state capitalism either. I think it’s a terrible idea. We win by being an economy that’s competitive and capitalistic. And there are certain ways to do state investment into industries that actually improves those, right? Look no further than Chinese auto. It is the most cutthroat competitive market in the world. Yes, there are some SOEs, yes, there were a lot of subsidies, but actually it’s extremely competitive.

And so when you look at fabs, obviously I would love for that to be the same case, but we’ve got some different realities. The first is the CHIPS Act, it was a... I don’t know. What’s your opinion on the CHIPS Act generally? Subsidize fabs a little bit and get them to be built here? Subsidize them enough to make it a little bit better than building fabs elsewhere?

Danny Crichton:

Chris Miller, Jordan Schneider and I wrote a paper pre-Biden administration, which in some ways became the CHIPS Act. Parts of that legislation were directly from that paper. For example, the regional innovation hubs idea was pulled from our paper. So I mean, I’m generally in favor. Generally the idea was moving the economics around, but it was not about ownership of companies.

Dylan Patel:

When you take a look at the CHIPS Act and how it was administered, there were certain things that had to be done. Intel had to ramp to a certain level of volume on their fabs to get the money. Samsung, same. Likewise TSMC. There’s no question TSMC is going to ramp to that volume in their fab as much as they kick, moan and scream about it being too expensive.

Now with Intel, it’s slightly different. Economically, it did not make sense for Intel to build an Ohio fab. But they lobbied for it, got the grant — or guaranteed loans is what I should have said. But Intel will not ramp to that volume because they didn’t end up getting a customer for their 18A technology.

Now, 18A is not a complete failure. They’re actually manufacturing some chips right now. New laptop chips will launch with it in the next six months. From everything the Dells, the HPs, and the Lenovos are saying, the chip actually looks okay. It looks competitive with AMD, which is on TSMC process technologies. It’s not bad, it’s not completely behind.

So Intel is at number two, but this industry is very, very capital intensive and very, very difficult. And the switching cost is absurdly high. If you’re in this race to build whatever chip in whatever market, do you just dedicate engineering resources to switching foundries to use a slightly worse node with higher risk on yield or do you invest in improving the architecture?

You only get X percent improvement from node, but then you get a larger improvement from architecture and then you get an even larger improvement from software. There’s these stacking curves, so you might as well focus on whatever’s further up the curve — which is as much investment in architecture as possible.

And so Intel has this really hard problem. They must leapfrog TSMC to get any customer business, otherwise it makes no sense. So then you’re stuck. Intel is the only American fab. It’s the only company in America that’s doing process R&D. They are behind — not absurdly so — but they are behind. And it is extremely strategically important that there are fabs in the United States. A Taiwan invasion is a true risk and a true possibility. There are many parts of the U.S. government that believe it will happen this decade, but many folks think it’ll happen in the next five years.

By 2028, if Intel does not get more money, I think they’re bankrupt. And so you have this really challenging problem where, effectively, you must subsidize them for geopolitical reasons. But how do you do that if the method we created via the CHIPS Act won’t work — they’re not going to meet the obligations? They’re not going to ramp up Ohio in time. In fact, they can’t even afford to ramp up in Ohio because their business has lost share faster than expected.

The fab side needs to be subsidized somehow, and the fastest mechanism was converting the failed obligation on the CHIPS Act to an ownership stake.

Danny Crichton:

Shahin, Michelle, I know you have comments, so what are your thoughts?

Shahin Farshchi:

Listen, this is all symbolic. This is all political. Ultimately, Intel doesn’t need to be owned by the U.S. government. There are many smart investors out there with tons of access to capital.

If you look at TSMC or these other great institutions, a lot of them have been supported by governments. A lot of these things just take time and a lot of capital. In the case of manufacturing chips in America, it makes absolutely no economic sense to do this. It’s all for geopolitical purposes, and it’s a cost-benefit question. We’re basically purchasing a very, very expensive insurance policy here.

And the question becomes how much we are willing to spend on this. So far, the amount that we’ve spent seems to justify the cause, but this is not going to be limitless.

Danny Crichton:

Michelle, you were on Capitol Hill for a period, as AI advisor in the Senate. What are your thoughts?

Michelle Fang:

I would echo what Dylan has said and throw a question back: Do you think that there is enough political will for a CHIPS and Science 2.0? What’s happening on the R&D side? A lot of the funding is actually supposed to be allocated for packaging and whatnot, but do you think that there is going to be an evolution of this? And if so, what is the role of the state in the next couple of years?

Dylan Patel:

Yeah, it’s an expensive insurance policy. But I guess again, how do you think through this? If there is a 10% chance that Taiwan is now completely influenced by China, and they can start to do to us what we did to them — cut off Nvidia like we cut off Huawei — what is the impact to the American economy?

Holy shit, it is way, way more than 10% times Nvidia’s revenue, which is $200-240 billion. 10% of that number is more than we gave Intel already. And then you add on Google and all these other companies, you add on the knock-on effect of all the AI spend, right? There’s also the data centers and the capex for all these other things. Laying fiber, all this. Oh, and the AI impact on the economy. The GDP is $400–$500 billion higher this year because of all the capital investment. Okay, 10% of that? Well then, in one year it’s $40 billion, $50 billion.

Our expensive insurance policy is actually quite cheap relative to the 10% number I just made up. Government people know way better, but I just know that people treat this very seriously.

Danny Crichton:

Well, we’ve gone around the world and across the stack, but Dylan, founder, CEO, and Chief analyst at SemiAnalysis, the must-read guide for everything going on in semiconductors and chips, thank you so much for joining us.

Don’t miss