The long game of American reindustrialization

Charles Yang on government service, specialization, and changing the discourse

Reindustrialization is the word du jour in American policymaking circles. The hope is that a reinvigorated manufacturing base will bring back middle-class jobs and ensure our strategic autonomy in what looks like a tough century ahead. It’s a towering task, and it will take many strategic decisions to undo the last several decades of deindustrialization.

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

Our “just take it” era

Can software platforms reverse enshittification?

Why AI safety is like a bolt in a croissant