Jennifer Pahlka on Reforming Government

Conversations with Tyler is produced by the Mercatus Center at George Mason University, bridging the gap between academic ideas and real-world problems. Learn more at mercatus.org. For a full transcript of every conversation enhanced with helpful links, visit conversationswithtyler.com. Hello everyone and welcome back to Conversations with Tyler. Today I am here live in person chatting with Jennifer Pahlka. Jennifer is founder and former CEO of Code for America. She worked in the Obama administration and she founded the United States Digital Service. She has a background in computer gaming, media, and most importantly for our purposes, I'm a big fan of her 2023 book called Recoding America: Why Government is Failing in the Digital Age and How We Can Do Better. Jennifer, welcome.

So great to be here.

Thank you. What is your model for why Congress has become so passive, and how important is solving that problem for reforming the bureaucracy?

Oh, you did start with a hard one.

They're all hard ones.

Yeah. So Congress is— I'm sort of new to Congress. So I spent many years really in the executive branch agencies. So I see Congress through the lens of somebody —people who are downstream of their actions. And I've only recently come to any theories at all about Congress. But one observation I have, I think secretly will admit that they know that the laws that they pass don't actually work. They don't want to admit this. That is not something you'd want to run on. But I think increasingly they're aware that they are turning the steering wheel, but it's not connected to the wheels of the car. They're pressing the gas, but The car's not going anywhere. And I think that certainly contributes to— it's probably maybe not the only reason Congress is powerless, but I think it has really contributed to this sense of what are we even doing here? We're supposed to be steering this country. And, you know, when we turn the dial, like, nothing's— maybe not nothing is happening, but less than we thought, or something different is happening. This law of unintended consequences really kicking in.

Do we have a way of getting Congress to be less passive?

Well, I'm hopeful that these sort of various strands of disruption that are happening right now will push them in some direction. So I think the challenge is to shape that direction. Certainly, you have something like the Loper Bright decision, which is going to disrupt a lot of the sort of way they have done work. If you really want to write a law, have it passed, and then have it have the impact in the real world that you intended, You kind of can't really do what you used to do. What that change will look like is a very big question. I don't have any simple answers. But I do think it's a moment where you can say, look, we have been living in this essentially waterfall world, right? The waterfall process of you do this thing and then it sort of goes into this stepwise procedure, gets handed off to others, and it's really a one-way a one-way trip, right? It descends now through the hierarchy of the bureaucracy to get implemented. That's not how stuff that works well in our real lives works. It works in a much more iterative way in which you're sort of coming back to the intent, checking on how it's going. It's more of a quick feedback loop than this long waterfall that descends away from you and away from any control that you have over it. So some element of the change, if it ends up being positive, I think will be away from waterfall process and to something that's much more agile, iterative, and closing the feedback loop. And I do think that speaks to some of Congress's anxiety, right? They are anxious because they feel like the bureaucracy doesn't do what they told them to do. And it's not a surprise that they don't because of the structures. But the change that we're talking about, I think, is really profound. It's moving from one paradigm that is very, very ingrained in how Congress operates, how the executive branch agencies operate, but also just how everybody thinks. It's such a profound change. We are going to need a fair amount of disruption to make that change happen.

Now, we're talking in March 2025. To the extent Congress stays passive, that residual authority, where do you want to see that reallocated to? To the executive branch? To the courts? Because a lot of what you write is very critical about the courts, right? There's too much of the judiciary. But this is March 2025. And where should that power go?

Well, I mean, your question about Congress being passive in this moment, of course, I didn't really answer. I'm sort of talking more generally, I think, before the inauguration. But no, I would agree. We have put far too much into the judiciary. We are having them make decisions that I think executive branch agencies should make. And one of the big problems with sort of punting everything out into the courts is that those tight feedback loops require you to act quickly, and the courts do not act quickly. That's going to be a very big problem, I think, just generally for this whole era where everybody's counting on the courts to get their thing done, right? Especially on the left where there's a desire to stop Trump. Well, You know, if the lawsuits can't get through the system in any meaningful amount of time, the harm is already done. So that's already an issue. But it's an issue even, I think, in less dire times because you can't operate iteratively if you're just always waiting for a court decision or always waiting to be sued. And I think that's sort of how we've grown up is like we're, you know, if you Work in the bureaucracy, you're just always waiting to be sued.

But does that mean we need something like DOJ? So I've lived near the D.C. for about 40 years of my life. I haven't seen anyone succeed with regulatory reforms. You can abolish an agency, but to really reform the process hasn't worked. And maybe the best iteration we can get is to break a bunch of things now. That will be painful. People will hate it. But you have a chance in the next administration to put some of them back together again. And maybe it's just in a large country, there's no other way to do it. We have separation of powers. The first 2 years of Doge will seem terrible, but 8, 12, 16 years from now, we'll be glad we did it. Is that possible?

I don't know what's going to happen. I do think this is the disruption that we're getting, whether it's the disruption we wanted. The question of whether it could have been done in a more orderly manner is a tough one. I just feel sad that we didn't try.

Are you sure we didn't try?

So— I don't think we really tried.

You know, the second Bush presidency, people talked about this, what we need to do. Al Gore, some of that was good. In fact, Reinventing Government. We've been trying all along, but this is what trying looks like.

Yeah, I think Reinventing Government happened at a time when we were just at the beginning of this digital revolution. And so it was trying with a very 20th century mindset. Fine, did well within that context, but we don't need that again. We need 21st century change. We need true digital transformation. We need something that's not stuck in the industrial ways of thinking. So I don't think we tried that. And I think the efforts have just been too respectful of old ways of working and the institutions. There was really not an appetite, I think, for what I would call responsible disruptive change. What is that worth? Is there such a thing? I don't know.

But say you're approaching USAID, where I think the best programs are great. A lot of it they shouldn't be doing. On net, it passes a cost-benefit test, but the agency internally never seemed willing to actually get rid of the bad stuff. All the contracting arrangements, which made American congresspeople happy because it was dollar sent to America, but way inflated overhead and fixed costs. And why isn't it better just to blow that up? Some of it is great and then rebuild the great parts.

It's so hard to say. I've had the same thought. In fact, you know, before inauguration, I wrote about the Department of Defense, right? It's the same thing. There's a clear recognition by the people in the institution, as you saw with USAID, that this is not okay, that this is not working. And it's just strange to be in an institution that large where so many people agree that it's not working from the bottom to the top, and yet nobody can make really substantive change. Like, you look at efforts like Unit X, you know, there's this book out by Rasha and Chris Kirchhoff about their efforts to start the Defense Innovation Unit.

He's been a guest on the show.

He was great. Yes, of course, of course. Right. So, like, unbelievable talent, unbelievable support from the Secretary of Defense and actually then subsequent Secretaries of Defense and like good outcomes for sure. But is the, is the Department of Defense fundamentally more agile than it was before? Not meaningfully so. So we do have a real problem that isn't even just like that a lot of people are resisting change. I think most people want change, but somehow the system itself fights back against any individual who's trying to change it. And I told a story about being at an event where the person next to me— it's one of these events, you know, where people are off the record. And this person was a pretty senior official in the Air Force. And I said that I thought we would have a safer country with a smaller defense budget. And people look at me and they're like, ah, she's a hippie or something. She just doesn't care about our national defense. I spent 4 years on the Defense Innovation Board, and I learned to care a lot about our defense and to be quite worried about it. And I felt a little awkward saying this next to this Air Force official because it sounds insulting, like I want to cut your budget. And he turned to me, and in front of everyone, he said, you're right, except it needs to be 50%. Because when we've had 15 or 20% cuts, it doesn't shock the system enough to actually change behavior. Of course, you can have a cut like that would be a shock to the system. A hot war would be a shock to the system. And we all know we will see change when that happens. So I have, in a sense, been an advocate for this kind of shock to the system. And that whole, I was sort of alarmed in a certain sense when Doge, you know, came and said, we're gonna, Or I think Vivek Ramaswamy, when he was going to be on Doge, was saying just, we get rid of half the bureaucrats based on the last number of their Social Security numbers. Like, wow, what he's saying, what I'm saying kind of rhyme. And yet I feel like I come at it from fundamentally a desire for government to work. And sometimes it's hard to see whether what's going on right now is about government working better or just being smaller. I think it could be smaller and be better.

If we think about defense—

That better is still really important to me.

You use the defense example. If we think about what has worked, it seems to me that outsourcing non-bureaucratic activity to say Palantir, Anduril has been very effective, cost-effective, probably militarily effective. We'll see. And also Christopher Kirchhoff's unit has worked pretty well. So why don't we do more of that? Just take government agencies and sidestep them, like create a new NIH, call it something else, figure it will be good for 10 or 20 years, then we'll need to start all over again. But in the meantime, we cut the budget of the current NIH, say in half, and create a new agency like a DARPA for health, which we have a version of already. Yeah. Why not just keep on doing that? Create new things and destroy parts of old ones.

Well, I mean, that does happen. That's the, what do we call it? The bureaucratic rigidity cycle, I think it's called. Right. We haven't, I think, been disciplined enough about dismantling the old when we create the new. I mean, I experienced that myself helping start USDS, the United States Digital Service. It's creating something new, but it's even up until you know, inauguration day, just had far less power than the existing infrastructures, CIOs of agencies, you know, the procurement frameworks, all the, you know, that stuff sort of has a life of its own and a power of its own. And if you, you know, are creating these sort of skunkworks to disrupt the old stuff, it has to actually disrupt the old stuff. It can't just, you can't just add. If you're just adding new layers on to things.

Yeah, so start with the cuts. But like, what's something where right now you'd be willing to cut half of it? Well, not knowing what will be added, like no package deals. You've just got to hope and pray for the best.

What would I cut if I—

if something you'd cut half of some agency, some program, not farm subsidies, like everyone would agree, but something that would hurt a bit.

Well, I guess I'm on record as saying that the Defense Department should have a shock to the system. I don't know if it's 50%, but it does need to be freed from the the constraints that it seems not to be able to shake.

And the rest of government other than defense?

Would I cut it 50%? Yeah.

And hope to grow, you know, new limbs for the lobster, so to speak.

I guess I could imagine cutting something quite deep, but I would do it far more discriminately than what's happening right now.

What would be your first target? The discrimination is you're in charge. We all listen to you. You're dictator of the bureaucracy. And you can, other than DOD, you can cut something by 50% and we're all gonna do it.

So my metaphor is basically that we are imbalanced between brake and gas, right? So in any given agency right now, you could talk to someone who's trying to create something new, stand up a new, okay, I'll talk about the Direct File team for instance, right? They were building something that was better, that works for people. They did it cheaper, they did it faster than how things normally go. So you've got, I'll call them, you know, the go energy. Around them, you have like 10 times as much infrastructure, people, rules that are stopping them. That's the stop energy. I don't believe we should have no stop energy. I just think that the balance between those two things is way, way off. So if you're going to come in and change, and what you do, what you want is government to be more effective, go faster, you know, build stuff. And I mean, like, build things in government, like, you know, Direct File, but I also mean enable the building of housing and, you know, energy infrastructure and transportation that we desperately need in this country. You've got to go identify Where is that stop energy and how do you reduce or remove that so that you can enable the go energy and the people whose job that is to get the thing done? You have to go do that analysis or you're just getting rid of the go energy first. Like, you know—

But a lot of the New Deal, it happened very quickly, right? There wasn't always much analysis behind it. It's like, let's set up a bunch of things. A lot of them did not run well at the beginning. I don't mean that as a criticism. It's just inevitable. Like who's on the staff. So if we're going to recreate government, doesn't it have to be at least as chaotic? Well, it is chaotic. Well, but there's something very orderly about the current bureaucracy. If we want to improve the SEC, HHS, whatever other agencies, yeah, don't we just need a flat-out chaotic approach? Because they were created in a sense with chaos and no one asked too many questions.

I am somebody who would agree that you're going to have to be far more tolerant of mistakes. Than we have been. And I think a lot of people that I worked with agreed with that. A lot of them are now in reaction to something that is just far more chaotic than it seems to need to be. And again, where the cuts are happening in the wrong places sometimes. But that's part of chaos, right?

You rebuild later on. Well, and to me, it doesn't feel that chaotic at all by global historical standards. Sure. We're in a remarkably orderly time, even with you know who as president. No. No, that's fair.

And I think the other thing that we should be thinking about— so everyone is very concerned, and I think understandably so, for instance, some of the younger and less experienced members of the DOJ team getting access to systems. And I think we should not measure government by the degree to which it can fail, but more by how quickly it can recover. And we haven't seen that yet. We have— I mean, we haven't actually yet seen sort of, you know, Social Security checks stopping going out, for instance. That very well could happen. What we should be looking for is, can it recover quickly if that happens? And under what circumstances? And are the— you know, one thing I worry about is there are people with specialized knowledge that will be very hard to pull back in those times. But I think we've always had too much of a focus on get it perfect the first time rather than be tolerant of failures, but main— reduce the scope of those failures. And that's what the whole Lean Startup movement was about. Like, when I came here in 2013 to help start USDS, we were talking about Lean Startup methodologies, and those really are like, figure out, like, fail fast, fail quickly, but fail on a very small scale. So you're not hurting a lot of people. And then by the time you've rolled it out to millions and millions of people, you've worked out those bugs. But the way government does it, they try to get, you know, serve everybody on the first day. They haven't actually tested these things and they're trying to do things at scale without having those little failures early on to like make a robust, resilient and scalable system. So in some ways, I'm worried in this moment when we're talking about all this chaos and people reacting to the chaos of all these cuts and people getting access to systems, that they're losing that view we had, you know, in 2013, 2014, which is let's actually be tolerant of failure. Let's in fact intentionally do things that will help us learn what those failures are. And we're just going to have a better government then. But if we just go into full reaction mode, we're going to lose all of that.

If someone says when it comes to regulatory reform, Accountability is not the solution, it's the problem. Do you agree? Accountability is not the solution, it's the problem. You put in accountability, everything has to be measured, everything has to have a process. Yes. Judicial review. So should we have less accountability in government?

In a certain sense, I would agree with that. I don't think in sort of an absolute sense. I think the way that we structure accountability is very flawed. And I think we are holding public servants essentially accountable to metrics that are not proxies for real outcomes that people care about. And when you have a very high focus on accountability that is really about fidelity to procedure and process instead of to the actual outcome, that's not accountability.

But outcomes are heterogeneous. They're tricky. They're long-term. When you ask people to measure, you end up with a lot more emphasis on process than you want. So maybe accountability is the problem. And to say, well, accountability for outcomes, that's just going to morph into accountability for process. That's what I observe even in private companies, big, successful, profitable private companies that we're all familiar with. They have the same problem, as I'm sure you know.

That they're held accountable to the— well, that measure—

well, yeah, that There's too far too much process, bureaucracy delays. They're slow. Look at construction productivity in the United States, right? It's terrible. It's declined. Yeah.

No, I would agree with that.

And that's the private sector.

I think one of the issues, though, is that there is more accountability to process in government than in the private sector, though, I believe.

More in government.

More in government. Yes, sir. In the private sector, if you don't get the outcomes, you are unlikely to succeed financially. But there's a profit. Clear goal.

And in government, it's not the same kind of outcome. So it collapses more into process.

Yes, it collapses more into process. Absolutely. I think also you have, what is it, Goodhart's Law that says once a target— Yes. Once a measure becomes a target, it ceases to be useful. And you see that everywhere in government. And I think that's part of what I mean about sort of the new public management and the reinventing government in the '90s was highly reliant on like, let's set a goal and follow that goal. There can be real value in that. I'm not discrediting it entirely. But you do have that kind of erosion of the value of those targets as people try to meet the target without actually meeting the outcome that was intended. And I think that a more sort of digital transformation approach that is instead able to change over time more quickly and say, wait, this target is no longer helping us get where we wanted to go. We've got to iterate on that. Let's change it. Can really, I think, get us out of that sort of industrial era thinking. How should we prepare government for AGI coming?

That is such a good question. I think I'm going to ask everyone this question for the next 2 years. But does it just mean firing a lot of people?

I don't know. I do see that. I mean, take social services, for example. We have a lot of people in government doing pretty low-value work. It's essentially paperwork. Like, all of our social safety net is paperwork. It's lots and lots of different kinds of paperwork at lots of different levels. And people who fundamentally just want to help another human being just do paperwork and IT systems and compliance. And, you know, what a social safety net is supposed to be is about actually fundamentally helping someone get through a tough time. And that requires humans, I think. So I think to the degree that it can, rather than just displace people, though it probably will displace workers, move people to higher-value work that helps people more. That is what we should be trying to craft it to do.

So we should shift humans more into what kind of jobs?

So imagine you're an eligibility worker for SNAP, and imagine you got to actually talk to your clients and listen to them.

And— But the AGI can do that at scale, right? And it will make better judgments than the human.

I understand that. And I think that I'm trying to get my head around AGI as opposed to sort of the chatbot that I think doesn't actually— Not yet. —fulfill that at the moment.

Soon, especially when government trains it on the data government has.

Yeah. And maybe there won't be any difference to somebody to know that it's an AGI versus a human. Right now, I think people want to know it's a human.

Well, people want a lot of things. I would like personal chauffeur service. But if you're getting money and an AGI hears your case and it's more fair than the human, I think if it's indeed more fair, people should just have to deal with that.

Yeah.

And I agree they won't like it.

And maybe AGI may be able to do this as well, but I'm not talking about just adjudicating the case. I'm talking about saying, things that I think our current chatbots wouldn't say, which is, you know, displaying empathy.

Oh, they're great at empathy. It's hard to get them to be pissed off at you.

Referring them, you know, to other services. I get that, you know, we could go that direction too. But I still think that there is higher value work that people can do. I think AGI being closer than we thought it was, say, even last week is— yes, it's disrupting my thinking on this too. But I also think that there is work that government doesn't do today at all, right? Like, if we can do a lot of these things, we're not going to magically print money and have more safety net dollars for people. But what we care about really is addressing the needs, right? And there's going to be, I think, a long time where it's going to take us Well, to figure out what are those needs that could be addressed by AGI? How do we actually get that technology to address those needs accurately? Test in, again, with limited risk, et cetera, not just roll stuff out willy-nilly. And then I think we get to what were we not doing that we can do?

Should we have public sector unions? Because they're going to fight AGI, right? Should we consider just getting rid of all public sector unions? They oppose reform, right? They don't really like accountability in the proper method. Their goal in part is to protect jobs.

I don't know if we should or could in the current reality. I do see that they are often, I think, advocating for policies that are actually not in the interests of their members. Sure. So something's gone wrong there that needs to go right if we're going to continue to have them. For me, one example of that is they're so protective of underperformers. And underperformers are a very small part of the civil service. I'm critical of the civil service in the sense that we just talked about, where there's too many people whose jobs are to stop things. But that doesn't mean that person is an underperformer. They may be doing a fantastic job at what they were told to do, which is don't let anything happen, or like, you know, make sure you're using these guardrails to the maximum extent, for instance. But there are actual underperformers too. And while there are not that many of them, they really drive the rest of the public service like crazy. And it's crazy to me that the unions don't say, actually, for the good of most public servants, let's make it easier to remove these underperforming employees.

But there's some asymmetric information embedded in the problem, right? So the people who are losing their jobs, they scream loudly. The other people don't have to pay the cost of that. And so they're willing to let the union be protectionist. But isn't that more or less intrinsic?

I think the other people are paying the cost.

Very indirectly, right? Asymmetric information. Those costs are very widely distributed. Whereas if you're losing your job for being an underperformer, you're screaming.

Yeah. That's right. But the unions, if they were performing properly to actually advance the needs and interests of all of their members, would take a different stance, I believe.

But, you know, any institution, you can say that, well, if the people were doing their job properly, but you need to design the institutions for the fact that they'll pursue their own self-interest. So we're in a university. There's university administration.— some things they do well, not well. But it's never an answer to a problem to say they should do their job properly. That's fair.

I don't know enough about unions to know how they would work if they were working properly to me. I do think that, especially in government— I mean, Philip Howard's made a bunch of points about this that I've warmed to over the years. There is a different set of circumstances for public sector unions than in the private sector. And I think that's part of the problem is that it's sort of been directly translated into the public sector. Philip Howard is much more articulate on this than I am. But I would, I think maybe slightly counter to Philip, say that there is a higher need to protect civil servants from firing for a certain set of reasons. We do want an independent civil service that can say, actually, the science says X, right? Now, an administration may choose not to follow that science, but you want to retain that person and their, say, scientific expertise or their bureaucratic expertise and not create a situation in which the entire civil service gets wiped out every time you get a new administration. That sort of seems to be where we're headed. I don't necessarily see public sector unions succeeding in protecting against that. But to the extent that they exist to do that, I do value that. I do think we need independence.

If we think about the United Kingdom, the country, their government, it seems quite stuck. Do you have an interpretation of how that's happened? They haven't had productivity growth since the great financial crisis. They have a big fiscal problem. It's hard to see signs of progress. The government appears paralyzed. They've had many different prime ministers. What are they doing wrong? I don't know.

It's not really my topic. I actually get a lot of inspiration from the UK because they started the Government Digital Service there. Which was the inspiration for USDS. And in fact, a lot of the work they did at Code for America also just picked up on their principles and their values. But I don't think that has been a salve for— yes, stagnation economically, that is a little bit out of my territory. But when I look at what's written about the problems in the UK, I see a lot of that here in the US as well.

But take Singapore. I've been to Singapore many times. It's arguably the best functioning government in the world. But when you look at Singapore, does that in any way give you pause? Do you just think, well, we should be more like Singapore, or you see what has been required to make Singapore work and it puts you off a bit? And I'm not talking about like how democratic is it or not. I know that's one set of issues, but just the talent drain from the private sector. There's a certain conformity in Singapore. Some of that comes from size. But when you have the super cooperative, well-disciplined civil service, that does come with some costs. And how well have they really done when you consider all possible factors?

Look, I think there are obvious benefits to the system that they've set up, the value that they've placed on civil servants. Yes, the sort of technocratic, like extreme competence that they display. You could also say the same thing for Estonia. There's always this desire to say the US should be like X or Y or Z. And the reality is, yes, you have a different culture, right? So would we be as orderly about it? No. Where it's just, we have a different culture. I think another thing people don't recognize when they're trying to make comparisons between our, sort of the way we run our government and any other country is that we have policy complexity and sort of legacy that just goes back further than they do. We didn't clean up for 250 years. We've added and very rarely subtracted. So if you look at a country like Estonia, their problem is not our problem. Like, they have a relatively greenfield environment in which to create the digital infrastructure, but also to implement in a policy framework. You know, you look at something like unemployment insurance, and it doesn't really work here. It still doesn't work. Like, we, you know, threw our hands up and screamed during the pandemic and said we would fix it, and we have not fixed it at all. It's going to be a disaster if we have another downturn, you know, another spike in unemployment insurance claims. And part of that is because it goes back to, you know, the 1935 Social Security Act. So you have 90-something years now of just adding, adding, adding, adding. You've got thousands, literally thousands of pages of active regulations that make something that should be a very simple program, like I get, you know, some portion of my salary under some circumstances, you know, for some number of weeks. It just shouldn't be that hard. But 'cause we have let the policy accumulate and accrete over decades. The challenge isn't digitization. You can digitize something. The challenge is creating a system that is resilient and scalable with that much policy craft. And I think both Estonia and Singapore have the advantage of just having sort of you know, less legacy than they do now. And they're small, right? And they're small. And they don't have federalism.

Sure. Let me give you a weird hypothesis. Tell me what you think of it. If someone argued, well, the US has the highest state capacity in the whole world because the number one feature of state capacity is that you can convince enough of your citizens to fight for your country and die for it. The US does pretty well for that. But to do well on that, your country in some ways has to be nutty, a bit militaristic, or maybe you need gun culture or diverse regions. A lot of the military comes from the South, which is a more conservative part of the country and so on. And that a lot of the other problems we have actually stem from our high state capacity because we've decided to emphasize talking people into fighting for their country, being willing to die for it. Maybe Israel is pretty good at that too. And that's just another way of diagnosing the problem that we're doing way better than we think. We prefer to be big and really strong and militarized, and the rest is collateral damage, but we live with it.

Well, I'm not sure I agree with the definition of state capacity. I could see that being part of it, but I'd also not accept the notion that because we can convince people to go in the military, though, you know, we're having terrible recruitment problems in the military. We're doing Poor America.

But compared to other countries, I mean, we're way ahead of just about anyone.

But that is not translating into an ability to deter our adversaries. Like, we lose our war games with China.

But we protect more of the world than maybe any country ever has before. We did. I would say. Well, it's somewhat less now. Somewhat less now. But it's still at an all-time world historical peak. And US soldiers are incredible fighters. You could have them in Afghanistan, Iraq, maybe very complex missions. Connection to life back home is highly unclear. And they fight incredibly well. And maybe that's at least half of state capacity that matters. And we're masters of it.

One thing which I agree with is that the separation, I think, between the parts of our country that are involved in the military and the rest of the country is pretty damaging. I think that's a, that's a really, it's a big problem. I'd also, I think, somewhat agree with the something behind that question, which is that when we have too much of something, we get bad at it in a certain way. Like, I think very often people, when I talk about capacity of the state to do anything, and my background spans from, yeah, food stamps to military readiness, people often go to resources. Capacity is money essentially, and people. And I think very often it's in fact the situations where there is less money and fewer people where you get services delivered better. And I'll give you an example of that. When we started working on food stamps in, I think it was 2012 in California, California, the second lowest participation rate in supplemental nutrition assistance in the country. The only state with a lower rate was Wyoming. Maine, on the other hand, was, I can't remember, somewhere in the top 5 or 10. Maine spent orders of magnitude less on the systems to enroll people. I mean, we had tons of money, you know, and you think about California being, you know, tech-forward, it's where Silicon Valley is, you would think that like the online enrollment system would be great. But no, because we could put a ton of money into it, we way overspecced it, gathered way too many requirements, you know, gave it to expensive vendors, and we ended up with an enrollment system that, you know, had 212 questions. You know, they just put everything into this. It took, you know, far longer to fill out that form on a computer than you could, you know, spend on a computer at the library, which is where a lot of people filled food stamps applications. So like when the computer times out after 30 minutes, you haven't finished your application. There's a lot of these like very little frictiony ways that we sort of kept people out of the system, which starts to account for that really low enrollment rate. And it wasn't lack of resources, it was too many resources.

As you know, government does not typically have its own AI, not in the United States. As AI becomes more important, will that lead to a de facto mass privatization through procurement? So the government will contract out services to the major AI companies. That will be done privately. Procurement will become far more important. The government itself, it won't quite vanish, but will be fundamentally different than what we've all grown up with. Is that just going to happen?

You know, I think that there is a world in which you get a government with enough technical competence to procure AI and manage AI. Well. The problem with our procurement today is that we buy software the way we buy pencils and cars, right? We've bought a thing. AI isn't like that. And actually, software isn't like that. This is a lesson we should have learned in the internet era. Now we need to learn even better in this AI era, is that you can't spend 5 years, you know, developing requirements and sending it out for bid and then having it come back and then say, okay, I bought this thing. You need the internal competence to do that process. This far better, far faster, and then to actually own and run that software, right? It has to be yours. And I think we could do that. And in some ways, you know, there are rumors that Doge is, you know, actually pro-insourcing more tech talent in government. I don't— you're not hearing about that now, and it may just be a rumor and it may not be true. So, you know, don't, don't quote me on this, but I certainly think that folks in Musk's world came in and looked at government and said, how do you even operate with so little technical competence in-house? That's crazy. And it is crazy. They're right about that. So AI, yes, will either end up, I think, even further privatizing, not just the software, but the whole operations, as the software and the operations are increasingly melted. Again, this is not new. It's— or we'll be forced finally to gain the internal competence that we have always needed and start to do this a little bit better.

Let's say you're the nation of Peru and AI becomes more important. You run a lot of your education system through American AI, a lot of your military, your cybersecurity, processing of social services. At the end of all that, after, I don't know, 15 years, are you even a sovereign nation still? Are we seeing this fundamental shift in what a government is? If the US government does something, at least it's US AI. Other than US and China, it's far from clear how many countries will have their own AI. What's going to become of Peru?

I mean, I can't answer that question, but it's an excellent question. But I think it highlights the degree to which you are going to see I think you may see countries realize this and take just more drastic steps to maintain their sovereignty, which, you know, I don't know how they're going to do that.

But they could just stay backward, right? They could just not use the AI. And probably some will make that decision. I think some will. But that's a huge cost. Yeah. And then there's countries like France where maybe they can have their own AI, but they haven't succeeded so far. But they could at least make a run at it. Like, would you tell France, yeah, make sure you develop your own AI so you're controlling your own government?

I mean, question back to you, is there some way in which the market fixes this? Because there's a market for AI that, you know, doesn't implicate their sovereignty.

I worry that the economies of scale are in conflict with the number of nations in the world today, and we're going to see a very radical shaking out. That will go fully unadvertised. But at the end of it all, geopolitics will be very different. And I don't know that France can build an AI good enough, actually. They had Mistral. They still do. It's not making money. People want to buy it. Macron says we're going to basically prop it up. But you can't just prop it up. It has to be good enough to work and be competitive. I don't know.

But won't there be companies that have enough AI expertise that they're able to cater to the needs of a sovereign nation and say, I'll, you know, I'll do this for you in the way that fits your needs?

Well, it could be France contracts with Anthropic, and Anthropic in a sense runs the French government. And maybe we trust Anthropic. I'm not I'm not necessarily opposed to that by any means, but again, it's a very different world.

It is a very different world and it is very hard to game this out.

And America and China for sure will have their own AIs and maybe those countries just become much more powerful. And then you've got to decide, well, whose network are you in anyway?

Yeah, this, this, this is all coming at us very, very fast. And I think the point's been made that this is the first time that such a transformational technology hasn't fundamentally come out of US government resources. This is not one that you can ultimately trace back to DARPA or DOD. And so our government really is at— at least then, even if it was privatized and the value, so much of the value came out of the commercialization of, you know, DOD or DARPA investments, since that's not the case this time, they don't really even understand it. And I think even before you get to the questions about other sovereign nations, we have the question of the relationship between the companies and the US government.

Of course. And how we're going to sort that out. Not to mention federalism, right? We have plenty of governments in this country. What's your favorite novel about bureaucracy?

A novel about bureaucracy? A novel.

So I would say Kafka, a number of Kafka novels, but maybe that's too negative a view.

Can I use a TV show?

You can use a TV show.

Well, there's so many, but I got it just for a positive vision. That's— I gotta call out Parks and Rec. Parks and Rec.

And what does Parks and Rec show you? I've never seen the show.

You've never seen the show? It just indulges in that banality of things like, you know, public meetings that are dominated by, you know, the loudest voice, and, you know, how ultimately, you know, humor and you know, just communities sort of get through it. But it's a very funny show. Yeah, it often highlights just how mind-numbing bureaucracy can be, but also, you know, how people end up making it sort of beautiful and funny.

The Kurosawa movie Ikiru is very good if you haven't seen it. Yes. I talked Ezra into watching it.

Isn't that the one that's been remade in the UK recently?

I'm not sure. Maybe.

Oh, I'm going to forget the name of this movie, but it is excellent. And I believe it's the— this is the British version of that about a civil servant near the end of his life. Yes, exactly. That's it. And the playground. Correct. It's coming back.

Yeah, I've heard it.

Okay, so I've seen the British version and I should have named that. Okay. Excellent, excellent movie.

That would be my favorite movie about bureaucracy.

Well, maybe we put in the show notes the actual name of the British version.

We will put it in the show notes. Yes. Do you still do gaming? No. How do you think about not doing gaming? This previous part of your life, like, I don't play chess anymore and I can't, I sort of feel I can't, I shouldn't, I won't.

You know, gaming takes up a lot of time. Like, it's hard to be a casual gamer. I mean, like, my gaming now is like, I do the New York Times games like every day, you know, Wordle and Scrabble.

So you are still a gamer.

Yeah, yeah. But I mean, since you're referring to me working in the video game industry, running the Game Developers Conference for many years of my life, Yeah, you can't play Doom for like 5 minutes on the bus, right? Yeah. So many, many hours have gone into many PlayStation games and Xbox games in an earlier part of my life, and it's just not part of my life anymore.

What do you think you learned from that being such a big part of your life?

Yeah, that's a good question. I, you know, I went from working in essentially little bureaucracies. So my first job out of college, I worked in a child welfare agency and then at a sort of a healthcare think tank. And then I went away for a year and I came back and I got this job in tech. And it was amazing to me how, you know, creative game developers are. They have this sort of completely unconstrained ability to just imagine entirely different worlds and of —of course, you know, like the economies of those worlds and like build the systems that simulate all sorts of aspects of our lives. And then we sort of moved from that into this Web 2.0 era and then kind of went back into government. And I think what I kept with me is that you can imagine fundamentally different things. This is an imagination that is so powerful. And the quote I chose for the book, Recoding America, is a David Graeber quote. You may need to open it to remind me exactly what it says. Sorry, I will pull it up. "The ultimate hidden truth of the world is that it is something that we make and could just as easily make differently." And I think I got that from the designers of Ultima Online. And my first husband was on the game Spore. It was a Will Wright game. You can have a whole different world. And once you realize that, all these constraints in our daily lives, you realize they're manufactured. I think we have a little bit of that right now. This is going to sound extremely boring compared to what I just said, but since we've been touching on federalism a little bit, you know, you talk about like the federal civil service and people are like, it's a thing, like it can't change. It's— it hasn't changed since 1978, really. I mean, little changes here and there, but we don't just have the federal civil service. We have 50 civil services in states that are wildly different. A lot of— not a lot— some states have at-will employment. We can't imagine that in the federal civil service. And I'm not advocating for that, but I am advocating for us getting out of the trap of like, this is the only way it can be, when clearly, like, even in our own country, there are 50 different models that we could pull from.

Which do you think is our best-run state?

I really can't answer that, but I'm very interested in Pennsylvania under Shapiro. Great paper out by Don Moynihan recently about the work that they've done to just speed things like permits. You know, these— maybe it's boring stuff, but it matters to people.

And— But their schools aren't that great, right? Their roads are fine. But he can't have changed that much, right? The state as a whole doesn't seem to me exceptionally well governed, maybe in the middle somewhere.

I'm more talking about where I see interesting change happening that others could follow. Colorado as well. I think what poll is—

that seems to me a well-run state overall.

Yeah. Utah. Yeah. I don't know enough about Utah to say that with great certainty, but yes, I think I could see some of that. I mean, there are states that I think have either tackled that procedural accretion or just not let as much procedure accrete in the first place that are going to have an easier time. California, sadly, is not one of those where I used to live. Tell me what you think of Virginia since we're in the state of Virginia and I live here now.

It used to be one of the top few. It's declined somewhat, but still well above average. It is genuine two-party competition, which I think is a good thing. Yes. It has this cash cow, the Pentagon, which creates jobs and revenue, yet no one in state government can touch it. You can't grab from it. You can't take corruption from it. It's not like Alaska oil revenue. So that's a huge permanent plus that is fully protected. So that's a great natural endowment. And again, it's one of the best states to live in, I would say.

Well, I chose well. Yes.

Where did you grow up?

I was born in a naval base in Maryland, but very quickly moved to Austin, Texas, and then moved to New York when I was in 4th grade. So I grew up in New York City. New York City. New York City.

How well run is New York City today, in your opinion?

Well, it's certainly cleaner and safer than it was in the '80s when I was growing up.

But is that because it's run better or just because it's richer?

Probably just because it's richer. And I don't, you know, to be honest, like, it was smelly and I got mugged a couple times, but I don't regret growing up under those circumstances. I think it made me who I am and I have a fond affection for New York in the '80s, even though it was kind of a mess.

I was there often in the '80s. It was a great time.

And I went to a public school. I went to Bronx High School of Science.

I grew up in New Jersey, right across the river. So, but say they call you into New York City. We'll do whatever you tell us. What do you tell them to do?

Okay. I have a great answer for this because we just wrote a paper. If you want your government to be able to achieve its goals, and I think New York's got some good goals that they're not achieving, you need to be able to hire the right people and fire the wrong ones. So you've got to look at that civil service system. New York in particular has a very bad way of hiring. They still are stuck in this artifacts from the Moses era where you do these civil service exams and they only happen at certain times of the year and all this stuff that just makes hiring really, really hard. You got to fix that. Second, reduce procedural bloat. That's sort of my favorite one these days. These go together. The procedural bloat is what creates that ineffective hiring system. The third thing is you have to invest in digital infrastructure. Lots to say about that. And then the fourth thing is that you have to close that feedback loop between policy and implementation. So what that would look like in New York City is— well, I guess it's the same that it would look like at a lot of different places— is that your relationship between law and policymakers and those that implement those laws has to change. And you've got to sort of stop firing and forgetting as you make policy and figure out how to build the affordances for that. Ongoing iterative feedback.

And to close, tell us about your work with Niskanen.

I am loving working with Niskanen. They were the only think tank that had anything on state capacity when I was finishing my book. And I hope I'm helping build a team there that can do much more to articulate how any government, especially in the US at any level, can increase its ability to achieve its policy goals. We have a wonderful team there led by Ben Bain. And we're getting a lot of interest from people that never used to pay attention to things like technical infrastructure in the agencies, even from Hill staffers and members of Congress. So I'm very hopeful that we'll start to make a real impact— well, not just in Congress, but working with states and cities around the country to sort of tackle these four pillars I just talked about.

Jennifer Pahlka, thank you very much. Thank you.

This was a lot of fun.

Thanks for listening to Conversations with Tyler. You can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. If you like this podcast, please consider giving us a rating and leaving a review. This helps other listeners find the show. On Twitter, I'm @TylerCowen, and the show is @CowenConvoz. Until next time. Please keep listening and learning.