TBPN | Tyler Cowen on AI, Economic Growth, and America's Debt Problem

How should, you know, nations think about the level of indebtedness that's appropriate?

United States is a special nation and we can get away with more debt. About half of our economy right now is just totally perpetually sluggish.

Tyler, how are you doing? Good to see you.

Welcome back.

I'm fine. Good to see you. Happy New Year. Congratulations on the Vanity Fair piece. I loved it.

Thank you so much. I'm so glad that you were able to check that out. It was a lot of fun. Very funny getting dressed up all fancy.

When are you going to be in Vanity Fair?

They really have not been doing their work.

I think I would love to see a— a Vanity Fair photo shoot with you as the subject. I think it would be fantastic. We should make it happen.

A style icon, right?

Yes, definitely. Well, I mean, they really should do more reflecting on Conversations with Tyler. The team's grown a lot. Did you just hit a recent milestone? I saw that there was an event. Can you just get me up to speed on the organization and how long you've been working on the show?

Conversations with Tyler is now 10 years. It's a few hundred episodes.

Great success.

And last year, we did almost An episode a week.

An episode a week. That is fantastic. Well, it's one of my favorite shows. It always has been for— I feel like I might have been a decade-long listener. Maybe not the first year, but I got on early, and I've been very pleased the whole time. But anyway, thank you so much for joining us. We first wanted to have you reflect on this point from Michael Burry that artificial intelligence might displace more blue-collar jobs than people are expecting because you can now take a picture of your toilet and learn how to do plumbing, or you can take a picture of an electrical panel and you can do your own electrical work. Does that resonate with you at all? Do you think that there might be some substitutive effect in the DIY community that might have an effect on those trade jobs that previously have been, you know, rumored to be very AI resistant?

Well, that's true, but there's still a net job boost coming in those areas. Because there'll be all these new projects, new data centers, new sources of electricity, and they need plumbers. They're not all going to use ChatGPT to take the photo and try to figure out how to fix the thing under the water, whatever.

Yeah.

So the house call plumber, maybe that goes down by 20%, but so many other new things happening, terraforming the earth, whatever, other countries, Africa developing, those jobs will be doing great.

Yeah. How are you?

That's it. I agree with his example.

Sure. Sure. How are you tracking the overall data center boom? There's been this narrative that data centers are the only thing that are propping up the US economy. It feels like we hear about new data center projects every day. The numbers are getting bigger. There's a new zero every time there's a new press release. And at the same time, the overall power generation, the big, big like US national number is not yet started to move, but it feels like this might be the year where we see more electricity generated in America than before. How are you tracking the overall impact of, of AI on the broader economy?

We need to do much better. But I think it's wrong to feel that without the data centers, the economy would collapse. Those resources would be used to do something else and different. It might be less risky. It might be more immediate. Actually, consumption would be higher in the short run if not for AI in the data centers. So what we're doing is investing longer term, taking some risk, boosting medium-term productivity. I'm all for doing that. But again, without that, we'd make more toys or more hot dogs or more something else. So no big deal.

How often are you listening to AI music?

I try it reasonably often, but I don't yet listen to it because I want to. It gets closer every time. I would say within 2 years, I think I'll be listening to it on purpose. Right now it's experiment and tracking the field, not for enjoyment.

Sure. Do you find the, the act of playing the game of trying to generate a song that you like more entertaining than the actual resulting music?

No, I don't find either entertaining. For me, it's a pain. You know, I have a very large collection of albums and compact discs. I can hear literally the world's best music at my fingertips when I want to. And the other is a distraction. I do it as my own investment in seeing where things are headed.

Do you think there's—

I had someone share an interesting anecdote recently, which is that a prof— you know, one example, a professional singer who made— who's sort of top of their field within country music. They made something like $400,000 last year, basically going into the studio and recording sample tracks. You know, basically a songwriter would make a song, they would sing it, and then they'd go out, or this person would sing it, and then it would get pitched to other artists to be made into like a hit, right? And actually be produced. And his work literally went to zero because of Suno this year. And he doesn't have the ability, nobody hires like a backup singer to go on tour. Like, you can be like a studio guitarist, and get live jobs. And so their work went from— because now people can just prompt with Suno and say, hey, make— use these lyrics, apply it to this track, and make it sound like this artist. And so that's just been kind of an interesting space. And I was talking to another friend yesterday that says basically everybody is using Suno at scale, and nobody in the— very few people in the industry are willing to actually talk about it.

Interesting.

So, you know, I played some for my wife recently and she said, I hate it. And I said, why did you hate it? And she said, I hated it because it was good. That's where we're at right now.

That is, that is.

But yeah, it's notable too that it's maybe not— it's not great for like the end listening stage yet, but it's totally good enough for like the sampling, for the professional kind of workflow side of things.

Yeah, yeah. Do you think there'll be sort of a a legal reckoning, figuring out how all the royalties flow through these models and saying that, okay, there's 1% Taylor Swift in this generative sample and 2% someone else. And then all of the money from like, sort of like a more complex rev share that might flow through these generative models.

I don't know if that's practical. I mean, the Beatles took from Chuck Berry, Buddy Holly, Carl Perkins, everyone.

Yeah.

Without paying, and the world went on, even whether or not it was fair. And I think we're going to redo that experiment. Yeah, I don't think it will be all that different. Most musicians don't make from recording anyway.

Yeah, that makes sense. Uh, have you tried Claude Code?

No, I have not. I've been meaning to, but I've been traveling, and I'm finally back home, and I will be trying it. Uh, everyone tells me it's amazing.

Yes, yes, people are, uh, maybe one, one click behind you. I remember you, you declared '03 AG.

AGI.

And now people were maybe hesitant to call the game there, but many more people have jumped in with affirmations that Claude Code is AGI because it can do so much more when it actually has access to your full computer. But it is a little bit more cumbersome to get set up.

How did you process Powell's video yesterday?

Oh yeah.

Well, it was terrible what Trump did. Fortunately, most of the markets did not react very much, but gold and silver went crazy. To me, that's a sign that the dollar is much less of a safe haven. And Trump is going a bit, you know, what I call the Captain Quee route, and he's just not reliable. And that's very bad. But I don't think it's really going to change inflation or interest rates very much.

Is that just because Powell will hold steady?

No, because we already wrecked the independence of the Fed, which I'm not happy about. But that's the ugly little truth behind this story. Yeah, that's why it's not been worse what Trump did. Is because it was already wrecked.

Why is Fed independence important?

Sometimes the central bank needs to do things that are not politically popular, and they need that shield to do it. Mm-hmm. But the basic problem is our debt and deficits are so high that over time we will monetize them to some extent and have higher inflation because we prefer that over higher taxes, no matter what we might say. And so the Fed independence is taken away through that mechanism. Apart from whatever Trump said yesterday. Oh, so just like I'm not telling you not to worry, but I'm telling you, you should have been worried to begin with.

Black Bill. So, yeah, you're saying that just by running high deficits, that reduces Fed independence.

And that's an old story, but it's now closer to a tipping point. And Trump is making that much worse. So I would say his actions On the fiscal front, do more to hurt the Fed's independence than his words, but they're both bad.

Can you walk me through your mental model for thinking through what the right level of debt for a nation is? Some of the numbers get so big, it's trillions of dollars. It feels very abstractly bad. But then if you look at it more like a mortgage relative to, you know, debt-to-income, it feels maybe more reasonable, like How should, you know, nations think about the level of indebtedness that's appropriate?

United States is a special nation and we can get away with more debt, which is great for our living standards, but it's bad for our political responsibility because our leaders all know we can get away with more debt. I think what we'll need to do at some point is have half a dozen years of something like 7% inflation. Get the debt down. It won't at all solve the problem. The problem will never go away, but it will give us some breathing room. Well, then lower inflation, maybe have a recession and start all over again. And it's highly unpleasant, and a lot of people will be thrown out of work and living standards will be lower. But we've already spent that money. We can't default. And that's facing us over the course of the next 10 to 15 years. But I don't think the world will end.

Is America's position in AI the biggest counterbalance to all that, like, sort of negative view?

It is. So if AI would help our economy grow 1 percentage point more a year, we could just afford the whole thing and we wouldn't even need to inflate.

Yeah.

Now, that would not be my best prediction, but it's not impossible that it could happen. And I've been predicting half a percentage point a year, which still puts us right on that border of can we, can't we? We don't know. I wouldn't want to stake the whole house on that. But again, there's some chance we'll squeak by due to AI and assorted productivity gains.

Yeah, but why not 1% growth? Why not 2% growth? These tools, they feel amazing. It feels like—

or triple digit as Elon Musk.

Sure, sure. But, but it does. It does. You use these tools and it feels like you're more productive.

And there's some studies that say, oh, maybe you can generate documents that don't get read.

Basically, is that what's going on? Like, like, like what is holding back AI from actually moving the needle on economic growth significantly?

Well, first, about half of our economy right now is just totally perpetually sluggish. Look at government. Look at most of the nonprofit sector. Look at higher education, parts of our healthcare sector. Add that up, you're at 50% of GDP. To some extent, they use AI already, but in pretty trivial ways like, oh, it saves people some time. They don't work as hard. They take more leisure. They hang out more at the water cooler. That's fine. It's fun, but it's not going to get rid of our debt problem. And then you have some startups, you have programming, you have the dynamic parts of our biomedical sector that are already using AI a lot. But the more they use AI, the more efficient they become and the more the inefficient parts of your economy are left over. And it's just really hard to grow much faster. Interesting. Unless you're like China playing catch-up.

Yeah.

Yeah. So the number's hard to estimate, but GDP is a huge mound of stuff, most of which is produced by people who are not neither white nor blackpilled on AI. Sure. And it will take things a long time to change.

Yeah. How do you think about the legacy of Thomas Piketty and the, uh, the new, uh, debate over Piketty in the 22nd century? I believe, was it 21st century, 22nd century that Dwarkesh wrote about?

21st century. It's in the future.

Talking about economic inequality, increasing gains to capital and increasing returns to capital.

Dworkin did 22nd century. Piketty did 21st.

He's too focused on the labor share. I think real wages will go up a lot. If real wages go up, workers are happy. They might be upset that Elon or someone else is a trillionaire,. But I don't think that will drive our politics. I think if real wages go up, we'll be fine. People will feel good. So my view is very different.

Wait, but, uh, I mean, real wages will go up, they'll feel good, but then they'll also be upset that there's now multiple trillionaires walking around. And so won't those emotions actually drive the politics? It feels like they need to do it.

I don't think people are that envious. They're envious about the people they went to high school with or maybe their brother and Most people are not that envious about Elon or Bill Gates or whomever else.

Yeah, yeah.

Envy is local for the most part.

Yeah, I was trying to—

Maybe you two envy each other, right?

Yeah, maybe.

You envy, you know, the sheikh, the richest sheikh in Saudi Arabia. I bet you don't even think about him that much.

He does have a great car collection.

Or if you do, you think about his role in the AI world, which is fine.

Totally.

But you're not upset that he gets however much hummus and you have less hummus than he does.

It's all about the hummus. Oh yeah, I was doing a thought experiment.

Don't you think envy is a factor in the current wealth tax, uh, sort of discourse.

California is a crazy state, but I think it's just more a money grab than envy, uh, and I think there's a very good chance they ruin the greatest engine of wealth creation maybe in human history. So I, I hope it fails, and soon. Even the risk of it, as you know, is inducing many people to leave or get ready to leave.

Yeah. What's the, uh, what's the retrospective view on the Laffer curve in light of what's happening in California today?

Well, at some tax rate, the Laffer curve is true. We're not usually in that range, but California is playing with fire and experimenting with that range.

Mm-hmm.

So I dearly hope they don't do it for their own sake. It'd be great for Austin, great for Miami, maybe good for New York. Yeah.

Yeah, interesting.

New York has a problem. What's your personal investing strategy?

Buy and hold. Focus on other things.

Investing is a— Yeah, but buy what? Because if you buy individual names, you can end up being wildly distracting and then you can't focus on the other things that are maybe more productive.

Diversified portfolio. I've thought now of buying some more gold just as a hedge. It's not that I think it will do well, but I see higher risk. And Bitcoin's not really a hedge. So maybe gold and silver. Again, I'm not saying it will make you rich. I'm just saying if everything else falls apart, you'll have something.

Yeah. Have you pushed any AI model, AI labs to develop taste or smell technology to help with the judging of food? It feels like culinary aspects are particularly AI durable or resistant. But I don't know if that's something that can be overcome with better technology.

Maybe it's cheaper for now just to use humans and have the AI record what the humans say and judge the food that way. That seems to work well. Judging food is not hard. It's one of the easiest problems for humans to solve. And humans will do it basically for free. How many food bloggers get paid? Well, they might get some free meals, whatever.

Sure.

But so many people do it for free. I just wouldn't put AI money into that for a long time. I'd work on almost every, every other problem first. You know, AmeriCorps, they wanted to hire some basketball analysts to make basketball commentary better by AI. I'd rather do that than food. It's harder. Sure, sure.

That's very funny. Um, What about this call for new aesthetics? What do you— do you believe that we're in some sort of local minima for the development of new aesthetics? Like, do you attribute this to just broad stagnation? What are the key sources that led us to this point where it feels like aesthetically we're stunted?

I think poor taste is a big problem around the world, but especially in America. You look at the older parts of San Francisco, the Victorian homes, they're beautiful.

Mm-hmm.

You look at the new buildings, I like some of them, but a lot of them are just awful. Or you look at parking structures, or you look at a new bank branch that goes up somewhere. Awful, awful, awful. Are we really so poor a society that we cannot afford to invest in more beauty? So Patrick and I are trying to get people to think more systematically. What can we do to try to make much more of our world just plain, flat, outright lovely. It absolutely can be done.

Is there an element of the financialization or, I don't know, like multi-party, like multi-pronged stakeholders that go into the development of a campus or a building now that maybe didn't exist 100 years ago where one person could just decide that they have some crazy vision for a building and they go and build Hearst Castle and it kind of destroys them over there life, but it's a singular decision. And now a company that builds a property or something, it doesn't have— there's not a single person that can be maybe authoritarian about the decision and the taste. Even if they have the taste, they get overruled by a committee of investors and stakeholders that all kind of put the kibosh on whatever they want to do.

That's one problem, and there are too many veto points. But if you go back earlier in time to the 1920s, say, You look at a neighborhood like Shaker Heights near Cleveland where the homes are being built by different families for the most part, they're still far more beautiful than homes being built today. So I don't think that's the sole main root of the issue. I think just bad taste is.

Do you think there's a technology angle where, uh, I— it feels like in the modern era a lot of young people lament high housing prices, but when I look at their choices, they'll often choose to live in a small studio apartment in a very dense city to be around other people. And when they go to their apartment, it is small. They don't have a library, but they have a Kindle. And they don't have a room to dance or something, but they have a TV where they can watch dancing. And they don't need a movie theater because they just have their phone to watch a show on or something like that. And so the— the number of places, you know, if you were rich, you know, a century ago, you needed a lot of spaces to do different activities. Now every activity can happen on the couch with a TV. Do you think that there's some effect there?

A black mirror view, John.

I don't know.

Well, if you do all that and you live in the West Village, you're surrounded by beautiful buildings.

Sure.

I'd like there to be more places in the country where you have that choice to be surrounded by beauty, that you don't just have to be all scrunched in that you can have reasonable living space and afford it, and what's around you looks nice. Again, people have done this in the past, even the distant past.

Yeah.

Even sometimes in medieval times. So to claim we can't do it now, it's simply a failure of will. There are laws that need to be changed, procedures that need to be changed. But I think the first step is just to wake people up and increase awareness.

Yeah, it feels like the YIMBY movement has a serious amount of energy behind it, and yet In my lifetime, I haven't seen that much movement on it. Maybe the ADU law in California is moving a little bit, but it still feels incredibly slow to get permits. Is there a dynamic where landowners, property owners are actually pushing back?

I don't even think that— my feeling is that the ADU, new support for ADUs doesn't actually really increase the housing supply because a lot of people are like, you'll let me put another structure on my property? Great.

I'm not—

doesn't mean I'm gonna— doesn't mean I'm gonna like suddenly like it's a new single-family home, right?

Yeah, I don't know. What do you think?

I think the YIMBY movement could go much further if it could promise it would boost housing and make neighborhoods prettier. Right now it's like, well, we're going to boost housing, costs will fall, that's wonderful, but maybe your neighborhood will get uglier. Sure. So you do it in Buffalo, you replace the old with the new. Yeah, the new is probably uglier. Yeah, I don't like that. I think YIMBY would have a lot more successes when it can promise beauty.

Yeah. Sorry. On the new aesthetics, I heard a critique or an idea that new aesthetics come from problem solving. And so it's less about the taste and opinion and will and more of there's a specific problem in society in a particular neighborhood. And then a design emerges to address that particular problem. The critique was that we won't get the new aesthetics until we identify particular problems, or maybe the bland aesthetics are solving for a particular set of problems that are in front of us. Does that resonate with you at all?

Not that much. I mean, I would say the problem is ugliness. If you look at older structures around the world, especially in Europe, but often in the US, they can be very ornate and have all kinds of flourishes and small details that are lovely. Those are not solving problems. They're put there because people think it will make the thing look better. So, it's not mainly about problem solving, but I'm not against solving problems, hardly.

Sure. Has there been any studies on the effect, the local effect of brutalist architecture? Like, does it make people like stay at the office longer or anything? Because I can actually— it's funny, I can appreciate brutalist architecture when it is surrounded by nature in a big way.

It stands out as contrast.

Yeah, it's a nice contrast. And it feels like a triumph of man in some way. But it oftentimes will make a neighborhood much less warm.

A shipping container home in a forest sometimes can look beautiful if you see it as a getaway.

Yeah, is there any— what made brutalist architecture be such a force in the world?

People don't like it once they live in it. Personally, as a tourist, I often enjoy it. I think it's interesting or sometimes creative, but it's not the model I would want to seek to spread through the world.

Yeah.

Because to live with it all the time, I think it wears thin on you. Why it ever got as far as it did, maybe England is the best example. The '60s, the '70s, It's just cheaper. They're in a hurry. Concrete is easy to manage. You have a bunch of planners who think they know better. And a lot of it was a big mistake. And, you know, I would say maybe England got the worst brutalism. Maybe parts of communist society got the best brutalism because they really needed the structures after World War II. But I don't think it should be our emphasis moving forward. We want to move away from that.

Do you think there's 8 billion people on Earth? Because there's a popular exact number of people on Earth. No, no, no. So, so tell me the back— no, the backstory is that there's, there's a growing conspiracy theory specifically on X this, this weekend, uh, movement of people that are just saying there's no possible way.

Google says it's 8.2 billion. Yes, but this is—

many people, many people are, are disagreeing with— and, and, and one of China's official data—

one of the arguments is that, is that in certain countries, China among them, there are incentives to to inflate your local population numbers so you get more resources from the state. You do that, you game theory that out across the entire country.

And there's a homegrown effort in China from people that are trying to prove this. And they're going and they're finding these cities that are just barren. It's a city that has, you know, 1,000 homes and 1,000 structures. That might just be overblown though.

But yeah.

Yeah, yeah, yeah. But there's like a homegrown movement to prove that China does not in fact have as many people as the official numbers claim.

What do you think?

I think there's a modest inflation there. If there's a betting market on this, I'd love to get in on it.

Okay, that's good.

I'm going to bet on, you know, $8 billion or more. That's my bet.

Yeah, yeah, yeah.

Okay. If it's $8.1 billion and not $8.2 billion, that I can believe.

Sure, sure. Yeah, that makes sense. Back on the architecture question and the aesthetic question, it feels like when people are dissatisfied with the current status quo of architecture or any aesthetic, there's often some sort of emotional response to return to a previous era, to just go back in time. And I'm wondering if you think that that is the solution. We need more Art Deco or mid-century or all sorts of different, I don't know, Gothic structures, or go back in time to just build new buildings that look like they're hundreds of years old. Or you think we actually need something that's entirely new and never before conceived?

Yeah, everyone draws from the past. Mostly I think we need freedom to experiment. It's not that I want everyone to follow my preferred path. The city where I love what they've done with new things and somewhat older things is Helsinki, Finland.

Hmm.

So personally, that's what I want. But do I want every city to look like Helsinki? Of course not. Or the modern parts of Copenhagen, I think are quite striking. And beautiful. So those would be my preferred directions. But most of all, it's about simply the ability to raise your hand and say, this is ugly, we don't want to do ugly things anymore. And we all look around for different ways of avoiding that outcome.

Yeah, I, I mean, the risk is that you just have nimbies that say, I think everything looks ugly and I don't want anything to build. I guess if you have the option to raise your hand and say that's ugly, we shouldn't build it.

I don't know.

Earlier on the show, uh, John was sharing how, uh, in 1947 there was a very, very small number of televisions in the US. 16,000. 16,000.

8 years later, it was 32 million.

So we were talking about how that was effectively a fast takeoff, a hardware fast takeoff. Do you think we can see something similar in robotics? There's a bunch of different exciting projects, a lot of which we've covered on the show. I saw a video of a robotic hand yesterday spinning a screw perfectly, and it looks like it's sped up like 10x, but it's actually in real time. So it's like spinning a screw 20 times faster than a human can.

Interesting.

But yeah, so I think people see the current state of humanoids today. They can do cool demos and other robotic form factors, but it doesn't quite feel like we could have a population of 100 million of these in 3 years. But who knows?

I don't think we yet see the killer app in the home. So if I ask myself, what do I want? If I could have a fully functioning robot that would be like a maid or a butler, that would be useful to me. But most things short of that don't quite seem worth it. Do I have a Roomba? No. I don't know. What if there's a mechanical vacuum cleaner that goes around on its own when I'm at work? Fine. It just doesn't seem that necessary. I do my own laundry. Uh, it's a welcome break sometimes. So I think we're still waiting to see the killer app.

Yeah. Any other technologies that you're excited about this year?

Well, everything in biomedical. I mean, even without AI, progress against cancer.

Yeah.

I think youngish people today will die of old age and maybe live to 97 or whenever the time is. And that's already in the cards. You don't even have to be a big AI optimist. I don't think it will be soon. To me, it seems like a 40-year process where you have some gains coming now, but it's mostly complete within 40 years that you just won't die of most of the things that kill people— cancer, heart attacks. There'll be accidents and there'll be deaths by old age.

Yeah, that's an extreme white pill. I love it.

Uh, and that's likely. And again, it doesn't have to rely on AGI though. Of course, AGI can help and accelerate that.

How do you think about your information diet, your information consumption? It's obviously AI-augmented, but I imagine you don't start the day with the prompt to an LLM, and maybe the, the LLM prompts come in once you've explored or read some source text and want more context. Is that how you're using them?

Oh, you imagine wrong. I wake up in the morning, I had questions when I was lying in bed, questions when I was falling asleep, and to ask an LLM something can easily happen in the first 10 minutes.

Okay.

Because I've been thinking and wondering, and I don't get up just to ask the LLM, but once I'm up, it's like, hey, I'm here, why not? And if the question takes a while to answer, then I scroll through Twitter, email, whatever. It's perfect. You want to set your queries in motion early in the day.

Yeah, I actually do find that often I'll kick off a deep research report before bed because I know it's cooking for me when I wake up. I love that. That's actually a great workflow.

Iran is in flux. Very, very hard to tell, you know, what the outcome will be there. But if you have a free Iran, how should people be thinking about how that would impact world markets?

I think it's unrealistic to expect a free Iran. I would like to see a stable Iran, which has been rare in world history, though Iranians are often very successful. There are so many ethnic groups in the territory, and they have expanded or contracted so many times. What they need is stability and some breathing space and a government that's not one of the very worst in the world. And I think the chances of getting that now are pretty high, maybe 50%. But we should set our sights a bit low and not actually reach for too much and let the Iranian people over time, you know, make it as good as they can and not have it be a question of outside influence. So, oh, we're giving you this, we're making you do that. I think that's counterproductive. Mm-hmm.

What about the looking back, the postmortem on the tariff and the trade war from last year? Obviously, the markets did very well even though they were tumultuous. During the process. What, what's the economic view on the impact of the tariffs in 2025?

Not as bad as we thought, but no gain, no upside. Manufacturing in this country, manufacturing employment, they're still falling. Our allies are mad at us. People pay higher prices. When I buy my favorite Rainier cherries from Chile, they're £9.99, £9.99 a pound instead of £6.99 the year before. What's the point of that? So I'm against it.

Against it. Well, we are certainly not against your appearances here. We love having you on the show and we appreciate you taking the time to come chat with us. Thank you so much for coming.

Always a pleasure.

And always a pleasure. I'm glad your 2026 is off to a great start. I hope you have a great rest of the week. We'll talk to you soon.

Take care.

Great to see you.

Cheers. Have a great rest of your day. Finn.ai, the number one AI agent for customer service. If you want AI to handle your customer support, go to finn.ai. And I'm also going to tell you about Cognition. They're the makers of Devin, the AI software engineer. Crush your backlog with your personal AI engineering team. And up next, we have Cerebras in the Restream waiting room. Andrew Feldman is the