软件股崩盘、Claude 的“命中名单”、国情咨文反响、Trump 关税转向

Today, we have a Conspiracy Corner episode for you. We're going to go over the 9/11 inside job, flat Earth, and the JFK assassination. It's going to be all conspiracy all the time after our amazing blockbuster episode during ski week. We're going all conspiracy here. Our guest today: Alex Jones.

How many views did it get? 9 views?

I mean, it's tough when you have 1 out of 4 besties. Michael Tracy is on standby.

Not true. I can carry an episode for at least 400,000 views—

I mean, you might.

By myself.

I liked your—hey, for people who don't know, Chamath has his own YouTube channel. He's got his escape hatch for when this train wreck burns to the ground. He started his own YouTube channel, and he's hedging his bets. Friedberg's working on his solo project. Everybody's doing solo projects. The band's got a lot of solo projects going on.

The Beatles are experimenting.

They're experimenting. We got a little Yoko Ono situation going on here. You know what the number-one topic for this show was by the All-In AI bot, Sacks?

What's that?

The number one was Dario versus Hegseth. The Department of War versus Anthropic was the number-one topic selected by our AI bot. We've got a full docket. The Claude kill list has expanded, and an AI fan-fiction Substack tanked your 401(k) on Monday. Let's get into it.

1. Anthropic Tanks Another Sector

Anthropic's generational run continues. They're now 3-for-3 in tanking different market sectors, Chamath, in February. Congratulations. This was like they took the mantle from Brad Gerstner, tanking the market.

The Anthropic list.

It is. On February 3, Anthropic announced, "Hey, we got a legal plug-in for Claude: Cowork." Thomson Reuters, LexisNexis, and LegalZoom were all down at least 10% since February 3. Then, on February 20, Claude Code Security was announced in a limited research preview. Stocks tanked again. CrowdStrike, Cloudflare, and Okta were all down.

Then, on February 23, Anthropic announced that Claude can modernize COBOL codebases. If you don't know COBOL, that's the oldest coding language in the world. That's where Sacks learned to code when he was in college in the '70s. It's used for banking, payroll, government—

Healthcare.

Healthcare. It runs 95% of ATMs in the U.S., and it powers Social Security payments. 85% of all COBOL code runs on IBM machines, so IBM decided they would tank 13% on Monday, their worst day since 2000: $31 billion in market-cap losses.

Let's stop here before I get into the fan-fiction piece. What's your take on what's happening in the market, Chamath? Is this simply people looking for an excuse to trim their positions because things have been at all-time highs, and people are just looking for an excuse? Or is this reality? Is this the go-forward reality—that AI is going to compress these kinds of stocks because it solves a lot of problems?

2. Markets Question AI Durability

I'm going to give you 2 explanations. I don't know what percentage I would allocate across the 2, but I think 1 is tactical and 1 is much more strategic, and I think both are happening. The tactical one is that we're at a moment in time where a lot of the smart-money hedge funds are starting to massively degross. What that means is they're trimming a lot of positions, and they're just taking on a lot less risk.

Hmm.

Why? I don't exactly know. It could be motivated by the second thing that I'm going to talk about, but the point is, in a degrossing cycle, you tend to be trimming risk and making your position sizes much smaller. The longs become less long, the shorts become less short, and you just shrink, so there's general downward pressure. That is a clear behavior right now.

But I think the structural change is the more important one, and this is sort of what I talked about this morning. In a normally functioning market, what we are always debating is when a set of cash flows go from becoming highly confident to less highly confident. It's a when conversation. So when will Coca-Cola's cash flows be impacted? When will Eli Lilly's cash flows be impacted? When will Meta's cash flows be impacted?

The answer to the when gets translated by the public markets into 3 things. Your price-to-earnings multiple, where, if you invert that number, what it's equivalent to is the yield on the money that you get. If you're at 20 times P/E, that's a 5% yield. The second is a revenue multiple, and the third is what's called your weighted average cost of capital.

If you look at the next 20 to 30 years of earnings and you want to figure out what that is worth today, you have to discount all of these back, and you have to assume a percentage of interest, effectively, that it takes to get there. The basic math is that when you have a high WACC, you're massively discounting these cash flows. When you have a low WACC, you're assuming that these things are very durable.

So what is happening? We used to debate when. This is no longer a when moment. The market is very much in an if mode: Are these cash flows durable at all?

Hmm.

Could they fall off a cliff in year 3? Is there some AI model that's going to come around the corner and obliterate this business without me knowing it? Because they've shifted into this if mindset, your risk becomes totally different. You have this event risk that you don't know how to price.

Whenever the market shifts into that mode, what you see is that the holders of those equities want a massive margin of safety. What does that mean? They have to take P/Es way down. If you used to trade at 40, you should trade at 20. If you used to trade at 20, you should trade at 10. They take revenue multiples down. You used to trade at 10 times revenue; now you're going to trade at 3 times. You take the WACC way up. It used to be a 6% discounted weighted average cost of capital. You know what? I'm taking you to 12% or 13%.

That's the market's way of saying, "I'm now debating if these things will even exist, and so I need to give myself a huge buffer to own this stuff." That's what's happening right now.

It has a lot of ripple effects that we can talk about. Friedberg and I have talked about this a lot. The most obvious impact is how these tech companies recruit and retain talent, because the biggest thing that it starts to eat into is the cash flow of a business, which really directly ties to stock-based comp and all this other stuff.

Let me just stop there. We have moved away from a when to now an if, and I think that is a very smart question to be asking. The answer may be that many of these companies will survive, but we don't know for how long. Until that becomes clearer, you have to give yourself room to be wrong.

You said "when," then "if." Did you mean "if" to "when"?

No. We've always debated when. When will these cash flows disappear? Now it's, "Will they even exist?"

Right. Okay.

We have always debated when. Now it's, "Will they even exist?"

Got it. Okay. So the second part of this story, Friedberg and Sacks, is that a Substack fan-fiction post taking place in the fictional 2028 Global Intelligence Crisis went mega-viral: 28 million views on X. It was posted Sunday night, and it made the market tank on Monday.

In this fictional Substack post, the author said there's essentially going to be a death spiral because of AI. How does that work? First, companies embrace AI. Everything goes right. They're able to cut staff, and their margins go up, similar to how Amazon has trimmed its white-collar staff.

Then they're so successful at this that they lose their customer base because consumers don't have discretionary funding to spend. That creates a death spiral where the companies keep deploying AI to try to hit their margins, cutting staff, and the entire economy collapses. Dr. Doom-level stuff. Unemployment is at 10%. The S&P goes down 38% from its highs.

After this piece came out, which speculated that agents would get rid of all the 3% interchange fees and move everybody to settle transactions on stablecoins, all the financial stocks got hit on Monday. AmEx was down 8%, Capital One was down 8%, Mastercard was down 6%, Visa was down 4%, and so on.

Finally, this piece got a lot of pushback. There was a silly section where they said AI agents would vibe-code their way to displacing DoorDash, and that's kind of silly, as anybody who's run a network-based business knows. Sacks, I assume you read this piece or at least saw the fallout from it. What's your take? Then we'll go to you, Friedberg.

I know this Citrini Research article got passed around like a joint at a Grateful Dead concert. But I'm starting to question how legitimately viral it really was.

There's some information that just came out that the attribution of the article has been amended, meaning the co-authors have been amended to include a short fund that was shorting some of the names mentioned in the article. This is according to another post that just came out.

According to this post, the authorship attribution of the market-moving article was changed after publication, and the co-author is a managing partner of a $262 million SEC-registered hedge fund.

Ooh.

...

who confirmed short positions in the companies the report named. So I think that's point number 1: I just wonder, did this article truly go viral, or did the authors do anything to amplify it? We just don't know the answer to that question. But regardless of that, let's just take the arguments on their face. I think one of the best responses to it was by another writer named Derek Thompson, who wrote an article called “Nobody Knows Anything,” which I think is a reference to a famous take by legendary Hollywood writer William Goldman.

In any event, what the article says is no one really knows what's going to happen with AI in 2 years, never mind 20 years, and so they resort to science-fiction writing masquerading as analysis. And the author here, Derek Thompson, says that the conversation about AI is really just a marketplace of competing science-fiction narratives. And he says, “That's not to say I think the technology is a parlor trick, but rather that the level of uncertainty is so high and the quality and supply of real-world, real-time information about AI's macroeconomic effects so paltry that very serious conversations about AI are often more literary than genuinely analytical.”

In other words, what he's saying is, look, this guy's writing very compelling science fiction, but there's no real analytics behind it to defend it. And yes, this could happen. Here's a prediction market on whether people believe the Citrini report's going to come true: something like 12% believe the Citrini scenario's going to happen. But the truth is no one really knows.

I mean, there are other dueling science-fiction narratives where AI's going to create such a world of abundance that we're not going to need for anything. And, just by the way, Derek Thompson is one of the abundance guys with Ezra Klein.

Yeah.

This is why the market's getting whacked. I think that—you're right, Sacks. Nobody knows. So if you can get 5% for owning government bonds, why are we taking excessive risk here?

Right. Let me just build on your point about SaaS. So the reason why there's so much uncertainty around SaaS is that SaaS used to be such an easily modeled and predictable category.

As a VC, we saw the same story play out across many, many different categories of software. You'd have this initial period where there'd be an experimentation phase. You'd have a bunch of different products that come to market. There'd be a battle, and then the market would eventually settle. There'd be a category leader, and they would capture most of the market share and the vast majority of the market capitalization. They would have very predictable metrics.

It was very easy to grade a SaaS business. You look at ARR, you see annual recurring revenue. You look at the net dollar retention. You want to see, depending on the phase—

RPO, RPO—revenue under performance obligation.

Right. And so these things began to be seen as an annuity with growth, right?

They were rock-solid, yeah.

Yeah, because a good net dollar retention would be something like 120%, which means that your cohort of existing customers, on balance, would all renew the next year, and they would renew at 120% of their previous year's contract values. And the reason you got that extra 20% is they would buy more seats, or there'd be additional products or features they would upsell.

It got to be very predictable. And so when people were buying software companies at, I don't know, 13 times ARR, they thought they were buying a growth annuity. And now all of a sudden, you have to factor into that: well, wait a second, what if AI disrupts the whole market? I don't think AI is going to get rid of Salesforce, but it could eat into its growth opportunity. We just don't know.

What if it changes the pricing model? It creates a whole lot of unknowns. I actually don't believe in the Satya Nadella or the doomer take on this, but I can see why the market would feel this level of uncertainty, given how predictable a category SaaS used to be just a year ago.

Yeah.

Well, chaos is a ladder, Friedberg, and this means opportunity. So if we look at this and SaaS has headwinds, then is there a winner? Is open source the winner, or is this all deflationary in your mind, Friedberg, and we just make less money and the earnings of these companies get compressed, the size of them gets compressed? How do you think about it?

I think fundamentally, if you're driving productivity with AI, you're driving leverage on human time and leverage on capital. The question is, how quickly can you drive that up? And that's a function of how much consumption there is, how much capacity there is for consumption.

On the one hand, I'll just speak broadly. I think humans have this desire to improve their livelihoods by roughly 10% every year. Meaning, your income and your ability to purchase stuff that's new relative to where you were last year has to go up by 10% for you to be happy. If it's less than 10%, you're probably unhappy.

Is that your anecdote, or is that like—

That's just an anecdote. I think that's sort of my rubric for thinking about why people are unhappy or happy. So if your earnings are the same, but things are getting more expensive, you're not happy. If your earnings go up by 10% and things stay the same price, you got 10% more than you had last year, you're going to be happy.

I just think all humans are driven by this need to consume more each year than they did last year. So I think, for me, that's the lower limit on consumptive capacity in the world. The question that we're now facing, which we've never faced in human history before, is: Is there an upper limit on consumptive capacity?

Hmm.

Because AI creates such a profound shift in productivity and in leverage, normally you would say, “Hey, when we get a new tool or we get new leverage in a system, we build a new technology, we can make more with less. Therefore, everyone gets access to more things for the same price, or the cost of things that they consume comes down by a certain amount.”

But there may be a situation now where the ability to make stuff exceeds the capacity to consume stuff. That is something that I don't think we've faced before, and I think that's where a lot of the models start to break: just general economic models, just general productivity models, and general social models. This goes to the point about what is everyone going to do.

In the same way that we've argued that maybe SaaS was a transitory business phenomenon that existed between the foundation of the internet and the era of AI, it may be the case that knowledge work in general is also a transitory phenomenon that only existed between the foundation of the computer, or computing tools, and the existence of AI, generally speaking.

And if all of that goes away very quickly, and all of those people can be redistributed and recast into doing other higher-level, more creative things, their productivity goes up by 100x, is there really a consumer on the other end of all of that productivity? Is there really enough consumptive capacity? I think that's the profound question that we all face. I don't think there's any limit to productivity.

Sorry, is that your way of saying that SaaS goes to zero, or is that your way of saying these companies go to zero?

I'm just saying knowledge work in general.

Yeah.

But is this just another dueling science-fiction take? Or what's your evidence for this? I think it's fine to have a sci-fi take about the future.

I think it's intuition.

Yeah, but is there data? Because I can show you some data that I think contradicts what you're saying.

In the sense that there are more leverage stacks that people are able to actually deliver.

Well, Jason, I want to hear what you have to say because I know you're experimenting with this, but let me just show you a few data points real quick, because I think this is relevant.

So we're really talking about the disruption caused by coding assistance, right? This is the first big killer app of AI. I guess after writing and research for chatbots, and we'll have agents later, but really it's all about coding assistance, right? It's the ability to more easily create code. That's what's creating the disruption to the SaaS category.

Let's just focus on the data we see right now around that. There are a lot of people pointing this out: Anthropic right now has a job listing for a software engineer on its website for $570,000. And a lot of people are pointing out, okay, so wait—what Anthropic is saying is that they're still trying to hire software engineers at a very high wage, but somehow they think these jobs are going to be eliminated?

Chamath might apply for that job.

Yeah. Austerity measures.

Sounds pretty good to me.

That's a lot of money.

Chamath might take that job and then just have AI do it for him.

No, I'm worried. I hope my AD19 team doesn't see that offer.

It's a big number.

Our equity is way higher, but our salaries are not that high.

You put these things together, and it’s like that equity is money. Good. The reality is, those guys are doing $5 to $6 billion in structured secondaries every year now—or they’re starting to, which means that they will. That’s like cash compensation.

Right. I think a lot of people are pointing out, well, this is a contradiction. Anthropic doesn’t really seem to be practicing what they’re preaching if they’re still paying enormous amounts for software engineers, even as they claim they’re obviously leading the entire category. Something doesn’t quite add up. Citadel Securities did a new report that rebuts that Citrini report, and they show a couple of stats here which I think are really interesting.

Job postings for software engineers are rapidly rising. They’re showing roughly a 10% year-over-year increase in the demand for software engineers. On a related note, they also show that company formation is rapidly expanding, and that may have something to do with AI making it easier to start a business or to get leverage, to your point, Friedberg. So there are a couple of competing effects going on here, and I think Aaron Levy had a really good explanation of why you might see something very counterintuitive happening. Again, it all goes back to Jevons paradox.

What Aaron says is that when you lower the cost of something that was previously supply-constrained, demand for that thing goes up. Software engineering is just one of the easiest examples to contemplate, but there are going to be many other jobs like that. Think about software engineering: even among startups in Silicon Valley, which I think are probably some of the most attractive places for software engineers to work, there’s always been a chronic shortage of them. Then you’ve got the Fortune 500 companies—non-tech companies—which have always had an even harder time hiring technical talent. So you have this massive unfilled need for software engineers across the entire economy.

Now you’re going to be able to get a lot more leverage out of software engineers. It doesn’t mean they’re going to get fired; it just means that now maybe you can have a lot more 10x software engineers, and those jobs are now being spread throughout the whole economy. I also think, just to put some numbers on this, the cost structure of the average Fortune 500 business is something like 5% IT. And that includes all of their IT, not just their software. What should it be? What should the percentage of software be in an enterprise cost structure? Elon describes companies as cybernetic organisms that are part software, part human.

I think—

If you think about the current Fortune 500 company being 1% or 2% software, maybe they should be 50% software. I think what Aaron is saying here is the market for software and software engineers was so constrained by the lack of availability that even if we 10x or 100x the productivity of software engineers, the demand will be there to absorb this new supply. And so it could lead to this explosion in productivity without the massive job loss.

I think you’re right. The thing that I would look at is that I would expect OPEX as a percentage of revenue to fall off a cliff. But within that OPEX, the percentage of it that you allocate to technology and technology-related things probably goes way, way up from what it is today.

Okay, Jason, the batch of people that are applying for Launch—has SaaS stopped? Has software stopped?

No. They’re AI-first companies, obviously.

Are they rebuilding traditional SaaS tools, just cheaper?

3. OpenClaw Automates Knowledge Work

Basically. Everybody’s building the best pilot in the world. As we talked about at the All-In Summit, some of these companies are trying to build the best copilot in the world, or Waymo’s trying to build the best driver in the world. People are now trying to build the best SDR in the world, the best salesperson, the best executive coach.

We have been obsessed with Claude Cowork, but mainly OpenClaw. What we did was—and I think it’s not developers that are going to do all this work; it’s knowledge workers. We have 20 people in our firm. We had 15 of them come in this weekend, and they all got trained over 6 or 7 hours. They each had to have their own OpenClaw agent, and we started building.

Every piece of software that we wanted to buy or build over the last 10 years that we never got to, my people are building in the last 30 days. As an example, when you’re selling ads for a podcast, you want to check all the other podcasts and what advertisers they have. We trained an agent to go take the top 100 podcasts, look through the transcripts, figure out who the advertisers are, check those advertisers in Pipedrive, tell us when we last contacted them, and put it into the sales room. That was an SDR job that we wanted to fill and software we wanted to build.

Sorry, hold on. Was that a human that you were paying money for, and now you’ve replaced it with software? Or does that human still exist, but now they just do it in a better way?

We’re redeploying that human. We had a human doing it; we’re going to redeploy them to do other things. The consistency and accuracy of this, Chamath—and then it’s doing it all night long. We have 7 of these agents in these kinds of roles.

The next piece we did was give my agent, which is like Ultron, root access to Gmail, Calendar, Zoom, Notion, and Slack. What it’s doing is giving each person, “Here’s what you got done this week,” and giving their manager, “Here are the emails you sent, here are the meetings you took, here are the contacts, here are the threads you were involved in.” And then it’s helping manage those people.

Okay, but all that, to me, says you, Jason, despite all your doomerism, seem like you’re growing, and you’re going to be hiring more people, and you’re more productive. Am I getting this wrong?

No, no, no, I’m not doomerist.

But you’re growing, and you’re going to be hiring more people.

No, we’re not going to add more people. Definitely not adding people. The people we have are becoming 10% or 20% more efficient every week because the software we would’ve paid for or bought from another vendor, if we had the time, or the custom software we wanted to build with 10 engineers, is being built by our OpenClaw agents.

As an example, when we make clips for this podcast and other podcasts, we have it look at an episode of This Week in Startups from 10 years ago, tell us the 3 best moments, make the clip, put the subtitles on it, and put the clip into the Slack room. That was something that was going to be a full-time job. So we’re getting 10% or 20% more efficient.

Then I started doing it at home. I had it pull up our last 10 Instacart orders, tell us what we order most of the time, and then automatically build a cart for us. Every single knowledge-work job is being automated right now. If you’re a business-process head, you know how to do a business process, and you can structure it and write it with an agent; it’ll just run it every day, every week.

We did another agent around the question, “How do you make better thumbnails?” We told it to update its skills every Saturday. When you build an OpenClaw, it has a soul file and a skills file. In the skills file, we told it, “Sacks, every week, go out and look for people discussing how to make better thumbnails on YouTube and how to make better titles.”

It found somebody at MrBeast’s company talking about how they’re using heat maps. It was an article I would’ve never known about. It added it to its skill, and now whenever we post a thumbnail, it tells us, based on its skill that it refines every week, how to make that thumbnail better. And it’s starting to make the thumbnails.

This is becoming recursive. So you keep the same number of people, but they get 10% or 20% more efficient. I don’t know what this means for the larger economy. All I know is it’s the most exciting time I’ve had online since the internet came out. It is so much fun to automate all this stuff.

4. Companies Need Cash To Adapt

The big question that I’m thinking about, that I haven’t gotten a good answer to, is—and I don’t know what you guys think—whether all these businesses are going to need to batten down the hatches and give themselves room to figure this all out, right?

If you take Sacks’s point, and if you take your point, JCal, the young, nimble companies like yours are going to be rapidly experimenting. The larger companies are going to slowly onboard themselves and start experimenting. All of that means we’re going to get much clearer answers to all of this, but what it also means is that you’re going to have to have time so that you can figure this all out.

Yes.

And if you want to buy yourself time, you’re going to need a ton of cash. If you’re thinking about saving cash, the one place tech companies literally incinerate cash is in how they do compensation. So I kind of think that, at some point, the next shoe will drop, and all of these tech companies will have to really look at stock-based comp because they literally incinerate most, if not all, of their free cash flow fighting the dilution from stock-based compensation.

If you want 5 or 6 years to just be in the arena, on the field, figuring this out, you’re going to want to be very cash-flow-generative and really conservative in how you spend your money.

Yeah, Sacks. The people who embrace this, I think, become 5 or 10 times more valuable than the people who are not. That's where I think the opportunity in the economy is. So unless you think humanity's going to run out of problems to solve, I think it's going to be boom. It's going to be boomtown, and I think people are going to start more companies because the barrier to starting a company is no longer 3 or 4 million dollars. You could just have 2 or 3 people, and you start setting up these agents—

Mm-hmm.

—and, man, you can—

You can cook.

—you can make software—

You can cook.

—you can do sales, you can do PR. Everything is getting faster and faster and faster. So the time between conceiving of a product and publishing it and finding a developer—you don't even need a developer. You can just publish software.

The wake-up moment for me was when we were talking to our agent about, “Hey, we want to get this functionality out of Slack,” and it was like, “Yeah, Slack doesn't have that, but have you considered Matterpost?” I'm like, “What's Mattermost?” It said, “Oh, it's an open-source project. I can spin it up this weekend, export your Slack instance, and put it there.” And I was like, “Well, don't do that.” We're only spending $6K a year on Slack—or $10K a year.

But the software is building CRM systems for us, it's building agents for us, and it wants to just build the whole software stack. So when you renegotiate with Slack or HubSpot or whatever company you're working with, you're going to be able to say to them, “Hey, we could roll our own.” And when you want to upsell us on this latest thing, like you talked about, Sacks, upselling is such a big part of SaaS, you're like, “I can actually build that software myself internally. I don't need you to do it.”

Ryan Peterson just posted on X, “Claude for legal seems to work just as well as Harvey, by the way.” Now the SaaS apocalypse—

He's dunking on it.

—the SaaS apocalypse is going after private companies now, too.

Well, I think for a while now, there has been a question of which layer of the stack is going to capture all the value. Is it going to be the model companies, or could it be the applications that are built on top of the models? Or if there's a lot of competition at both those layers of the stack, do the chip companies get it all? I think it's an unclear question, but—

Totally unclear.

Yeah, I think for any given vertical application, you do have to defend why you think your value proposition will be sustainable as the underlying foundation models just get better themselves.

And it's open source. This week, we put up Kimi 2.5. It can do about 80% to 85% of the jobs, so we lowered our token bills massively when we stood that up. All right, listen, this is TBD. We've got a lot more to think about on this topic.

Just on this point, a lot of these debates about AI are dueling science-fiction narratives. I just think that the doomer narratives are inherently more appealing to people. I mean, partly, you look at most sci-fi movies: They're dystopian, not utopian.

In addition to that, I think we have a bunch of heuristic biases in favor of the doomer narrative. One of them is the seen versus the unseen. It's a lot easier to see the jobs that already exist that could be obsoleted than it is to imagine the new jobs and the new business models that haven't been created yet, and that will likely take some great innovator or a genius to think of in order to create. So we have that huge heuristic bias of not being able to see the creation that's coming. It takes way less creativity to think about the potential destruction.

I think it's a big—

And then finally, I think the other heuristic is just the whole fixed-pie fallacy. Most people do tend to think of the economy as a fixed pie. This is why you see so much anger against millionaires and billionaires, because of this idea that if someone's getting rich, it must be at the expense of someone else. That's not actually the case. The economy itself could be growing larger as a result of someone inventing something new that increases production.

A really good line from another article that was written just a couple of weeks ago was, “The economy is not a pie, it's a garden, and technology is rain.” So again, all of this technological innovation is going to increase the growth rate of the garden. It's not a fixed pie. And just because you see an expansion in productivity in one part of the economy does not mean that you're going to see job loss in another part of the economy.

Yeah, I think the job people are not seeing, but I'm seeing right now, is the person who creates agents, manages them, and is the maestro of the agents—the person who can take the business process, explain it, and train the agent to do it. And there are certain people in business who are just really good at operations. You were one of them, Sacks, running companies. And that person who can fire up an agent, train the agent, figure out how to manage it, and figure out how to increase its sales—

Sure. Look, with any new technology—

It's like a great job.

So—

And it's not a developer.

Look, with any new technology, there's always a huge change-management aspect with enterprises because it's hard for them to adapt and change. The people in the organization who can lead that change management are the ones who are going to create an amazing career opportunity for themselves. But it's hard to do, and that's going to slow down the rate of change—just the amount of inertia in the economy.

One other constraint is going to be that, at some point here, we may be token-constrained, right? I mean, we may not have enough energy, like we've talked about. Even though the chips are getting so much better, so that tokens per second, tokens per watt, and tokens per dollar are all increasing very fast, we're still probably going to be constrained in the next couple of years on some dimension, whether it's land, power, and shell, or just energy production, or maybe chip production. There are real-world constraints on just how fast we can scale the infrastructure, and that will—

Hmm.

—mean that these hyper-utopian or hyper-dystopian narratives will be wrong. I don't think there's time in the next few years for the whole economy to change in the way that the extremes would present.

I think you're right. I think you're going to see a 10X increase in the demand for tokens, but I also think you're going to see a 90% price reduction in the cost per output token, probably by the end of this year. So I think that, to your point, it's going to just create an enormous upswell of demand because we're going to be able to cut the prices of an output token so dramatically. And I think that's going to—

And by the way, that discussion we had, Sacks, last week, when we talked about the tokens outpacing employee salary and just where are these tokens all going to come from, that was our most-viewed clip, or one of the most-viewed clips, in the history of this podcast. So people are actually really focused on this. Yeah.

I had my team at 89. We redid our cost model, and now we have that as a line item when we think about the fully burdened cost of employees. We now factor that in because we're at a place where some of our engineers are just racking up ginormous bills.

And then separately, just general runs that we do for general-purpose stuff that we need to run our product—it's so expensive. So I am waiting with bated breath for what Sacks said, which is that we need an explosion in the capacity that's available, because I do think that the silicon solutions are coming that will cut the cost, but we need a large block of land, power, and shell ready to then turn all of this stuff on so that we can actually take advantage of it.

Rumors are that the new Mac Studio is coming, will have an M5 chip in it, and will be language-model ready. So the rumor is that they're building it for models. That could be an incredible turn of events: Everybody's desktop running a local model. Sacks, you want to have the final word here? Friedberg, before we—

5. Data Centers Face Political Resistance

Just to go back to what Chamath was saying there, you've got political forces that want to stop the construction of all data centers in the United States, so—

Which is crazy. Yeah.

—if that gains steam, then that's going to be a huge constraint on any change whatsoever.

Can I tee this up for you, J. Cal? I went back this weekend, and I looked at the number of data centers that have faced local opposition and whether there were patterns, and I posted it on X, so Nick, maybe you can put this up. But it was really a very small minority that was pushing back on data centers and getting them canceled.

We had about 25 projects total, of which 20 were just in Q2 alone. There are 100 data center projects right now that are facing some form of local opposition. So interesting, because these—

Where if you take that 40% number and you apply this, and then you multiply it by the number of megawatts that they have announced, last year we lost almost 5 gigawatts in terms of canceled projects. This year, coming in 2026, we have about 7 that could be canceled if you use this math.

If then you flow that through, OpenAI's CFO, Sarah Friar, said this: Every gigawatt for OpenAI is about $10 billion of revenue. So if you assume that that's roughly accurate, plus or minus $1 billion here or there, what that means is that in 2025, the industry as a whole lost $50 billion of revenue.

And this year, if 7 gigawatts gets canceled, it's about $70 billion. Now you're talking about $130 billion of lost revenue over these 2 years that'll go forward in time that we miss out on. I think that's really bad. We need to figure out a way to nip this in the bud.

It's so confounding because we were sitting here 5 years ago, 10 years ago, and local municipalities were fighting and giving discounts to try to get these data centers open to get the jobs and the revenue. Now we've got people trying to stop them. This is a perfect transition for the State of the Union.

One of the big topics, and I think something you're working on with President Trump, Sacks, is this energy pledge. I've been seeing rumblings about this. Explain what's going on in terms of getting the country in sync around these data centers and energy.

Well, the President announced in the State of the Union last night that he supports a ratepayer protection pledge, which requires the major tech companies to provide for their own power needs for AI data centers so that residential consumers do not see their rates going up.

I think this makes total sense. Chamath, to your point, this is the reason behind a lot of the opposition to new data centers: local residents fear that their electricity prices are going to go up, and that shouldn't be the case. The President said that he's committed to not allowing residential rates to go up as a result of data centers.

It's pretty straightforward. You get the big tech companies, the hyperscalers, to pay for the increase in the electricity cost, or you let them set up their own power behind the meter. The President's been talking about this for over a year: our biggest AI companies would also become big power companies if we let them stand up their own power generation behind the meter.

These data centers don't even have to connect to the grid. They could just co-locate themselves. But also, I think that with this ratepayer protection pledge, what you're going to see is that it could actually bring down consumer prices, because when these data centers set up their own power and connect to the grid, they can give the excess back to the grid.

They will also make investments in scaling the infrastructure. Although electricity is priced at a metered rate, the cost to generate it is not all variable. There are a lot of huge fixed costs in there. So when you increase scale, you can actually reduce the metered rate.

Again, this is really, I think, the rebuttal to Bernie Sanders, who just wants to stop all progress whatsoever. I saw a funny post calling it BANANAS, which is “Build Absolutely Nothing Anywhere Near Anyone.” BANANAS is replacing the new NIMBY, so you just can't build absolutely nothing.

That sounds sustainable.

I think the President's approach finds a very good balance here, which is: We can have progress; just don't make residential consumers pay for it. Let the big tech companies pay for it themselves, and I think you'll see more coming out about this from the White House next week.

Yeah, it's bananas.

Quite a deft move. Friedberg, how should America be thinking about this great data center build-out and energy usage if you expand it out over the coming decade? And how do you sell that against the backdrop that you talk about—the socialist movement? How do you think about those competing forces? You've got the socialists saying, “BANANAS, NIMBY, slow down, decel,” and then you've got this incredible race we're in for efficiency and this opportunity and abundance. How would you sell it to bring these 2 sides together, or is it just impossible?

The data coming in and out of data centers moves at roughly the speed of light, so you could put them anywhere. I think that our policymakers need to be very cognizant of that fact.

We connect the internet using high-speed cable and high-speed fiber optic throughout the world. So theoretically, if we don't embrace and allow the economic development of the data center industry—and it will fundamentally be an industry because it is almost like the new sort of oil—where are the oil rigs going to go? Where are the railroads going to go? Where are the telegraph lines going to go? Where are the factories going to go?

If we don't put them here, someone else will put them on their shores. Someone else will put them in their country. Someone else will put them in their jurisdiction. A lot of the economic value that arises from the people who will build those facilities, the energy that will be installed to produce power for those facilities, and all of the second- and third-order industries that emerge as a result of those installations—that value will accrue elsewhere.

Such a good point.

So it's not going to just go away. The demand is there. The economy's moving forward. AI's moving forward. We live in a world with 196 countries, and data centers do not take up a lot of space. They're very small relative to the economic value that they produce.

If you zoom out on the map of the world, all the data centers in the world fit under the tip of a pin. This is a very small footprint, and if we're going to give up hundreds of thousands of jobs and many billions of dollars of economic value creation, we're being pretty silly and pretty obtuse in our view of the world.

I would just encourage the system that I think is the right system, and we talked about this last time, where, provided that data centers are producing their own electricity, that means that you're taking electricity consumption off the grid because they otherwise are not being used on the grid. That will reduce the cost of electricity for other residential and industrial users.

So it's silly to think that we need to put a moratorium on data centers. As soon as you do that, the companies that use data centers are not going to slow down. They're going to go put them somewhere else, and we're going to miss out.

And it's such a good point, Chamath, because you were recently in the Middle East, and I've been there a bunch—in Saudi, the UAE. These are the folks who built a large portion of those oil refineries, and they are savvy to this. What are they doing in Saudi, the UAE, Qatar, and all of these regions? They're 10Xing their data center builds.

So to your point, Friedberg, either we build them or they're going to go somewhere else, and there are people who are willing to underwrite these. They're willing to take out the red tape from the process here and move quicker than us.

So I think this is a pretty deft move by President Trump to say, “Hey, you guys should all just guarantee that consumers don’t get impacted.” The water thing is a total hoax. The water is recirculated. That’s a hoax.

I think this is really smart. I think that what the president’s doing and what Sacks is doing is really smart. The thing to keep in mind is that there’s still a risk that prices go up, and it has nothing to do with these data centers. It has everything to do with the business model of being a utility.

Because what happens is, in order to get a license—a monopoly license—in an area to provide energy, to generate energy for a community, the exchange works in the following way. You go and you present a CapEx plan to the Public Utilities Commission. That’s effectively your budget. That says, “Here are the lines I’m going to upgrade. Here are the generators I’m going to upgrade.”

Independent of data centers, the reality is the electricity consumption of individual Americans is going up because we have more devices, we have cars, and we have all of these other things. So what we also have to do is look at how utilities’ business model actually incentivizes them to increase prices by making all kinds of investments.

We have to do a good job of making sure we hold everybody accountable. Otherwise, what you could see is the data centers taking on the burden for themselves, but prices still continuing to escalate because a utility says, “I need to spend $1 billion this year to upgrade my infrastructure.” What that allows them to do is take that $1 billion and essentially invest it for a return. That’s the business model of a utility.

And this is really happening in blue states. Micron has a $100 billion megafab in New York, and there’s a lawsuit by six—

This is shameful.

Six concerned citizens.

That’s shameful.

Jacob Helberg

And the project has taken—

That’s shameful.

1,200 days between their announcement and the groundbreaking, and they spent 612 days on the environmental impact study. People, wake up. Just go to Texas. Elon built his factory here, the Gigafactory, in under 18 months. This is the great state of Texas. Come here. We’ll build it for you, and you’ll be done.

Yeah, I don’t know why anyone bothers with the blue states anymore. They make it too hard to build.

It’s retarded. It’s so dumb. It’s such a self-own, too. Don’t you want to be part of the future? You’re literally kneecapping the entire country to scratch the odd itch.

By the way, there are a lot of people in New York who want to work. This is not actually a case of this new fab being unpopular. The majority of people in the area actually want this plant being built. They want the jobs that are going to come there.

A lot of people say data centers don’t create a lot of jobs. This is actually a chip fab, so it will create a lot of jobs, a lot of good, high-paying jobs. People want it.

Totally.

But six people can stop it with a lawsuit—

It’s not—

—after it’s already been through a 2-year environmental review.

It’s not blue and red states. These are nonprofits that get organized to create this kind of chaos. I remember looking at a massive lithium investment in Nevada, and the whole point was to domesticate lithium production.

What was interesting is this enormous deposit that’s just sitting there, ripe for development. Right before they were about to get environmental approvals—or right after—there was a lawsuit by people who wanted to protect the sage grouse. It’s seared in my mind that what was likely the sage grouse of Nevada is the reason why we do not have domestic national security around lithium.

You have to ask yourself, why is this possible? It’s possible because you have these environmental nonprofits that can go and create this chaos with absolutely no risk to them. Zero. They can fundraise around it, and they can create this chaos.

I mean, Sacks, to this point, this is an example of Greenpeace. Specifically, here they were pushing back on an oil pipeline to such a degree that they created so much chaos that they were sued. A North Dakota judge just said that he’s going to order Greenpeace to pay damages that should total almost $350 million in connection to those protests.

Nice.

It should not be the case that six people can slow down a $100 billion investment package. That’s not right.

Well, I think there’s—and I just want to highlight this important point—not a lot of logic and reason. You guys are right, but I do think there’s a lot of emotion, and there’s a huge aversion to big tech, a huge aversion to wealth creation by select individuals and select companies, and a huge aversion to economic growth that doesn’t benefit everyone.

There’s a fundamental, underlying left-behind emotion that drives a lot of this. I’ve said it before, but I think unless there are systems or mechanisms that get folks to come along with the value creation ahead and help them connect their own lives to the value creation that’s being realized, they’re not going to be supportive.

There is this kind of diametric opposition toward big tech, toward the wealth gap, toward value accrual to a select few companies or select few individuals, and this fuels and feeds that. So I think, fundamentally, maybe it’s not just about giving the data centers their own power capacity, but there have to be mechanisms and tools that help the broader population understand, recognize, or get some benefit from it as well.

They have to be an owner in it or a participant in it because they have the power, as we’re seeing. They have the power to stop it. Therefore, they want to have some benefit for providing the authority to do it.

These six people are concerned about housing costs, worker exposure to toxic chemicals, pollution of air and water, greenhouse gas emissions, energy consumption, flooding of the wetlands—all these things that obviously could be mitigated. All right, let’s keep moving here. We’ve got a lot more docket to get through.

6. The State Of The Union

The State of the Union came in at 108 minutes, and it’s the longest in 60 years. Actually, the longest since they started tracking this. The theme of President Trump’s State of the Union this year was “America at 250: strong, prosperous, and respected.”

Trump took a bunch of victory laps: inflation, jobs, closing the border. All those have gone really well. But this comes against the backdrop of Trump’s approval rating being super challenged. He started his first year at +11.7%; now he’s at -14.3%, a 26-point swing.

The economy started at +3.4%, down to -18.2%. Trade started at +5.9%, and we’ll talk about the tariff stuff later, and went down to -22.7%. So let’s call balls and strikes here, gentlemen. Favorite moments: what were your favorite moments from the State of the Union, and what were your general impressions of 1 hour and 45 minutes of Trump going to town?

I thought it was great. I had a couple of favorite moments. One was the Ilhan Omar–Rashida Tlaib death stare, and them just losing their minds, screaming. I thought it was so un-American.

The second was when he was calling for law and order and focusing and prioritizing on American citizens—

Over illegal aliens.

Yeah. None of the Democrats stood up. I thought that was kind of foolish. It was obvious things, and the Democrats wouldn’t applaud, but this time they did, like they did for the hockey team, which I thought was the right thing to do.

The fourth thing is just a shout-out to our friend Brad Gerstner, who got a big shout-out from the president. I don’t know, Sacks, if you engineered that or not, but that was—

No, I—

That was fantastic.

That was incredible.

And it was like—yeah.

He got a double shout-out. It was like a double tap.

Yeah, that was really cool.

That was surreal on the group chat.

Our group chat went crazy.

It lit up, yeah.

It was really cool.

Those are my 4 highlights. Great. Here’s your 20-second clip of Democrats not standing for Americans over illegal aliens.

Donald Trump

If you agree with this statement, then stand up and show your support: the first duty of the American government is to protect American citizens, not illegal aliens.

Jacob Helberg

Why wouldn’t you stand for that? That’s an easy one to stand for. It doesn’t make any sense.

Would you stand for it, Jacob?

Yeah, I can be anti-ICE, but I’m pro-American and I’m pro-reasonable immigration, like 90% of the country is, so it just doesn’t make any sense.

But do you think American citizens should be prioritized over illegals?

Jacob Helberg

Well, of course. Of course, yes. And I also think there should be a path to citizenship for people who’ve been here for a while, and I think that’s what the majority of the country thinks as well.

Right, but your point is, I can hold 2 thoughts in my head, so I would have stood if he asked me.

Yes. Obviously, we should take care of American citizens first, and we should deport violent criminals. We’ve been over this a million times here. This is consensus in the country.

Yeah, but what do you think is going on in everybody else’s head when they say, “We cannot stand for this”?

This is out there”?

These two sides—I think it's like the tariff thing; it's like the ICE thing. These two sides cannot work together. It's just the most polarized it's ever been. Trump is not the kind of guy to reach across the aisle, and the Democrats are now digging in. So we just have a dysfunctional government where, in a more functional time period, like under Clinton, let's say, or Bush, people would have gotten together—and I'm jumping ahead to the tariff discussion—and they would've said, “Yeah, of course, tariffs are done in Congress. That's the law, whatever. What are your thoughts, Mr. President? How can we support your tariff program?”

But now it's like, oh, well, we don't work together. We don't actually have discussions anymore. There's no bipartisan collaboration. All these politicians are disgraceful, disgraziato, across the board. They should be working together for the American people. If the president wants to do tariffs—

Well, how do you—

They should be reasonable about it and give him the power to do reasonable tariffs, and he should be reasonable and say, “Hey, I understand that's your power. Let's get together, we'll chop it up, and let's have dinner together.” But they're just too polarized. It's just disgraceful where this country has gotten to. I blame both parties.

Whenever the Democrats get smoked out as being radicals and extremists, you always want to basically say, “A pox on both your houses,” and blame the Republicans and Democrats equally. The fact of the matter is, the President said to the audience, to the members of Congress, “Hey, if you agree with this statement, stand up.” And, of course, every single Democrat sat there stone-faced and refused to applaud or acknowledge what he was saying. This was a very easy test for the Democrats to pass.

Which is what I just said.

In fact—

What I—

It was a—

Literally what I said.

In fact, it was a political risk for the President because it was so easy for the Democrats to demonstrate—

They're dumb. Yeah.

—that they're operating in good faith and that they're willing to be bipartisan, and that they're not extremists, and they're actually commonsensical and logical. They completely failed the test.

Yeah.

And, by the way, it wasn't just on that one. Let me just tell you some of the other ones where they refused to applaud. They refused to applaud the grieving families of innocent American women and children murdered by criminal illegal aliens, including the mother of Irina Zhuravskaya.

That was very sad. That was sad.

That was unbelievable. They refused to applaud for securing our homeland and ending the invasion of criminal illegal aliens, killers, rapists, gang members, and traffickers. They refused to applaud for unifying against political violence. The President mentioned the assassination of Charlie Kirk, and they would not even do a polite clap for Erika Kirk in unifying against political violence.

They refused to applaud for keeping violent criminals locked up. They even refused to applaud for lower prescription drug prices for millions of Americans because it was President Trump who orchestrated that policy. There were so many other examples like that. I think the reason why this speech was so effective—and it's not just me saying it—was that something like 2/3 of the people CNN polled, two-thirds of CNN viewers, said it was highly effective. Something like 3/4 of CBS News viewers said it was highly effective.

Yeah.

It's because the President laid out 80/20 issues one after another, right? Or even 90/10 issues or 95/5 issues. These were all issues where the overwhelming number of Americans, I think, agree with the policy the President laid out. In every single case, the Democrats opposed it and wouldn't even give it polite applause.

That is different than in the past. You can say that's because of hyper-partisanship and polarization, but it's also because of another thing: the Democrats have become a party of radicalism and extremism. The viewpoints that they expressed through their aesthetics the other night, they do express those things in policy and in speeches all the time. So it's not just a one-off or somehow—

It's pretty crazy.

—we have a misconception of who these guys are. I think the big line of the night was when Trump said, “These people are crazy.” He said it in an almost mournful and regretful way. He doesn't want them to be crazy. He wants them to be rational so he can work with them.

I mean, just to—

But I think that point resonated—

I'll still point to the other side, which is this has been going on for a couple of State of the Unions. The Republicans often didn't stand for the Democrats, and it's a bit of showmanship. But the truth is, Trump is the divider-in-chief. He is always attacking people. He's always mocking people, so they don't want to play ball with him. So—

He's counterpunching.

I do think you can both-sides it.

You have no choice in politics. You gotta counterpunch.

No, you can—you can—

No, I don't think so. That's actually the problem with that philosophy, Trump's philosophy of, “We have to counterpunch, we have to attack, we never have to apologize, we never have to be reasonable.” That's part of what's broken down in our politics, and these two sides should work together. We should go back to a bipartisan—

Yes.

Really? How are you going to work with Ilhan Omar—

Listen—

—when the President talks about—

I'm not saying it's going to be easy, but she is the mirror—

Hold on. What, what—

Hold on, I'll finish my statement. You asked a question. I think Trump is the mirror of that. He has been hostile toward these Democrats. He doesn't give them an inch. They should be more collaborative.

The balance of power between the executive branch and these congressmen, Congress, and the Senate—this is how it's supposed to work. These two sides need to learn how to get back to listening to each other, understanding each other's positions, and then finding a middle ground. That's why the Democrats lost last time, because they didn't have the common sense to say, “Hey, everybody wants the border closed.” To your point, it's a 90% issue, and Kamala Harris was too dumb to just say, “Yeah, we should've closed the border. It's closed now, and we've got it.” Anyway, the whole thing is a mess. I understand you gotta fight for your team—

Well, you just—

I don't like the counterpunching. I like collaboration.

You just actually made the key point. You made the key point, which is, underlying the optics and the polarization, you have issues. On those issues, President Trump is on the side of the American people—the issues where 80% of the American people agree.

Some huge percentage, I don't know exactly what it is, thinks that the Somali daycare fraud in Minnesota was an outrage, and the President is right to point that out. And what's the Democrats' reaction? You've got Ilhan Omar screaming from the audience at him.

Yeah, she's a loon. At the end of the day—

But the rest of the Democrats are not that different.

He did have Elizabeth Warren stand for stopping Nancy Pelosi from trading stocks, and that gave Trump his best one-liner of the night. That was his best one-liner, clearly. And they stood for Iran too, and stopping Iran from becoming a nuclear power.

I give Elizabeth Warren credit for that.

There you go. You don't have to punch her back.

Hear, hear, hear—

You don't have to punch her back.

Donald Trump

We will have a Stop Insider Trading Act without delay.

Yeah. See? That's something bipartisan. Look at that, Zach. That's what you need to get the country back to.

Okay, but that—hold on. But this disproves the point you were making before. You said that it was polarization—

Donald Trump

They stood up for that? I can't believe it. Did Nancy Pelosi stand up, if she was here?

Can't believe it.

Donald Trump

Nailed it.

That's why he's so good: he's in the moment and reacting to what's happening in the chamber. He's not just—

He's got good timing. Good timing.

He's not just reading from a teleprompter—

Nailed it.

—and he nailed it. But look—

Good timing.

—that moment disproves what you were saying, JCal, because this is not just about polarization. On that issue, Elizabeth Warren was willing to stand because she actually, to her credit, wants to ban insider trading by members of Congress.

Yes.

But on the rest of those issues, like securing the border, she did not stand. Why? Because she does not agree with the President on that issue.

We just have to get back to these sides working together. That's my personal feeling. Friedberg, any thoughts on the theatrics and Trump's first year writ large, and the back-and-forth? Is there any hope that these 2 teams could collaborate at some point on something like, say, the ballooning deficit, which Trump has not gotten under control in his first year, and it's going to be $2.5 trillion added? What are your thoughts here on them collaborating on anything important, Friedberg?

It's probably one thing they can agree on: just keep the money flowing.

Got it. They’ll both give a standing ovation for burning more capital and putting us more in debt. Well said, my guy, David Friedberg. Sultan of science, it is your time to shine. The world’s greatest moderator has decided we’re going directly to Science Corner. This is your time to shine—

Sacks is waiting.

—and it’s Sacks’s time to drop a deuce.

Sacks is waiting immediately off camera.

Yeah, he has to drop a deuce.

I’m here. What’s up?

He’s still here.

But what do you need me for?

This is very important for you.

Friedberg’s going to talk.

Lightning round for Science Corner. Go, Friedberg.

Wait, wait, wait. Are we doing any more topics after this, or can I just leave?

Yes. Tariffs.

Yes, we’re doing tariffs.

Wait, why would we do that?

Three—

Because to keep you—

Because I want to get paid for Science Corner.

You’re going to put the audience to sleep.

Two—

I’m not saying—

One.

Science Corner doesn’t have its audience—

Listen, you can go take—

—but why wouldn’t we do— Yeah, exactly.

Go take your deuce—

Look at you. You’re screaming, you’re screaming—

Let him do his work. Three, two—

Let him cook, let him cook.

Friedberg, tell us about this Harvard study.

7. Yamanaka Factors Reverse Aging

Speaking of science, I think there’s a very important moment happening right now. We’ve talked a number of times on this show about Yamanaka factors. These are 4 proteins that were discovered by Shinya Yamanaka. We found later that, when applied to cells—mammalian cells—they can actually reverse the age of those cells, reset the epigenetic clock, and reset the epigenome, which is the little markers on top of the DNA that turn genes on and off, back to a youthful state.

This extraordinary, groundbreaking work won the Nobel Prize and led to the foundation of several companies. There’s a Harvard scientist named David Sinclair. He’s a bit of a controversial character. Do you guys know him? I think 1 or 2 of you may have met him. Chamath, have you ever met him?

I’ve followed him. I’ve seen his stuff.

Yeah.

David Sinclair is somewhat bemoaned by the scientific and academic community for being a little too overhyped—a snake oil salesman, as some folks have claimed—because years ago he sold a company to GSK saying resveratrol would reverse aging, and he made $720 million on that. It didn’t end up working, and he’s promoted certain supplement companies and so on. I want to preface with that before I underwrite what he’s saying with this next thing.

He’s a co-founder of a company called Life Biosciences, and they’ve reached a major agreement with the FDA to be the first company to treat humans with Yamanaka factors. Specifically, what they’re doing is delivering these Yamanaka factors—these proteins that rejuvenate cells and make them youthful again—into the eye.

Their first indication is to inject them into the vitreous fluid in the eyeball. They’ll affect the retina in the eye to address people who have gone blind from glaucoma or one of these stroke-like diseases that happen in the eye. The expectation with this Phase 1 clinical trial is that the delivery of these Yamanaka factors into the eye will rejuvenate the retina, make it youthful again, and restore vision.

If it works, which it’s expected to because we see this result happen in animal models, it could be an extraordinary breakthrough—not just in terms of blindness, but in terms of the first human application of Yamanaka factors to reverse aging.

The way they’re doing it is by packaging DNA that will make these proteins into a virus—an AAV virus—that is delivered into the eye. The virus will then go into the retinal cells and deliver this payload, the DNA to make these proteins, into the eye cells. It can be turned on and off.

Amazingly, they’ve created a switch mechanism in which the production of these Yamanaka factors can be turned on and off by taking an antibiotic called doxycycline. The person who receives the drug takes the antibiotic, turns on the production of these Yamanaka factors, and then, theoretically, their eye cells will de-age, become youthful, and their vision will be restored.

The Phase 1 clinical trial is underway. It’s the first time in human history that we’re seeing Yamanaka factors being delivered into humans. It’s literally the tip of the iceberg. There are now more than a dozen startups trying to deliver Yamanaka factors—which are these proteins—or some other sort of protein that can actually reverse aging by restoring the epigenome in cells and making them young again.

This is the beginning of a wave that I think will be the most extraordinary revolution in human therapeutics, and ultimately could lead to what some people would argue is the fountain of youth.

Is this just a tox study, or is it a Phase 1A?

Phase 1. They’re not—

1A?

Yeah, they’re going to see results. They’re going to keep dosing low, but you will see results.

God, that’s going to be incredible. Hmm.

It’s going to be incredible. By the way, a number of other folks are gearing up for Phase 1 using, if not the Yamanaka factors, other factors that they’ve identified or designed as an alternative to Yamanaka factors, again to rejuvenate the cells.

Just to remind folks, the way this works is that it was discovered that these proteins, when they go into a cell, take all of those little markers that sit on top of your DNA and turn genes on and off, and create a system that causes them all to move to the right place.

Hmm.

So it resets the markers so that those cells—

Incredible.

—will start to operate like they’re supposed to, as when they were young again.

That’s going to be wild.

It’s going to be incredible. Yeah.

Fucking wild.

Do you think, in all seriousness, people’s knees or joints—where do you think it could flow to next?

Arthritis.

Yep. And, by the way, when applied and distributed in the skin, they’ve seen some results in monkeys where wrinkles go away. It literally makes—

Hmm.

—these cells all work youthfully again. A lot of the damage that happens over time is not damage to DNA; it’s damage to the epigenome. These are the parts that sit on top of the DNA and turn genes on and off, and they get moved to the wrong place as you get older.

By resetting them and getting them back to the right place, boom, the cell is young again, the organ is young again, and suddenly you look, act, and feel young again. It’s an incredible technology. We’re just at the early stages, the early innings, of turning it into therapeutics. Again, the discovery goes back to 2006, and now we’re starting to see it get into the clinic.

All right. Let’s rejuvenate some hairlines on this podcast. That would be next up.

Speak for yourself.

I don’t know. You’ve got a little peak going there, my brother. Little peaks here and there.

What are you talking about, bro? My hairline’s incredible. I’m 50.

I mean, it’s not bad for 50. I give you credit. You’re holding your own.

8. Trump Pivots On Tariffs

All right, let’s talk about our final topic. SCOTUS struck down Trump’s emergency-powers tariffs last Friday. SCOTUS voted 6–3 against President Trump’s IEEPA tariff: 6 justices voted against it—3 conservatives, Roberts, Barrett, and Gorsuch, and 3 liberals.

According to Bloomberg, this is the biggest rebuke of existing executive policy in 91 years, since SCOTUS struck down FDR’s first New Deal in 1935. A UPenn Wharton analysis says the tariffs collected about $175 billion to date. Fifty percent of all tariff duties might wind up being refunded. This is going to take some time to sort out in the courts. 2,000 importers have already filed for refunds.

We talked about it here. I think the majority of people felt like this is the way the decision would go, and we talked about how there were other options for President Trump to pursue. He immediately said he was not deterred and invoked a 15% global tariff across the board via Section 122 of the 1974 Trade Act.

Here’s your Polymarket: Will the court force Trump to refund tariffs? There’s an 18% chance, but it spiked to 40% after the SCOTUS decision. How will Congress react? Polymarket says there’s a 3% chance Congress passes any tariffs by March 31.

So again, as I referenced earlier, these 2 sides just can’t seem to work together, and that would’ve resolved the whole thing. Sacks, do you want to give us your take here?

First of all, I don’t think the tariffs are going away. What the court basically indicated, especially in the 70-page Kavanaugh dissent, is that there are multiple alternative bases in law for the tariffs under existing law.

For example, Section 122 of the Trade Act of 1974 enables temporary 150-day tariffs of up to 15% to address balance-of-payments issues, and the president has already invoked this. So we are now operating under that.

What the 150 days is going to do is buy the administration time to substantiate, via studies and agency reviews, what it needs to prove in order to invoke more sweeping tariff authority under Section 301 of the Trade Act and under Section 338 of the Tariff Act.

Section 301 authorizes tariffs responding to unfair foreign trade practices. Section 338 of the Tariff Act allows tariffs against countries discriminating against U.S. commerce. The Kavanaugh dissent actually provided a roadmap for the administration to put tariffs in place using one of these alternate bases.

So I think that, one way or another, the tariff policies of this administration and the favorable trade deals that they allow us to strike with many nations will continue. I think the court seems to know that because the majority's opinion and concurrences have collectively said nothing about how the administration should go about refunding the tariff revenue already collected. I think that if they expected this decision to end the tariff policies altogether, they probably would have said something about that.

And I think that brings up a really important point just on the merits here: Why would we want to give back hundreds of billions of dollars to a bunch of importers when we're trillions of dollars in debt? I'll just say that the people who originally predicted that somehow these tariffs would be catastrophic for the economy—all of those predictions proved not to be true.

So I think that this is ultimately going to be a popular policy. The administration will figure out a different way to do it, and I predict that future administrations, whether they're Republican or Democrat, will keep some version of the tariffs in place because I think that they will ultimately be popular on a long-standing basis.

Chamath, your thoughts?

I think we've proven the experiment has been successful. What was the experiment? We needed to smoke out what the right balance of trade should be between the United States and all of its partner countries. I think what we uncovered is that, for the most part, there were structural imbalances that were made not because they made economic sense for America, but because it was just part of a hodgepodge of globalist drivel that people just bought into.

If you strip all that stuff away, we had a hollowed-out manufacturing class, and we have a hollowed-out middle class, and the tariffs will create more equality for the American worker in the end. So now I think the debate should be about how to implement these in a structural and permanent way. I think we talked about this before, Jason, that—

Yeah.

—this was sort of expected, and there are many other mechanisms. I think the president activated one of them immediately. I don't think this is going away, and I don't think it should go away. So I think now the point is Congress really should ratify these things because it is clear that it was the right thing to do.

And if they don't, then the president still has a lot of room to get these done. But these make smart economic sense, in my opinion.

Friedberg, any thoughts on the ruling? Does it give you some respect for the courts that they made a judgment not along party lines for once?

Yes.

Yeah, okay. Expand on that.

Yeah, and I think that all Americans should feel assured and comforted in the fact that a lot of people view the Supreme Court as having a high degree of partisanship. The fact that the president, despite having a majority of what others would think were politically aligned appointees on the Court, had a ruling that he did not want, I think should give everyone good faith that the system that the founders set up is working.

There is a judicial branch that adjudicates the law against the executive branch when they think that it doesn't map, and I think that that was very important to see.

So, clearly, the debate about tariffs, the economic effect of tariffs, the security, structural, and trade-relationship effects of tariffs, and the importance of that is a separate conversation. But I do think that the read of the law being what I would say is nonpartisan with respect to the court's action is important and probably very valuable.

I'll reiterate that. This is a great moment, I think, for the Supreme Court to make a thoughtful decision, and I think we need to think about executive power a whole bunch, whether it's Biden with student loans or Trump with tariffs.

We have this beautiful system set up by the Founding Fathers. I know it's frustrating. Gridlock is frustrating. Having to work together is frustrating. Trust me, we all come to this podcast every Thursday, and we have to work together. It's hard to work together.

But you have to learn to work together, and we don't want an executive branch that can unilaterally just roll over the other branches. And that's going to end. I think Trump's going to lose the midterms, and we're going to get to more chaos again. We might as well start this reconciliation process, with these 2 sides stopping their lawfare against each other and working together for the American people on the important issues.

On the tariffs, there are some fundamentally important things that Trump was doing there, and they were working. They could have been chaotic. That's a reasonable criticism of them because business owners didn't know what to do. So Trump did it in a chaotic way. That's just a fact. He should have done it in a more thoughtful way, and Congress should have been alongside him saying, “Hey, what tools do you need? How can we help support this? We know that there are trade imbalances. We know that people are being unfair. Let's work together as one America to negotiate these things.”

So both sides should start having dinner together, start playing cards together, and do what we do here on this podcast, which is fight it out and argue, but then come together and try to find some resolutions for this stuff.

So they should go and tell Trump, “Hey, we'll approve all the tariffs you did. We will not force you to get refunds.” Congress should come out and just say that. And then they should say, “Hey, when you want to do them in 2026, just run them by us or ask us for some parameters that you want, and let's just be thoughtful about it. These are our concerns.”

That's it. Thank you for coming to my TED Talk.

Can I just do one response on that?

Absolutely. I'm sure you have some debate-club points that you want to point out. Go ahead.

Well, I just want to make one point. Do you think Susan Rice is going to respect your call for comity and basically working together, kumbaya? She just—

I want the esprit de corps. No, I don't. I think both sides—

She just had a diatribe where she basically said that Republicans, and actually not just partisan Republicans, but even tech companies that merely were working with the administration, should expect to get prosecuted. She was basically outright saying that she and the Democrats are going to pursue lawfare as soon as they get back in charge.

They're absolutely going to do that. Just like when Trump got in, he went after Comey, he went after Jerome Powell. The lawfare is happening on both sides. Both sides need to drop the lawfare. We need to get rid of these pardons. They're ridiculous. And we have to be a team.

So let's just get some esprit de corps and teamwork going in Washington, D.C. And that's what we should vote for in the midterms. We should vote for moderates who want to work together. And in 2028, we should have some kind of moderates on tickets that want to work together. That would be better for all Americans.

This kind of chaos is not good, folks.