债券危机逼近？共和党放弃 DOGE，Google 用 AI 颠覆 Search，OpenAI 收购 Jony Ive 的 io

The bond market is the captain. Apparently, the Treasury Department sold $16 billion worth of newly issued 20-year bonds on Wednesday afternoon, and there was weak demand. This pushed yields higher across the board. The 10-year, which Friedberg has said to focus on because it's the benchmark rate for most borrowing costs—mortgages and stuff like that—spiked, and here it is.

It actually hit a 5-handle at one point. A lot of people were wringing their hands about this, and the S&P dropped 1.5% in about 30 minutes. Here's that chart as well. The 3 major indices were all down between 1.5% and 2% on the day.

Obviously, in related news, the House passed the Big Beautiful Bill at the 11th hour last night. This makes the TCJA tax cuts permanent, and it's estimated to increase long-run GDP by 60 basis points if all the cuts were implemented. It's also going to reduce tax revenue by $4 trillion over 10 years, according to the estimate, and it's going to add between $3 trillion and $5 trillion to the national debt over 10 years. Friedberg, what's your take on the BBB, the weak bond market—all of it?

I just want to give a quick primer on how the government gets funded. I know we assume everyone understands it, but I think it's important for folks to really grok it. For the federal government to make payments to employees and contractors and buy stuff, they need to put money in their bank accounts. The way they put money in their bank accounts is by issuing bonds.

These are Treasury bonds. They'll sell Treasury bonds to the public, and individuals buy them, companies buy them, banks buy them, and foreign governments buy U.S. Treasuries. They transfer or wire cash into the federal government's bank accounts, which it can then use to pay for stuff. The Treasury Department needs to continuously sell Treasuries to raise cash to fund the government.

When folks don't show up to buy Treasuries, that's a bad thing. That means the government needs to increase the interest rate that it's paying on those bonds. What we saw on Wednesday was a really weak demand signal for Treasuries on this modest Treasury auction. Selling $16 billion of bonds is not a lot; there are hundreds of billions being sold each quarter. This was not a big number, but there was no buyer. The market was really dry, and everyone who participates in financial markets saw this and freaked out.

The big thing to understand about the relationship here is, to your point, that this big tax bill and spending bill is passing out of the House. That bill, as we talked about last week, has a high deficit, and that'll run up U.S. debt over time, which makes it more difficult for the U.S. government to pay its bills because it has to issue more debt. Interest payments have to be paid every year, and so on.

Now, if you just pull up this one slide, this is something I thought was really worth sharing. The CBO estimates, which is what you referenced, Jason, are estimates of what the cost is going to be over time. This is from Jessica Riedl of the Manhattan Institute, who put this chart together. The Congressional Budget Office, the CBO, creates an estimate of the budget, the spending, and the deficit, and ultimately the incremental debt that the U.S. government will need to issue to fund its obligations over time.

This chart shows that, on a baseline basis, the expectation is that over the next 30 years or so, U.S. debt-to-GDP will climb to 203%. But what a lot of people don't know and don't talk about is that, in the CBO estimates, they assume in all of their models that interest rates are at 3.6%. With the bond selloff yesterday, what we're now seeing is 30-year Treasury interest rates at over 5%. It's 5.1% this morning.

For every incremental 1% above 3.6%, you're spending an extra $350 billion a year in interest. That's $350 billion a year. As rates go up by 1.5% over the next 10 years, we're spending another $5 trillion just on interest payments. That's half a trillion dollars to $1 trillion a year of incremental interest on the difference between 3.6% and 5.1%.

That's an incredible recursive problem. I say recursive because the incremental interest that we now have to pay—because interest rates just went up, because the market is demanding more payments from the government to fund this—means that the government has to issue more debt, and it quickly gets away from you. There is a nonlinear relationship between the deficit and interest rates, which drives up the debt problem in a nonlinear way. It gets away from you, and you can't fix it.

That's what the market is telling us: The current bill that's being passed out of the House is showing such an extraordinarily high deficit that the market does not want to buy the debt from the government. Rates are now climbing, and that creates a massive problem for the government.

Okay, Chamath, you've been tweeting about this. We thought we were going to get an administration, and maybe some focus from the government, that would bring more austerity measures and balance the budget. Now it looks like we're going to pour gasoline on the fire. What are your thoughts?

I think we have to be careful here. The political calculus will be for the president to decide how much credit he actually wants to take for this bill, because even though it has his name on it, the contents of the bill are, in actual fact, different from what I think he intended.

What I mean by that is, when you look at what happened at the 11th hour last night, it's disappointing. This thing is anti-DOGE. If DOGE was meant to be a reflection of the American voting population's desire for meaningful reform in government, cost controls, some form of austerity, and getting this debt spiral in check, this is the opposite of that.

What happened was, at the 11th hour, you had a handful of people abstain. You had 1 person who passed away in the last few days. You had 1 person who fell asleep on the floor of the House, so he wasn't even woken up for the vote. In the middle of all that chaos, all kinds of things were added, attached, and canceled at the last minute.

What happens is, you have to have these puts and takes, as Friedberg described. If you want to spend over here, you have to find a cut over there. But I think what happened was, there wasn't a lot of financial literacy used to decide what to actually put in and what to cut. That lack of discipline is going to create, I think, a negative set of consequences.

So what are those consequences? Today, the 10-year is around 4.5%. At the rate at which it's escalating since Liberation Day, by the end of this year, we're going to be past 5%. The 30-year is on a rate now to get past 6.25%, maybe even reach 6.5%. Those are way beyond what most people thought was a reasonable place to be for the United States economy.

What will the implications be as rates go to those levels? You'll delever from the United States.

You’ll sell U.S. debt. You’ll own things like gold and Bitcoin. If you’re curious about what’s happening to gold and Bitcoin, they started to spike in the last few days. You’ll have ratings organizations that add to this cascade by downgrading the United States. That happened on Friday.

You’ll have very smart people starting to signal that this is a much harder problem than they initially thought. That’s how I personally interpreted Elon’s comments over the last few days. So, what was the House supposed to do? I think what they were supposed to do was implement some form of austerity. They were supposed to, by the will of the people, pass a rescission bill.

They were given that rescission bill. It was just $9 billion. They couldn’t even pass a $9 billion rescission. Instead, they passed a $4 trillion inflation to our debt.

Now you hand this to the Senate. The Senate is in a very difficult place as well. Do they want to, quote unquote, claim victory and say, “Here you go, President Trump. Here’s your bill”? But they’ll further bastardize this thing, and it will be even further away from what I think benefits MAGA and benefits Main Street.

So who does it benefit? On the current course and speed right now, this bill is about traditional Republicans and traditional Democrats circling the wagon and putting on a platter a set of things that I think will be hurtful to average Americans. You’re going to see energy prices spike. You’re gutting the number of electrons that will be available for things like AI. You’re going to increase Medicare prices, and the math is wrong.

So when you sensitize this thing to a 4.5%, 5%, or 5.25% rate—meaning, not what the CBO used but the real conditions on the ground—this thing is an albatross. And I think, unfortunately, for President Trump’s agenda and for a MAGA movement, this is the worst of all conditions. The financial markets will punish this.

And then the last thing I’ll say is, now to top it all off, I think that Jerome Powell will see the writing on the wall. Many aspects of this thing are inflationary, and if they’re not handled well by the Senate, he has a lot of room to actually increase interest rates.

I would just say that the Senate has an incredibly difficult job. I think the president has an even more difficult job about what to do right now. But the House did nobody any favors. They did not do anybody a favor yesterday.

Here’s the prediction from Polymarket on what will happen to Fed rate cuts in 2025. It’s across the board. It seems like nobody can tell if it’s going to be 0, 1, 2, or 3. The probabilities are 20%, 22%, 25%, and 15%, respectively, for those 4 options.

Sacks, you don’t speak for President Trump. You’re the czar of AI and crypto specifically, so let’s put that out there. But I’m curious about your thoughts on what happened with DOGE. Why are we continuing to spend ourselves into oblivion, and what will the downstream issues be here? Because this sounds like it’s going to be inflationary. This sounds like it’s going to be hard for people to buy homes. It’s going to create more austerity in the future. How bad is this? Do you feel it’s as bad as Dave and Chamath are framing it?

Well, look, I wish we had 435 members of the House who thought like Friedberg about spending and deficits. We don’t. The Democrats all want a lot more spending and a lot more taxes. Remember, if it wasn’t for Manchin and Sinema, we would have had that $4.5 trillion Build Back Better bill.

Among the Republicans in the House, unfortunately, we have a bunch of Republicans who are pretty soft on spending. You can call them RINOs if you want. We only have, what, a 3-vote margin in the House, and the Democrats are not cooperating in any way. So you don’t have that many votes to spare.

What I’m saying here is you can’t make the perfect the enemy of the good. I think you have to be realistic about how much we can get done here. I do think that this bill contains a lot of good things in it. There are a lot of priorities for the administration in here.

Do I wish it cut spending more? Yes. Do I wish that it made all the DOGE cuts permanent through rescissions? Absolutely. I think it’s outrageous that there were enough House Republicans who didn’t want to back up DOGE that that wasn’t enacted.

But there are a lot of good things in the bill, and I should just highlight them for balance here. The number-one thing is that this bill will permanently extend the 2017 tax cuts. If we don’t pass this bill, let’s just assume that we do nothing. If we do nothing, you’re going to get the largest tax increase in decades, and it won’t just affect the rich. This affects the middle class as well.

The 2017 tax cuts increased the standard deduction, which is important for the middle class. They raised the child tax credit. Again, if you do nothing, you get a huge tax increase as soon as the 2017 tax cuts that Trump passed back in his first term sunset. So that’s number one.

In addition, there are promises here that were made during the campaign to eliminate taxes on tips and overtime that have been enacted. The president promised to do those things, and he is accomplishing those things. So I think that’s important.

There’s funding for the border wall here, 10,000 new ICE officers, more detention beds, and so on. On energy, it repeals a methane tax and unlocks new oil and gas on federal lands. There are a bunch of other things as well, so there are a lot of good things in here.

Now, to your point, will this bill be bad for the Republicans’ prospects in the midterms?

I guess what I would say to that is, I don’t understand how you expect to do well in the midterms if Republicans preside over the largest tax increase in decades—a tax increase that Joe Biden tried to get during his administration. Biden tried to raise taxes and couldn’t get the votes for it.

But if we don’t make those tax cuts permanent, if we allow the sunset to happen, then you will accomplish what Joe Biden could not, which is to raise taxes on the American people. No bill like this that’s passed by a 1-vote margin on straight party lines, when you only have a 3-vote margin in the entire House, is going to satisfy everyone here.

Well, let me ask you a question there. You have Trump at his peak power. You have all these threats of, “Hey, we’re going to primary anybody if they don’t do what we’re saying here.” Why isn’t Trump just coming in and using his power and saying, “Hey, not good enough. Get back to work. We need to cut spending. Period. Full stop”? He’s got that power. I don’t hear him saying that.

I do think that there are ways to enact the tax cuts and make them permanent, and I think that that is a good part of what this bill did. The problem was what happened from 11:00 p.m. to 6:00 a.m. That’s all of the nonsense where, I’ll just be honest, I think the people in the House are not nearly as financially literate as they need to be. I think what they wanted to do was pass something, throw it over the wall, and say, “Mission accomplished.”

Now it’s up to the Senate. The good news is, the Senate has these 6-year election cycles for a reason. It allows them to think strategically and see past what’s right in front of them. So I still think that there’s a chance for the president to have his cake and eat it, too.

But I just want to highlight that, as a supporter of his and a supporter of that MAGA agenda—which I think a lot of people voted for—I want to be honest and say this is not it. There have to be some meaningful reforms in the Senate version of this thing that gets sent back to the House.

I’ll give you an example of one. Eighty-one percent of the incremental energy that was generated in the United States last year came from private enterprises that were investing in short-term forms of power that have a tax credit and a transfer market tied to them. This is just the financial machinery. This is how Blackstone and Goldman Sachs and Brookfield and all of these big players move tens and hundreds of billions of dollars around.

They changed those rules at the 11th hour. What is the impact of that? Elon foreshadowed it. Nick, maybe you can show the headline of the article. He said, “We could be running into a power capacity issue by mid-next year.”

Now, if we knew that, why would we take the short-term incentive away to generate more electrons? All the great things that David has done, that the president has done on AI—I mean, we’ll talk about what happened in the Middle East—but when the rubber meets the road and you actually have a shortage of electrons because of these financial actors, they’re not—and in fairness to them, they’re acting rationally. You take the financial incentive away and you can’t underwrite this thing. What are they going to do? They’re just going to stop doing it.

So in the absence of electrons, what happens? Prices go up. It’s inflationary. Now you’ll have to allocate electrons. Do the electrons go to an AI data center, or do they go to a home?

In a different example, actually on the residential side, places like Florida—I think there was an incentive for batteries. So if you’re about to get hit by a hurricane, you can buy these big battery systems from Tesla, as an example. But there are other ones, Generac and whatever other brand, and that was cut out.

So in Texas, all of these impacts will have negative consequences. My point is, I think you have to really look at what happened in the 11th hour and fix those things, because the problem with these plans is you can’t look at gains that are in years 6 through 10 to justify pain in years 1 through 5. You can’t do that.

I understand how that mathematically makes sense. It doesn’t make sense practically for the average American. That’s just not how life works. You don’t go into a J-curve. We do that as investors, right? But you don’t do that as a normal American, which is be on the wrong side of the economy for 4 or 5 years in the hopes that your 6th and 7th years are good.

That’s the point that has to get fixed. I asked Chamath the question: shouldn’t Trump use the capital he’s built up here and say, “Not good enough. Make some cuts here”? What would you say to President Trump if he was asking you for your advice, Friedberg?

I think the writing’s on the wall that the system is winning. The rebels came in and tried to fix the inefficiencies in government with DOGE under Elon’s leadership, with our friend Antonio’s participation, and with a great group of individuals who are true patriots spending time trying to help reassess how the government operates.

As I’ve said from day 1, from my first visit to D.C. during inauguration weekend and every visit I’ve done since then, I always leave after meeting with members of Congress deeply worried that they won’t actually do what they need to do to make the changes. The last line of defense, ultimately, is the president, with his veto authority and his ability to tell people to go back, rethink this, and cut spending.

I feel like we’re in a last-ditch effort. I’ll tell you why I say it’s last-ditch: I think there’s a relationship between what’s going on in the United States and what’s going on in Japan. Nick, if you’ll pull up the Japanese government 40-year bond-yield chart, just so you guys can observe it, in the last couple of days we’ve seen Japanese bond yields soar from 2.5% to 3.5%.

Remember, Japan owns $1.1 trillion of U.S. Treasuries. I’ll just read a quote from a guy who’s the chief global strategist, named Albert Edwards, at Société Générale. He said, “Recent developments in Japan may signal the end of the favorable investment cycle that’s been established since the 2008 global financial crisis, potentially reshaping global financial landscapes in the coming months.”

Can I translate that? Because it’s incredibly important what Friedberg said, but I’ll translate what he is saying. So, this is a guy from SocGen saying that since 2007–2008, every now and then the markets give you free money. One of the most obvious free-money trades was the yen carry trade.

In order to execute that trade, you’re effectively leveraging Treasuries against the Japanese bond market. What he is saying is that there’s increasing risk that that trade unwinds. When that trade unwinds, what you’re going to have are net sellers of up to, as Friedberg said, a trillion-plus dollars of U.S. Treasuries.

What will that do? It will further exacerbate the market’s reluctance to own U.S. Treasuries, which means yields go even higher, which means interest rates for businesses and individuals go higher. Correct. And the only way to keep those interest rates low will be for the Federal Reserve to try and step in and buy those bonds, which will devalue the dollar.

In his commentary, he went on to say that the 20-year Japanese bond auction that just took place, similar to the U.S. Treasury auction that just took place, was the worst they’ve seen since 1987. That’s how significant this event was in Japan over the past week.

So, the massive climb in bond yields in Japan indicates a massive sell-off in credit. Remember, if you look at the historical perspective, when the global reserve-currency nation sees its debt sell off, it usually is part of a global sell-off. It’s not just the U.S. that gets affected. Any one of these markets can cause a cataclysmic follow-on to the rest of the global financial markets. If Japan sells off too much, we are going to start to see a lot of unraveling happening.

That’s the consequence, I think, of this particular bill, because it signals a lack of willingness to actually address—

I called Friedberg today on the way here. It’s very rare that he and I agree about this kind of stuff, because normally what I say is, “You’re catastrophizing. Sure, calm down.” But this is where I will say I’ll put my money where my mouth is, and I’ll put the chips on the table to say this is the moment.

This is the moment where we either re-underwrite what we voted for or we mistakenly pass this Christmas-tree concoction. It does have President Trump’s signature on the top of it. It does have some features that he asked for. But what I hope he sees and realizes is that the details here matter.

Without some technical maneuvering in the Senate, this is a bad bill as constructed. If it passes, forget what any of us think: the bond markets will act decisively, and they’re going to go in one direction. It’s away from us. I just hope that we figure this out.

Seems like a classic diffusion of responsibility here, Sacks. You have Congress, the Senate, and POTUS not willing to step up and say, “Let’s cut the spending.”

Well, I don’t think that’s fair. Look, the president backed up Elon and DOGE. First of all, he created DOGE. He gave it the U.S. Digital Service. He backed it up in every Cabinet meeting and in every public pronouncement.

There’s a limit, though, to what he can do. Congress has the power of the purse. We said on a previous episode that, at the end of the day, for the DOGE cuts to become permanent, Congress has to incorporate that into the budget. The reality is there are still a number of these old-bull appropriators on the Republican side who just are not wired to make spending cuts.

Why aren’t we hearing him vocally say, “Stop this madness”? He is the most vocal guy in the world who mixes it up on every topic on the planet. Why isn’t he saying it?

Let’s give him the benefit of the doubt. He’s probably working it all in the back channel. There are things that he wants, and he’s in the midst of negotiating as well.

But my point is just this: the level of financial illiteracy in this bill will come back to bite America in the ass. Here’s the problem: what you guys want to sell right now is austerity. The reality is that the politicians in Washington just are not buying austerity right now. I mean, that’s what you want—is austerity.

Yep.

And the reality is the only way you get true austerity—a painful set of changes—is when both parties link arms and jump off the cliff together and make those changes. The Democrats right now have no willingness to do that. They’re opposed to doing this.

This is a straight-line party-reconciliation vote. I think you’re trying to overload this vote with way too much stuff. You’re trying to basically say that this vote has to fix America’s fiscal problems. It’s just not going to happen.

I agree with you that the fiscal situation needs to get fixed. This is just not the vote where it’s going to happen.

You cannot exacerbate it.

Well, I don’t know. I mean, look, I understand that this doesn’t have as much stuff as we want. Like I said, I want to see the DOGE cuts made permanent. But the alternative to this bill—let’s say the alternative is just doing nothing—is actually meaningfully worse, because you will get a massive tax increase. Massive.

I think President Trump inherited a Biden economy that was already feeling pretty shaky. I don’t know what the argument is for a massive tax increase in an economy that’s not doing great. I’d rather see the tax cuts made permanent. I’d rather see the no tax on tips and overtime implemented—important campaign promises. I’d rather see funding for the border wall. These are all important things that we need to have happen. I just don’t think that this bill is how you’re going to start.

Good. But what I’m saying is, why do you pay for that with electricity supply for homes and AI? Why do it by doing Medicaid cuts? Why?

I don’t know. I mean, you’re talking about a pretty esoteric issue, Chamath. This is not the same as solving the U.S. budget situation.

I’m not saying that. I think all cuts are great. Cuts across the board.

Yeah, cuts across the board. You’ve got to decide what you’re opposed to. Are you opposed to cuts, or do you want more of them?

I want more cuts. Given the fiscal situation, if someone showed up and said, “Hey, we’re going to increase taxes,” I’d say, “That’s what we’ve got to do right now. Everyone’s got to take a hit on this.” But I know I’m going to be alone in that, so I’m not going to try and make that argument.

You’re not alone in that. I agree with you. Well, let me ask a question, Sacks: what do you think is the moment when the 2 parties lock arms and jump together? What are the conditions that need to be prevalent for us to take draconian austerity measures?

Probably the bond market forces it on the politicians, forces it on Washington.

Yeah, I think you’re right. You’re an economist by background. Do you see that happening right now? Do you see this sell-off that’s happening not just in the U.S., but everywhere that’s linked to the U.S., with Japan and so on? Don’t you think the market’s telling us that this might be the moment?

It could be the moment, but I remember the first week of April, when Larry Summers came on the pod and said that the markets were melting down. That was a different moment that he was trying to sell in favor of his policy. So my point is just that there’s market volatility, and we won’t know for a while whether this means anything. It could just be a blip, right?

You ask me what the conditions are for getting austerity. I think that austerity will have to be imposed on Washington from the outside. Washington’s not going to find the will internally.

We had a wonderful moment here where you have Elon come in with DOGE, which the president made possible. No other president would have even done this. He let a total outsider come in, bring a team of young, big, bald geniuses, and let them go through these departments line by line and start deleting things.

They made, I think, great progress. I think they found $160 billion a year in cuts that can be made administratively. Beyond that, they need Congress to approve it. And I will acknowledge that there are still a number of soft-on-spending RINO Republicans, and there's just not a majority for the types of austerity and cuts that you guys would like to see. I mean, that's just the reality.

But I still think that the question with this bill is just how pure you want to be. My point is just: Are you going to make the perfect the enemy of the good? I don't think you're going to get austerity out of Washington right now. It's just not in the cards. The question is whether this bill is basically better than the status quo, and I think it is because of the tax cuts.

To your point, without opining on it, I think what's going to happen is the bond market will sensitize this budget at real rates. They're not going to use the 3.6 that Friedberg pointed to. They're just going to re-underwrite this at 5%, 5¼%, or 5½%. That has a very different risk exposure, I think, and so, to your point, they'll react. If it looks like it passes as is, then that's what they'll do. And, to your point, David, we'll know in the next 60 to 90 days.

Let's say that this is just a blip in the market and it's not some larger, once-in-a-century-type event. The reality is we don't know exactly what's going to happen. Look, I don't like America's fiscal picture at all. But what if the AI and robotics revolution plays out in the most optimistic way over the next decade and is massively deflationary, and we get basically AI-powered robots expanding the economy massively? You need power for that. I get that, and we need to do power generation for sure. But my point is just: What if the new technology provides an answer to the fiscal situation that's not currently on the table?

I think, again—so this is my point—Washington's not going to embrace it.

I understand, but you're going to have brownouts and blackouts in average human homes to do that.

No, you shouldn't have to make these decisions. Those are dumb trade-offs. You're not going to be able to solve the power-generation problem in a reconciliation bill.

I want to just add my 2 cents at the end here. Leadership starts at the top. This president acts unilaterally all the time. Whether it's immigration or DEI, he has no problem. And this is what he is doing: He's endorsing this bill. He's saying it's big. It's beautiful. He is saying this is historic. I think this is bad leadership. I'll just say it straight out. He was elected to balance this budget and to have austerity. He's going to put more onto the debt, and we sat here on this very podcast, the 4 of us, and said this was the most important issue in the world. Trump is putting gasoline on the fire. I think it's bad leadership. Let's continue on the docket.

I suggest you look at the Constitution, because the power of the purse rests with Congress.

He's endorsing it. He's not even putting up a fight. When it comes to immigration, DEI, and these other areas, the president has far more unilateral power, and he's used that every chance he gets. But when it comes to spending, you have to get Congress on board. This is a bill that passed with a 1-vote margin. I don't know where you think the opportunity is to squeeze out more concessions from Congress.

I don't think he could speak up.

He could speak up and say, "Do better." He's doing the opposite.

This bill passed by a 1-vote margin. It was hanging on by a thread. And if it failed, you get basically hundreds of billions of dollars in tax increases, which we don't need right now. In terms of Republican prospects, look, people vote for Republicans to do 2 things: cut taxes and cut spending. I wish that this bill did 2 out of 2. It does 1 out of 2, with some of the second one, like the Medicaid stuff. If we didn't at least cut taxes, there'd be no reason to vote for Republicans. None.

You all are much closer to him than I am. You all should be challenging him not to endorse this and to do the opposite. That would be the profile in courage. Then we get the biggest increase in decades. Great. No, he should just go to everybody and say, "Do better."

I don't agree with that tactic. I think the better tactic is to do what he did, but then work behind the scenes and help the Senate implement the technical guardrails that make the bill better than it is as is. Look, nobody thinks that this bill is going to pass as is. It's going to go through its own set of changes in the Senate. Now, I think the opportunity is to understand these puts and takes with a little bit more detail and a little bit more mathematical precision and fix it. So, I still hold out hope that we're going to get to a better place. But he's going to have much more influence, Jason, by actually taking the positive path than throwing everybody under the bus.

The real question will be, in 30 or 45 days, what is the final version of this thing? And whatever gets passed and signed by him, there will have been enough time to understand it. And you'll see the real-time reactions. All I'm trying to indicate is, I think that this was a moment where people were very pensively optimistic that the approach of DOGE would really be embraced and that we'd see some broad-scale attempts to at least—and Friedberg said this, just to go back to 2019—even if we don't do anything else, just go back to 2019. Pretend COVID didn't happen.

Yeah. And we would be, by the way, if we did that, we would have a budget surplus, and the bonds would be trading down to 2% yields again. I mean, we would be in such an incredible condition. We've got 2% inflation right now.

Anyway, we don't need to beat a dead horse. Let's move on. I wish there were 218 members of the House who thought like you.

I hear you.

Oh, gosh. Listen, this guy is saying he's going to primary anybody who doesn't support him. He should be speaking up right now. I'll tell you what: Give him the line-item veto and then let's talk, right? How about saying, "Guys, spend less money?" He doesn't have to throw them under the bus, Chamath. It could be something very subtle. But you make it that way. The way that you position it is like you make it such a black-and-white thing, and "I speak up" is black and white. There's a lot going on.

What I'm saying is, he is—I'm guessing, I don't know, but I'm going to assume—working behind the scenes. And what you're saying is he should virtue-signal and prognosticate in public.

I never said virtue-signal. That's what you're asking him to do.

No, I'm not.

I'm asking him to say, "Reduce spending in the bill, because that's good for America." This bill cuts $880 billion for Medicaid over a decade, which is already politically tough and controversial. It imposes work requirements for able-bodied adults. This is similar to what Bill Clinton did back in 1996 with welfare reform, basically saying that you can't be a layabout and get welfare. So, these are relatively tough things to do politically. Do I want to see even more cuts? Yeah, absolutely. But again, when you have a 1-vote margin, whose vote are you going to pick up by doing more cuts? I guess you get Thomas Massie, maybe. But you probably have to go all the way to balance the budget to get his vote. I don't know what it takes to get a Thomas Massie vote. So, I don't know what votes you think you're going to pick up by cutting even more. And we can't afford to lose even 1 vote. It's literally the art of the game.

We've got to move on.

I will move on. But if you're going to tell me that I'm asking him to virtue-signal, you're going to lie about my position, and I will correct it. It's been 40 minutes. My position is not to virtue-signal. It's to go flip a couple of people and say, "Do better."

It was a huge week for AI again. Google and the ghost of Steve Jobs took center stage. Let's start with Google here. They had their I/O conference on Tuesday. This is where they show a bunch of new stuff and get developers to come together to embrace their products. The stock ripped 5% on the day, which might be a turning point for Google, and it's up again today.

Search was the big announcement, and we had a discussion with Sergey about that in Miami. We had multiple discussions about it here over the last couple of months. They demoed something called AI Mode—no, not Founder Mode, AI Mode—and they compared it to regular Search. Here it is on the screen if you're watching us on YouTube or Spotify. It's a very elegant app.

Here you have somebody searching, and what you can see is that in AI Mode, it looks like a nice, comprehensive search instead of 10 blue links and tons of advertising. You can flip over and see all, or you can do a Perplexity-like search. Here we're showing you what this search would have looked like: the AI summary at the top and then your 10 blue links. It's distinctly different. It is exactly, Chamath, what you talked about. Somebody has to have the courage to flip the switch here. And so, that was the big drop.

I think that in your interview, Dave, didn't Sundar say we're going to integrate ads into that kind of result?

Yes, the comprehensive search result. I think there were a couple of things that Sundar said in the interview that showed up in a really important way at I/O. The first was exactly this, which is to make AI Mode more ubiquitous in Search and effectively, over time, replace Search with this AI Mode experience, which they've been testing in a small group and now they've expanded it. Very specifically, to Chamath's point a couple of weeks ago, they should try and flip the switch. Well, it appears they've done that.

The key question and the challenge has always been: What's the revenue per query? How are you going to make money? And I think one of the other interesting announcements that we saw come out of I/O, which indicates the business-model opportunity here, is not just AI Mode and Search but some of the other tools that they launched, bundling them together. And they have this product offering called AI Ultra for $250 a month that includes YouTube Premium.

It includes 30 terabytes of storage on your Google account. It includes access to Flow, their movie-creating AI model, where you can use Veo 3, Gemini, and Imagen. Those are the 3 models that contribute to this movie-creation studio tool. They’ve launched a number of other high-value Gemini models as standalone research apps, and you can get access to all of these high-powered tools for $250.

I do think this sets a new direction that Google is likely testing as a business model: subscriptions. They’re asking, “Can we make a very high-dollar-volume subscription model for consumers that, over time, could create a very meaningful shift in Google’s revenue mix away from ads and more toward subscription revenue?” We already see that in some of the consumer services like YouTube and YouTube TV.

This is a really important turning point. I would say if anyone were to identify the week that Google really pivoted into the AI business model, it might be this week. They launched over 15 products at this thing. Another important one I’ll highlight: Sundar talked about this in the interview I did with him.

In the early days of Google, from 2002 to about 2011, Google had this kind of playground of products called Google Labs, where they would introduce new stuff. I don’t know if you guys remember checking out new stuff. They relaunched Google Labs about a year ago, and as Sundar said, they’re putting a lot more stuff in Labs.

Labs is now becoming the new test bed, and they launched a bunch of the announcements in Labs. That becomes the place where they’ll say, “What’s the business model? What’s the use case? Do people like it? Do they love it?” If they do, it graduates out of Labs into full production. I think that opens up the opportunity for some of the things Chamath was talking about, where they could experiment with new ideas and new modalities, and then productize them if they work. They discontinued Labs in 2011. They just brought it back last year.

They just brought it back. Yeah. Sacks, you mentioned that they have the “I’m Feeling Lucky” button here. They put the AI search up in the top bar. We have Images, Video, Shopping, and News. We’re all familiar with that. What do you think? Would you just swap out “I’m Feeling Lucky” for “Delight Me” with AI mode, or maybe, if you were a product manager there—you’ve done a lot of great product work in your career before you were in politics—would you just A/B-test this and say, “Anybody who, let’s say, doesn’t click on ads, let’s send them to the AI search and just see how they do”? They don’t click on ads anyway, so just let the ad clickers keep going to the ad product.

I think we were all calling on previous podcasts for Google to risk disrupting its dominance in search by moving to more of an AI-based model. I think they’ve taken an important step in that direction. They’ve threaded the needle here between keeping their old UI and product and also incorporating the new sort of AI UI.

This is a compromise. They’re trying to have their cake and eat it, too. It ultimately feels a little bit like a transitional move. I doubt that this will be the endpoint, but given their need to protect their search business while also developing their AI business, this feels like a pretty good compromise.

It certainly shows that they’re in the game and they’re not going to get caught totally flat-footed and let the whole world disrupt them while they’re trying to figure out AI. They are responding now, and the market seemed to like it. I think they were up about 5% on this news.

Yeah, and then up again today. Chamath, you talked about having some design sense and maybe some bravery here. It looks like they’re pretty much on the cusp of doing that. The search results in AI look elegant. They look, dare I say, Perplexity-like when you look at them. I would like to have that as my default when I’m doing search. What do you think of the product? We’re showing it here again. And what do you think the roadmap will be? I gave one possibility.

Sacks is totally right. I think this is not the destination, but they’ve taken a really important step, and now they have to follow through. We all know what has to happen, so I think we’re all just going to debate when it happens. What the market is betting is that it’s going to happen in the next 18 months.

What is it exactly? It’s when AI Mode becomes the default for a large swath of existing Google users. How will we know that happens? They’re probably running A/B tests right now. They’re probably gauging the behavioral patterns of which kinds of users, what the actual impacts to search are, and what impacts exist to CPCs, or cost per click, which is one of the inventory types that they sell.

I think this is the beginning of a process. I suspect it’ll be done in less than a year. For a large swath of users, they are going to put AI in the front door. I think that’s a fait accompli. We’re just debating the mechanics of getting there, as Sacks alluded to. That’s a really powerful step.

But here’s what they also have to keep in mind. What you don’t want to have when you’re going through an innovator’s dilemma is to have the competitor be onto the next lily pad while you’re figuring out the current lily pad. The reason I bring this up—and I’m sure we’ll talk about this—is that it’s impressive what Sam Altman is doing right now at OpenAI.

I don’t know if it was a troll or not, but OpenAI bought io, Jony Ive’s firm, at the end of Google I/O. I read an article that said OpenAI had been trying to do this for a couple of years. What Sam alluded to in that video was some next-generation, whiz-bang device that they’ve been working on. That was the industrial logic for the M&A.

I think what you don’t want to have happen, if you’re Google, is to finally get the search thing right and then have to play catch-up on some device. I would encourage them to just run the A/B tests and shorten the window of measurement. We all know what has to happen, so it’s probably better to just do it sooner rather than later and then start allocating the incremental resources to these new form factors and other things, so they shorten the distance between them and OpenAI.

Let me tee up exactly what happened. OpenAI, the makers of ChatGPT, bought Jony Ive’s startup. The startup is called io. They bought it for $6.5 billion in an all-stock deal. If you remember, Jony Ive was Steve Jobs’ collaborator and designer: iPhone, iPod, iPad, all that beautiful stuff.

OpenAI is buying io. io is not LoveFrom, if you’ve heard of that. That’s his design agency. They did a really interesting rom-com video. I don’t know if you guys saw this. It’s incredibly ridiculous. Let’s play it.

Here are 2 people having a meet-cute. Dude, you’re right. It’s like a love story. It literally is like a meet-cute. Here comes Sam. He’s walking around San Francisco, and he’s a genius, according to this video.

Here’s what I have said about Jony: “I’ve never met a man who’s so brilliant and so humble.” I mean, it’s embarrassing, the language. “The responsibility that Sam bears is actually, honestly, beyond my comprehension.” He’s a co-founder of io. Yeah, OpenAI already owns 20% of it. Yes—no, no, 23%.

As with all Sam Altman deals, if you’ve learned anything, they’re conflicted. But the copy here—I just want to get your general comment on what I see you worrying about: other people, customers, society, culture. To me, that tells me everything I want to know about him. I’m surprised you guys are such haters on this. I’m excited to see what this thing is.

I just thought this was like—you’re falling for it the deepest. I don’t know why we’re even watching, but they’re going to watch the 9-minute video. I wanted my 9 minutes back. Nothing happened.

I want the device. They say they have a device. Okay, I’m giving them the benefit of the doubt. They’ve invented something that they say is world-changing. Okay, but this is, again, I’m just using their words—and he’s sipping a macchiato there. So that’s important.

Oh, my God, you are so falling for the espresso video. This is your problem.

Cham, look at the glasses. I think the clue is in the video. Look at Jony’s glasses.

The fact that we’re even talking about this is such a problem.

Let me tell you my 2007 Apple iPhone moment. Steve announced the iPhone, and I got one of the first few hundred or few thousand. I was going to Vegas, so I got it on a Friday and had it in my hand. The way that people looked at me for those 3 days—I’ve never been looked at like that ever before or since.

They liked you. People liked what you were holding. It was love and admiration.

All I had was music. There were no multiple screens. You couldn’t buy apps in the App Store.

But it was so mesmerizing. My point is, you’re right: I may have fallen for it, which may seem crazy, but when I saw that video, I thought, “Hey, hold on.” All I’m saying is, if there is a next-generation device that they—or somebody else—are working on, my only comment is that I think Google is so wonderful. I love the people there. I just want them to be on the front foot, and I just don’t want them to be reacting for the next 5 years. That’s all.

Here’s what Sam said about Jony: “Jony is the deepest thinker of anyone I’ve ever met.” I mean, this was a love story. I’m not justifying what they said about each other.

Okay, let’s handicap it here. I’ll tell you exactly what’s happening. It’s the glasses. He’s wearing the glasses in there. They’re making smart glasses.

Sacks, what are they making? What are these 2 guys up to? You know these guys. You’ve known both of them for 20 years. Come on now. You know what’s going on. Just spill the beans.

Let me spill it. Let me express a hot take here, but let me express it in the form of a question, because I don’t want to be too definitive with the rhetoric.

Okay, go: Is Jony Ive just a media creation? I mean, seriously, did Steve Jobs manifest him? I know he did good work at Apple, so don’t get me wrong, but the software was done by someone else, right? He did the beautiful industrial design. The hardware looks beautiful with a beveled edge, a painted glass with a beveled edge. Great.

He has this beautiful English accent, which everyone associates with intelligence, right? So he could describe anything and it’s going to sound amazing and taste 15% smarter, right? I don’t know. Maybe he’s parlayed this incredible reputation that he’s built in the media into 2% of OpenAI. It’s incredible. What a trade. Do you think it was done at that $300 billion price? Do you think it was done at a cheaper price?

Yes, yeah. I mean, there is an argument. I’m going to go in Sacks’s corner here. I’m going to go into Conspiracy Corner, and I’m going to say Dieter Rams designed the iPhone, not Jony Ive. Pull it up, Nick.

Well, I’m going to say Tony Fadell designed the iPhone. But if you don’t know who Dieter Rams is, if you look at the transistor radio he made, it’s the iPod. If you pull up the Braun radio, here is the Power Mac. If you look at the iMac, basically—

Are you accusing Jony Ive of ripping off a German industrial designer?

What I’m saying is, immature artists copy; mature artists steal. I’m saying 90% of what Apple did was steal Dieter Rams’s design. I will die on that hill. Ninety percent stolen. Jony Ive has been knighted. He’s Sir Jony. He took Dieter Rams’s knighthood. That’s what he did. Sir Jony.

I mean, Sir Jony did put the handle on an iMac, right? So you’ve got that. Stop, you guys. This is so brutal. So great, dude. Look at this one. Pull this one up, Nick. Everybody knows the iconic iMac. I hate to break your hearts, folks, but it was totally ripped off from Dieter. It was all inspired by the Germans. Sorry.

So you’re saying Dieter Rams is like Keyser Söze of our entire physical world?

Look, here’s the speaker on the left. Look at the iMac on the right. These guys took every design Dieter did in the ’60s and ’70s, even the ’50s, photocopied them, and put an Apple logo on them. I couldn’t stop. Sacks, am I wrong?

I didn’t know about the Dieter Rams angle. Listen, it is what it is. Everyone at Apple did amazing work; there’s no question about it. But I don’t quite get this design-god reputation that’s been created, a little bit. I would say that it’s certainly a brilliant idea to parlay that into 2%.

By the way, is that what San Francisco looks like? I never thought—every time I go to San Francisco, it’s so disgusting, and it’s cloudy and cold. Did they find the one day where the sun was shining in that hellscape?

Yeah, there were no homeless people, and there was no fentanyl. I don’t know what happened. Yuck.

On a serious note, maybe what this deal indicates is that OpenAI is going to do something with hardware. It could be a new kind of phone that is AI-first, as opposed to the wall of applications. Or maybe it’s some sort of pendant that tracks what’s going on around you and feeds it into AI so it can give you more context, maybe like in the movie Her. Who knows? But it seems like they’ll do something with devices in the physical world. So that would be the play, I guess.

I guess the question you have to ask is, will it make OpenAI 1% more valuable? If it’s a $300 billion company, 2% more valuable. Yes, thank you. Will it make it more than 2% more valuable? If it makes it 1% more valuable, it was worth the swing, I guess, right?

I heard an interesting factoid yesterday from somebody who was looking at OpenAI equity. All of the rounds before the $300 billion round from SoftBank had essentially a feature where there was a hurdle rate of return and the ability to exit or something. No, but the ability to put the stock back if they weren’t able to complete the conversion to a for-profit company, et cetera, et cetera.

In the $300 billion round, apparently it’s much cleaner, and it’s essentially like a sub-NewCo that is owned. I thought to myself, at those valuations, I would have expected the terms to ratchet up. Instead, it seems like the terms have ratcheted down. That’s either commentary on the negotiating prowess of OpenAI and its team, or the risk-management philosophy of SoftBank—and the lack thereof.

I do think he is a great designer, and he will make it 1% or 2% better. Design matters. There’s a chance, right, Sacks, that he could make the product 5% or 10% better through design.

I don’t think it’s going to be 1% or 2% better. I think it’s going to be either zero or a doubling of the company’s value, something like that. Either he develops an iPhone-like, amazing hit product—some new consumer device no one’s seen before that really could take the company to another level—or probably they do something that doesn’t work. So it’s more of a 2% option against a strike of doubling the value of the company.

I’m surprised it hasn’t leaked, because in Sam’s update, or whatever, he said that the executive team was shown the device and that he has it. Who? So what is it?

It’s definitely going to be a pendant or glasses. It’s got to be one of those 2 form factors.

He was the investor in Humane, remember? Wasn’t that the pin from the Apple people?

Yeah. I will take the field against that.

What are the odds, and how much money do you want to bet?

Okay, here we go. I think it’s a pendant, like a wearable. We’d have to define that—so, like Humane. I’ll say it’s not a pendant.

Okay. Glasses. I get it. What are you going to give me? You give me odds. I’ll give you—

Give me odds.

We’ll just do even money. Let’s do even money.

Even money? I don’t know. Whatever you want. However many thousands of dollars you want to bet.

How about you give me 2 to 1, 10 dimes? Because you have every option; I have to pick them.

No, no, no. That’s a pretty good bet.

This is my—You want to get in on the bet here? All right, let’s move on. We covered Trump’s trip to the Middle East last week, but our bestie Sacks was there as well. You got a couple of photos to show, and you met a lot of friends there. Tell us about the experience. Did you go, first of all, on Air Force One? That’s what everybody wants to know. Or did you go on AirSacks?

I took my own plane. I know that sounds a little bit like a flex, but—

Were you in formation with Air Force One? Were you right behind it, or were you on your own travel schedule?

I was on my own travel schedule.

Did you have the option to go on Air Force One?

I probably could have if I wanted to, but I actually wanted to get to the region a little bit early because I hadn’t been before. I wanted to check it out, so I was waiting there when the president arrived.

Oh, wow. So it was your first time in the region?

Yeah.

Tell us about your impressions generally. What were you trying to accomplish there as the AI czar?

I call it AI diplomacy. First of all, these countries are obviously very resource-rich. They’ve got a lot of capital to deploy, and they’re also very interested in diversifying their economies. They’re very interested in high tech, and they want to do things in AI. They’ve got very significant aspirations there, so I was over there to listen and learn, see what they’re interested in doing, and see some of their tech scenes.

Oh, you went to Digital Garage in Riyadh, right? I’ve been there.

Oh, you’ve been there? Okay. They’re really pushing entrepreneurship, trying to generate more ideas and teach people how to start companies. There are co-working spaces and all kinds of incentives, including visas to come to Riyadh and to the UAE—Dubai and Abu Dhabi specifically.

I’m curious about your impression of the individuals who are leading a lot of these different groups. I was really taken aback by how much time they had spent in the West. It seemed like every single person I met had taken one of these scholarships to go to Oxford, Harvard, Caltech, or wherever it was. They had all spent massive amounts of time in the West and then come back to Saudi, the UAE, Oman, and so forth. Maybe you could talk a little bit about the people, their knowledge base, their motivation, and so forth.

I had the same observation. The elites of all these countries have all been educated in the West, usually in the United States, with some in the UK. The leadership tends to be young, visionary, future-oriented, and intensely interested in AI, like I was saying. They want to do big things in AI.

The thing that’s been in their way is that, in October 2023, the Biden administration basically put a blanket ban on the whole region acquiring semiconductors and GPUs. It required that every export of a GPU, or a server that contains a GPU, had to get a specific license from the Commerce Department. It put a major damper on their efforts to do things like build data centers in the region or have their own local AI efforts.

They’ve been in this holding pattern where the Biden administration decided to alienate them. It wasn’t just AI; there were other things, too. Remember the fist bump? The Biden administration was very—I’d say hostile, standoffish, and cool to, and even hostile toward—these states in the region. From a geopolitical standpoint, I think it was just stupid because we’re in an intense competition—a high-tech competition, a security competition, and an economic competition—with China.

And these states want to be aligned with the U.S., and we are pushing them into China's arms. If we don't give them the ability to buy the American tech stack, it's not like they're going to sit on their hands and do nothing. They're going to be forced to buy the Chinese tech stack. They're going to participate one way or the other, right?

China still has some limitations on how many chips it can produce, but they clearly are building their own Chinese tech stack. It's a Huawei-plus-DeepSeek tech stack. There's actually a story in Malaysia from just the last couple of days where they were talking about building a local data center using the Huawei Ascend GPU plus DeepSeek. They walked back that story because I think they're afraid of getting in trouble with the U.S. government.

But the point is that if we drive these countries away, if we alienate them, they will end up partnering with China instead of the United States. I don't understand how that's in the United States' interest. As part of this trip, we rolled out a new framework for an AI acceleration partnership with these countries, and there are a few very important terms that I think it's worth going into.

First of all, the purpose of the framework is to replace the Biden export controls in the region with a framework that gives them the ability to buy GPUs and build data centers. There are a few very important provisions. Number one is a matching-investment provision: For every dollar of investment they make in building out data centers in the region, they have to invest a dollar in the United States in building out our AI infrastructure. So it's a huge win for the United States in terms of accelerating our own infrastructure buildout. That's point number one.

Number two is that, even with respect to the data centers they're going to build in the region, at least 80% of the chips have to be owned and operated by American cloud service providers or American hyperscalers. So even with respect to the data centers they're building, the vast majority of the compute is going to be run by American companies. To me, this is just a huge win.

Was there criticism? Were people criticizing this expansion? I guess Ro Khanna has said, “Hey, we should build more data centers here.” That was sort of a light critique, maybe, or a misunderstanding that Ro was making.

Ro is a friend, but I responded to him on X: We are building out data centers in the U.S. We want to do as much as we can. It's the Trump administration's policy toward energy that's making that possible. We're making it easier to spin up new power generation, and we're making it easier to permit. Both of those things were virtually impossible during the Biden administration. So we are building more here in the U.S.

In addition to that, you've got these countries that are resource-rich, and they have their own AI dreams and aspirations. We can either partner with them or drive them into the arms of China. The deal we've made is to enable their aspirations while getting them to fund more AI infrastructure in the United States.

Even the infrastructure they're building over there is going to be on an American tech stack. We want it to be NVIDIA and AMD and those types of companies. That improves our balance of trade. It also ensures that these data centers across the world are going to be built on American technology. We want American technology to become the standard, of course. If you allow those data centers to be built on Huawei plus DeepSeek, that will become the standard.

I think this is something that people in Silicon Valley intuitively grasp but people in Washington don't. The way that you win these technology battles is you create the largest ecosystem. You want to have the most partners, the biggest app store, the API everyone uses, the most data and the most power.

Exactly. So the point is, we want to involve the whole world on our tech stack. I think export controls on China make sense, but we don't want to reduce our market share in the rest of the world. We want them getting hooked on our platforms, making those platforms the standard and getting locked in before China can catch up to us.

Sacks, did Jensen just say that he thought export controls were kind of pointless? Did he just say that recently?

Yeah, he's against all export controls. Where I would split the baby here is that I think most people in Washington believe that export controls on China make sense because we simply can't allow our most powerful semiconductors to be used by China because of the dual use.

Where I agree with Jensen is that I think we should not place undue restrictions on the rest of the world using the American tech stack, because we will simply concede the market share to China. So I think he's right with respect to the Middle East.

The other concern that you heard a lot is around this so-called diversion—the idea that somehow the GPUs or the IP, if we allow it to be used in data centers in the Middle East, will somehow find its way to China. The expression on your face, JCal, shows that this makes no sense. These guys have unlimited resources and unlimited ambition. That's just a naive point of view. They want these for themselves. They understand the power of AI. They don't want to give it to China; they want it for themselves, period.

They make trillions of dollars selling oil and gas. They don't need to smuggle our GPUs. It doesn't make any sense. The other thing people don't understand is that they're called chips, but really these things are now the size of mainframes. The NVIDIA data center product, NVL72, is this giant server cabinet. It's 8 feet tall and weighs 3,600 pounds.

You're not putting them in a suitcase and getting them across the border. That's definitely not happening.

No, it's not happening. All you have to do is send an inspector to the data center to count the servers, and you can see that they're there. Ridiculous. So this idea that somehow they're going to get smuggled is a totally fake concern that's been created. It's almost like a totally fake narrative.

Then people say, “Well, they won't physically smuggle the servers, but the IP will somehow be transferred to China.” Look, there's nothing in a data center that China hasn't already seen. They have at least one of everything. The issue is that they can't reverse-engineer the advanced semiconductors just from the chips, because it's about the process technology it takes to create an advanced semiconductor. There are hundreds or even thousands of steps in that process, and you can't figure out the recipe just by seeing the end product.

I would like to say that I'm very proud of my bestie. I want to tell you, Sacks, you crushed it. I think you put your best foot forward, and people were really impressed. I got 3 calls from the powers that be. They were very psyched to meet you and spend time with you, and they thought you were super impressive. So it was awesome. Thank you for doing that.

Yeah, it was pretty impressive.

Well, I appreciate that. Yeah, it's an incredible region. I think you're doing a really patriotic thing. If you look at that region, like you said, it could tip one way or the other. They could be involved with the West. We could be building businesses together, and if we build businesses together, that's better than them building businesses with China. Pretty obvious, I think. I mean, totally.

And this is the thing: I just don't get the opposition in Washington to this. I especially don't get how the opponents of this can call themselves China hawks, because at the end of the day, this is a tech battle between the U.S. and China. We can have these countries aligned with us and in our orbit, or we could push them into the arms of the Chinese.

The choice is: Do we want these countries to be the piggy bank for American AI or for Chinese AI? Let's say they have an investment level that dwarfs anybody else in the world, with the exception of Norway and the United States. They can put a lot of money to work, and here's the thing: They want to.

It just hit the wire that OpenAI is partnering on a 5-gigawatt data center cluster in Abu Dhabi with G42.

Let's hold on. Let's just be clear about this. The 5-gigawatt cluster is more of a campus that's going to support many American hyperscalers and cloud service providers, and OpenAI is going to be one of the tenants. I think they're saying that they're going to have 1 gigawatt. There's still 4 gigawatts to go that they plan on supporting.

The most interesting part of the deal, it said, was that as part of the deal, G42 plans to make a reciprocal dollar-for-dollar investment in AI infrastructure in the U.S. So to your point, Sacks, it's perfect. Cooperating like this, I think, just makes a ton of sense.

Absolutely. And this is not dumb money, by the way. These are very smart people who are investing very strategically with a very crisp group of individuals, as you said, who have been educated in the West. When people explain to me, “Oh, yeah, just go to Abu Dhabi and come back with a suitcase full of money,” no. These are folks who want to be on the boards of companies. They want to build companies. They see themselves as owners in business, not just investors.

They want to do it domestically and abroad. The amount of investment that they're going to do in the next 30 years is going to dwarf what we see out of universities and retirement accounts. I don't think people understand the scale of what's happening. This is why I just don't understand the point of the Biden policy of alienating these countries and making them feel uneasy about their place, because you don't know who it was that made the decision.

Yeah, let's face it, that guy was confused. We can have hearings. I've actually got a pretty good idea at this point of who was controlling the autopen on AI. There was a group of 6 very powerful staffers in the Biden administration, about 6 people who controlled AI policy.

The interesting thing is that none of them had ever worked in Silicon Valley. They're all basically lawyers, and they don't really have a good understanding of the idea of a platform, a technology ecosystem, or a stack. I don't think they realized that by alienating these very resource-rich countries, they were creating a huge opportunity for China to come in and serve the region.

This is why I think this—what I call AI diplomacy—is so important. It boxes out China. It's good for America, it's good for our companies, it's good for our trade balance, it makes us the standard, and it boxes out China. What's not to like about this? I really don't get it.

I think they were living in a world where they put people into very rigid buckets. If you wanted to look at human rights as an issue, which I think is what some people were doing in that administration, I look at the progress that's being made. If you go to these societies and look at what's happening in that region, the pace at which human rights and individual freedoms are growing, and the rate at which those societies are moving forward, it's going in the right direction.

So what do we want to do? Do we want to steer people toward collaboration and working together to build the future, or do we want to have it devolve? It's such a stupid decision by the Biden administration. It was perplexing. I really don't get it. I don't understand what the point of alienating these Gulf states was.

Well, I like Trump's approach there, when he said, "Listen, you're going to make decisions in your own time." That's what they want to hear. They want to be respected, as they respect us. "Hey, we're going to make decisions in our own time, what's right for our people. That's a process that we'll decide."

Now, if you want to collaborate on building a company, a data center, or a technology platform, hey, we're here for that. But you can't tell us how to run our society, just like we wouldn't want them to come here and tell us how to run our society, right?

So I agree with what you just said, JCal, and I thought the best part of the president's speech—the whole thing was excellent, but the best part was when he talked about how the neocons and the interventionists had come into the region, lectured people, and come in with this very moralistic point of view. They made all these interventions that basically wrecked countries.

It wasn't the neocons and interventionists who led to so much progress in the Middle East and in the Arab world. It was the countries themselves, whose own leaders set the path and had the vision. That was an incredible speech. It was incredible. Trump, for me, did a great job. Credit due.

That's the message they've wanted to hear for a long time, and that's the message that moves society forward. All these other messages just set it back. I think I'm super hopeful about the region.

Okay, let's do Science Corner. Oh, by the way, I was just checking X here, and it looks like there's an announcement about what the OpenAI device is going to be. You guys see this? Here we go.

Oh, the puck. The little puck. I saw that puck, but it hasn't been confirmed yet. It's a puck you put on your desk, and it's like a puck with a camera and a hole for a mic. Oh, that's a pin. That's a pendant. It's a pendant. I could have made 10 times straight up versus a pendant.

That's a puck. It fits in the palm of your hand. It's too small to be a puck. A puck's about the size of your hand. It looks like your thumb.

Okay, this is AI-generated. It's not actually the device. They haven't shown the device from Jony Ive yet. Somebody just took what's in the Wall Street Journal report and mocked it up. This is their best guess as to what it might look like. This is already out.

By the way, this is going to get you punched in the face. You show up with a listening device to a party or a meeting, I'm going to knock you out. This is ridiculous, and it's got a camera in it, too. I don't want this surveillance device anywhere near me. How are you going to do it? Like this? Yeah. Get that puck away from me. Anytime you're ready: How dare you record me and violate my privacy?

I was literally at dinner the other day. I kid you not, somebody came up with one of these things, and it was clipped on his lapel. I said, "Hey, what's the thing on your lapel?" He said, "It's an AI recording device. It summarizes what we're doing." He said, "Oh, do you care?" I said, "Do you care if I punch you in the face for covertly recording me?" He said, "Yes." I said, "Well, then turn it off."

Let's talk about science. Let's talk about science.

You raise a good point, JCal, that these new devices require a level of trust that is kind of unprecedented. You already had to trust Google and Apple and so on, because they do have a huge level of access to your data. But the new—let's call them AI-connected devices—are surveillance devices.

Their whole point is to listen and see everything that you can hear and see, and then give that data to AI to do something with. You better trust that company completely.

Oh, yeah. Which company is doing it? Oh, right, that company. The company that had every leader leave because they didn't trust the founder.

Okay, all right. Go ahead and wear the puck. By the way, with these pucks, I'm going to make one more point, Friedberg. These devices need to have a red flashing diode on them when they're recording. We need to pass legislation requiring that. If it's recording, it should have a flashing red light on it.

Let's go to Science Corner, because everybody—Cham loves Science Corner. Friedberg, some breaking news, some innovations in CRISPR.

Yes. It's gotten a lot of press coverage, but there was a baby born named KJ. The research paper that was published says KJ was born with a mutation inherited from the mother on a specific gene called CPS1, and the same mutation inherited from the father, which resulted in a deficiency in this gene.

CPS1 produces a protein that's part of the urea cycle, basically breaking down nitrogen or protein compounds in the blood so that they can be excreted. In the absence of this protein—if this protein isn't being made in the body—ammonia accumulates in the blood, and that has extraordinarily harmful consequences, including brain damage and ultimately death.

The gene had a G that was mutated into an A on both copies of the gene, which meant that the protein wasn't being correctly produced. As a result, the breakdown of nitrogen and ammonia wasn't possible. This leads to a very short life for the very few people who have ever had this double mutation that this particular child was unfortunately born with.

We have CRISPR gene-editing technology today. In fact, CRISPR gene-editing technology is what we use to do some of the work we do at the company I run. Some advanced forms of gene-editing technology include what are called base editors, where you can change a single letter on a strand of DNA into another letter.

In this particular case, changing an A to a G was the goal. The physician Rebecca Ahrens-Nicklas deserves a big shout-out for the incredible work she did in partnership with scientists at the University of Pennsylvania. They developed very quickly a very specific gene-editing target to go in and find the letter A and turn it into a G in this particular patient's body.

Now, remember, we have the same DNA in all of our cells, but specific cells have specific functions, and genes are expressed in only those cells. In this case, it's the liver that makes this CPS1 protein. They had to get the gene editing into the liver cells to edit them. If they could edit the liver cells, then those liver cells would start to become more functional and produce a functional protein.

They quickly tested dozens of base editors. These are different proteins that can go in and edit. They tested different guide RNAs. These are the guides that tell the editing protein where to go, what part of the DNA to go to, trying to get it to this exact same site.

They tested it in Petri dishes, then they tested it in mice, then in monkeys to make sure that the selection they came up with worked. They then put it in the child. Basically, by putting it in the blood, it found its way into the liver.

The fortunate benefit here was that, at the end of the day, you're trying to edit specific cells, and we can get to the liver by putting the CRISPR and the guide into the blood. It made its way to the liver, and they coated the gene-editing system with lipid nanoparticles, which helped it get into the cells.

It was absorbed by the liver. It went into the cells. It found its way to the site on the DNA.

That's amazing.

It converted the A to G, and as a result, those liver cells became functional and made the protein that breaks down the ammonia in the child's blood.

They did the first dose just to make sure the child would handle it. It worked. The child didn't have any adverse effects. They then did the 2nd dose and were able to take the child off some of the medicines and compounds they were giving the child to break down ammonia in the blood. They gave the 3rd dose, and they're monitoring the child.

This is the 1st time that we've seen this kind of custom CRISPR gene editing for a genetic mutation applied to a patient in vivo, in the blood, for a specific target and a specific treatment.

I just asked Grok how many diseases are caused by an A-to-G point mutation, and it says it's hard to estimate through the Human Gene Mutation Database or ClinVar, but it estimates that between 20,000 and 30,000 conditions are caused by an A-to-G point mutation. Isn't that incredible?

Yeah. I'll tell you one other story about this. I saw a geneticist here at UCSF a couple of years ago, and he was telling me about all of these kids who come in and they don't know what's wrong with them. He can see it on the kid's face. He can see a disease condition. They have all these weird lab tests, and the kids are dying or they're really deficient in some way, and they can't afford to do the DNA sequencing because the child is on Medi-Cal or some other sort of health care plan. It costs about $5,000 to do the DNA sequencing needed to figure out what genetic mutation that particular child has.

There aren't easy, off-the-shelf targeted tests or lab tests you can run to figure out when people have a genetic disease. This is an incredible, undiscovered set of problems. Chamath, to your point, there are likely millions of people out there who have some health condition that they've not yet been able to identify because it's a genetic mutation, and not everyone can afford to go out and just get their genome sequenced to figure out where the mutation might lie in their body.

As we discover more and more genetic mutations, we build targeted tests to identify that people have them, and hopefully, over time, we can start to build more targeted therapies like this one to go after and make the changes at specific sites to fix the genetic mutations.

Badass. It's incredible. Okay, somebody wake Sacks up. Sacks, just so you know, one thing about America is that we've got a pretty good advantage in gene-editing technology today. China is on its way to coming after the U.S. on this technology front, as well as AI. You woke him up. You should pique his interest.

China's coming for your anus.

You'll see some very good posts recently from people talking about how a lot of technology that's been developed in the United States is being transferred to China. They're building biotech companies and biotech solutions that take American IP and put it into the market, and this has also had a very big dampening effect on the biotech market. So what do you think we should do about that? Should it be in the trade negotiations? Should we have tighter IP controls?

I can tell you, Sacks, that I had a conversation with the CEO of a very large company, and they said this exact thing: The amount of IP that's available for purchase or licensing, of questionable provenance but of uncertain effectiveness and quality, is causing them to figure out how to establish more of a presence there to just buy stuff.

What that means is that, let's say you have an R&D effort in the United States. You can almost hedge your bets by having an outpost in or near China—let's say Hong Kong—where there's an equivalent project for almost anything that we're working on happening there at the same time, in real time. If you can't figure out how to get it here, you can get it there. It's much murkier, the laws are sort of unclear, and it's just different enough that you can get to the same place.

So it's just to say what Friedberg said: The amount of R&D that the Chinese are doing is roughly equivalent to the amount of R&D that the United States is doing.

Yeah. As a result, everything that we discover, they discover. It really is a footrace, and we're kind of neck and neck.

For sure. This is a perfect time to make sure that we have an electricity deficit. Perfect time.

Well, I agree with you. We need to build out power, but I think the administration is committed to doing that.

Drill, baby, drill.

I'm not questioning Trump's desire. I'm talking about the actual congressional bill details. I saw yesterday that the Tennessee Valley Authority filed the first-ever application for a small modular reactor. Let's go. Hopefully, it gets approved very quickly.

This is, by the way, something I will watch. I think others should watch this to see how quickly they're able to get through the regulatory process to get that thing approved. If it gets approved and they can get fast permitting and fast deployment, that would be an incredible insight into how quickly the U.S. will be able to scale power production for AI.

Over-under 2032.

What exactly?

Let's do a Polymarket. Great. Let's do a Polymarket, and you should create a special Trump designation, like a speed-run designation. When something's super important, he puts the speed-run designation on it, and they actually attempt to do it as fast as possible.

I don't know if you guys watched my interview with Burgum, but he said that there's a permit for a copper mine that's been outstanding for 30 years. They've been trying to get through the permitting process—30 years to get this copper mine started and running. Now he and the team are trying to fast-track this thing, but it's been sitting there for 30 years.

Here's the good news: The Manila folder has made its way down the elevator. It's down there in the subbasement.

Let me ask Friedberg a question. I don't want to get rid of DOGE—I'm joking—but given the amazing advances that you see coming with CRISPR, biotech, AI, and robotics, what are the odds that we can grow our way out of this fiscal problem?

Yeah. I do. I'll be honest: I think the limiting factor—so, the CRISPR side is going to have profound effects on longevity, human health, and food abundance.

I think about abundance in 4 ways. One is an abundance of food, or calories. One is an abundance of labor through automation. One is an abundance of lifespan, or longevity. The fourth—and I think this is the most important and gating factor—is energy, which leads to everything else.

You can't have the labor savings or the productivity improvements without the energy. The thing I watch the most, Sacks, and that you and I have talked about, is the critical linchpin for all of those other points of abundance being unleashed. It's the energy equation.

This is where I look very clearly and plainly at what China is doing versus where we're at today. We've got stated intentions, but the actions are where we're still—we've got to start showing up.

That's it. But let's assume that we have enough power. My point is really that if we have enough power, the $38 trillion of debt doesn't matter.

Really? You're saying that in a very serious way. That's a pretty important conclusion, because if that's true and we can grow our way out of this problem, then that's an alternative to having this austerity approach, for which there's no political will.

No, I'm very serious.

Sacks, listen: If I saw us adding 1 terawatt of electricity-production capacity per year in the U.S., I would shut the fuck up about the debt. I wouldn't care.

A terawatt? We're talking about gigawatts.

Yeah, I'm talking about 1 terawatt.

No, I know. But I'm saying, China—you guys know China right now is dealing in gigawatts.

Yeah, I know. Have a terawatt, and China is dealing in terawatts. China is moving. We've been through the numbers, but they're adding an entire United States every 18 months.

Every 18 months. Yeah.

Every 18 months, China adds all of the power-production capacity of the entire United States to its grid, mostly in nuclear and solar. Solar is a big component. Hydro is—look at the number of times they're adding hundreds of gigawatts just on the Yangtze.

Look at the number of times Elon has said this. The amount of energy that's available if we just harness it from the sun and redirect it with some storage would solve all of our problems. But the problem is that we're at just a little over 1 terawatt. We're inching along in gigawatts, whereas China is in terawatts, adding terawatts.

Yeah, they're in that chart. They're at 3 terawatts in a corner of Arizona or Utah, going to 8, and we're at 1 going to 2.

I've shown this many times before on this show. This is total production over a year, so this is terawatt-hours. It shows 4,000 terawatt-hours, which is where the U.S. is at. If you break that down into basically continuous production, you measure it in watts. The number of watts that the U.S. can make at any given point in time, or is making at any given point in time, is 1 terawatt—1 trillion watts.

Meanwhile, right now, China is making 3 terawatts, and it's scaling up to make 8. We are at 1, going to 2 over the next 15 years.

So you consider this graph to be more important than the budget-deficit graph?

I do, and I'll tell you why. I've got another chart I can pull up for next time.

Now, do you think it's esoteric, Sacks, to add these line items to this market?

Well, I don't. I just don't. You're going to have to provide a lot more information about what exactly—

No, but look, it's obvious. It's obvious. And I just want to translate this: It's so obvious that, in any reporting on that, Jason, I just haven't seen anything. It just happened last night. This is the point.

I'll just say one thing because I want to add it at the end: Automation, which is unlocked by AI. We are going to see an explosion of robots. They're not all going to be humanoid robots, but robotics allows humans to get an incredible amount of work done.

Imagine having the cost to build a giant building go down by 50x. That's what this unlocks. This unleashes all of this.

Optimus doing it. Yes, we'll have Optimus, but we'll have these robotic devices doing it. You can see it in China. Go look at how they build bridges. It takes like 3 days to build a multi-mile bridge in China because the whole thing is automated.

They do automated drilling. They do automated mining. They do automated building. All of that, if we unleash that capacity in the United States, can help us build things that are economically productive for America. All of this gets unleashed. The technology is here today. The only thing that's missing is the power.

David, I would just then ask folks: Go talk to Duke Energy, go talk to Constellation Energy, go talk to the drillers, but they'll all tell you the same thing. If you want a natural-gas turbine, we have to wait until 2030. Why? Because China owns them all. They make them all.

So if you say, “Hey, Chamath, drill, baby, drill, and use that gas,” and I say, “Okay, I'm waiting in line for 5 years to get a turbine,” then it's like, “Okay, what can I do?” Nuclear? I can't, because that'll take till 2035. So then I'm like, “Okay, I'll go into the tax-equity markets and I'll finance myself with residential or industrial solar.” And when that goes away, there is no energy.

Somebody should just go and figure this out. By the way, this is our Manhattan Project. There was the Manhattan Project, then there was the Apollo Project. This generation's Manhattan and Apollo Project is energy-production scaling in the United States. That's it.

We don't need the government to do AI. We don't need the government to do automation. Private market. Let us do the energy. We need to figure out how to get energy, and you can't decapitate the markets that finance it. We're all willing to put up the risk capital. That's the crazy part. We don't even need the government.

I don't know any of the details of what Chamath's talking about, but, generally, you think you're a chairman dictator.

Yeah. Chairman dictator.

Who do you dictate for, your chairman dictator?