The Bulwark Podcast - Catherine Rampell and Stephen Witt: Can We Avert Economic Disaster and the Robot Apocalypse?

Episode Date: September 16, 2026

The Fed raised interest rates to curb inflation, as expected, but mortgage rates are surging and flailing Scott Bessent remains the treasury secretary. Meanwhile, the alarm raised in the AI community is not an op despite the backlash from people like Nvidia's Jensen Huang, who seems to get everything he wants from Donald Trump. Plus, the Europeans who once parked their gold in the US to hide it from the Nazis are now trying to protect it from Trump. Catherine Rampell and Stephen Witt join Tim Miller.show notes: Subscribe to Catherine Rampell's Receipts podcast and newsletter on Substack or follow her new show wherever you get your podcasts: Apple Podcasts | Spotify | YouTube | More  Stephen's "This Is Really Bad" piece in the NYT The latest episode of "The Next Level" Dan Selsam's statement on Twitter that Tim read from "If Anyone Builds It, Everyone Dies," the book Stephen mentioned Huang's "impromptu" phone call with Trump Closing song - Orville Peck - It's the End of the World NOBL gives you real travel peace of mind — security, design, and convenience all in one. Head to NOBLTravel.com for 46% off your entire order! #NOBL #ad Bulwark LIVE: Benefit for SpringfieldSome moments require you to step up. Some require you to show up. That’s why The Bulwark is coming to Ohio.  Join Tim, Sarah, JVL, Sam and friends Friday, September 25 at the Dayton Masonic Center for an evening of politics and fellowship.

Hello and welcome to The Bullwark Daily. I'm your host, Tim Miller. We have a massive show today. It's a double header. We're doing economics. We're doing AI. So I don't even have time to get to my guy Thomas Massey bringing the articles of impeachment against Peter Brian Hegseh. So go check out the next level. It's already up. We had Eager in for Sarah Longwell. We get into impeaching Hegset, other impeachment ideas that we have and other political news of the day. On this show in segment two, I bring in the author of a New York Times column, type. This is really bad, which is just squarely targeting us here at the Bullwark. You know, I mean, that's that, that is clickbait for me if I've ever seen that. This is really bad. It's about the AI vibe shift. And I think it's a worthwhile conversation, particularly if you're trying to make sense of the changing narrative and discussions around the AI threats. But first, she's the economics editor here at the Bullwark, author of The Receipts Newsletter. And coming up on Friday, a new podcast also called,

receipts. It's Catherine Rappell. What's up, Catherine? Hey, thanks for having me back. All right. Welcome. The new pod is going to be in its own feed. So people don't have to go get that. Go sign up on Friday. Make a little note for yourself right now. And we're really excited about it. What's going to be happening on the pod? Do you have any little teaser for people? For this week, well, we shall see what happens with the Fed today. We're recording this. I should let our audience know before the Fed announces what it's doing with rates. It's probably going to raise rates today on Wednesday. And my expectation is that the president will flip out when that happens, but I don't know. But assuming that that happens, I think that'll probably eat up most of the

conversation like for the next week. So we'll be talking about plenty of that stuff and other crypto scams and various other things happen in the world. A lot of great economic news. So we'll do a little teaser of it right now. Let's talk about the potential state of play. I mean, you said kind of what we think is the baseline expectation for the Fed. But just kind of backing up for people, why does the Fed feel in this moment that they need to raise rates? I mean, we're going to get into sex, but bond rates, mortgage rates already going up, you know, for reasons unrelated to what the Fed is doing. Well, sort of related.

Yeah. Well, yeah. What is it about what's happening in the economy that is going to lead the Fed to raising rates? Well, we've had high inflation for over six years at this point. We have a new Fed share who most of his life has been an inflation hawk, meaning that he has a bias towards raising rates higher and to deal with inflation. That's usually the prescription for dealing with inflation is higher short-term interest rates. So if you take him at his word that he deeply believes that the solution for inflation for getting stable prices, which is the Fed's responsibility, is, tighter monetary policy, you know, which he criticized his predecessors for not having. If you think that he deeply believes that and he wasn't just saying it for political reasons while there was a Democrat in the White House, that is what you should expect the Fed to do today. And he basically said as much in a recent speech where he kind of boxed himself into a corner

where he told markets more or less like unless we see something, some kind of major improvement on the inflation front in the next few weeks, it's up to the Fed to deal with it. How do you think you're balancing? I mean, at some level, like, inflation has been high, but it came down. I mean, it depends on kind of which sector we're talking about. Obviously, we're saying a ton of energy inflation, but that's, there's some external reasons for that.

But job growth is also, it's not great. I mean, it beat expectations a little bit in the last report. But how do you balance that, right? Like, it's not exactly like we have a booming job market right now. Yeah, well, that's basically why the Fed has held off on raising rates for a few years, despite the fact that inflation is high. Like I said, normally the prescription for dealing with high price growth is you raise interest rates, which makes it so that money is a little bit more expensive.

People don't spend as much money. They don't go out and buy as many cars and houses and whatever stuff they might put on their credit card, et cetera. That's the mechanism. Sure. However, because we have had... kind of a stagnant labor market. And because we have had a bunch of these other interventions from this president

that have slowed the economy more broadly, I'm thinking about tariffs, for example, that has made the Fed say, look, I know we're worried about high inflation, but the remedy for dealing with high inflation might kind of have these unintended consequences of killing the recovery, killing the economy. And if we're worried about the health of the economy, maybe we don't want to deal as aggressively with inflation. It's this term stagflation, which people from the people who are around during the 70s may remember. It's the idea that you have high price growth and you have kind of a weak economy. And the thing that you would do to deal with one of those problems will make the other problem worse.

So they've kind of been in like a holding position where they're just not raising, they're not neither cutting nor raising rates. And so that's the situation they've been in. But now the economy, the job market, like, seems like it's holding its own. It doesn't seem like we're in a recession. It doesn't seem like we're losing jobs. We're not gaining a ton of jobs. We're also not losing a ton of jobs. Maybe that means, again, if you're a hawk, if you are someone who is biased toward

being more worried about inflation, you'd be like, okay, I'm less concerned about these, like, knock on effects, these unintended consequences. I'm more concerned about, like, the keeping our eyes on the, prize and dealing with inflation, raising interest rates. And again, that is how Kevin Warsh has generally been understood to think about these problems, which is part of the reason why it was so weird when Donald Trump chose him for this job, because Donald Trump, the only thing he cares about is cutting rates. And I think part of the reason he held off on making a decision about who to appoint as FedShare is that, like, markets maybe seem to trust Kevin Warsh, but he was viewed

as a hawk, which is the opposite of what Donald Trump wants. And so the real question is, like, how does Kevin Warsh navigate this really difficult position that he's in where clearly the guy who thinks he's Warsh's benefactor is expecting him to cut rates and markets and everyone else is telling him to raise rates and his own like general outlook on the world would suggest he wants to raise rates? And so that's the thing that I'm wondering about is like how is he going to deal with the inevitable ire and potential lawfare from this president. that his predecessor dealt with. All right. Well, if Donald Trump has a diaper tantrum today, you'll be around on Friday on the new show to discuss it. Yeah, yeah.

The other Warsh thing is just a totally a side point because I saw it this morning is there's a brewing scandal in Greenland. You know, it's Warsh's father-in-law, water that was like one of the big proponents, friend of Trump, of the gambit that we need to go seize Greenland. It seems like there's a corruption scandal up there where we're paying off, you know, maybe some of the pro-America factions

or a pro-America person in Greenland. And Lauder was one of the people that was feeding money to a pro-Trump, whatever, advocate in Greenland. So potentially a lot of drama around the family table over the next week. Wow. I had missed that. It's like it's all one mono scam in some ways, you know? It's like threatening to invade our allies also intersects with debasing our currency, also intersects with destruction of rule of law,

Like all of these things are related in creative ways that a very clever author couldn't have come up with. We couldn't have drawn up. And the bond market is acting, I guess she said, it's related, but is also acting in some ways independent of what's happening with the Fed. We have here from this week, the 10-year note rose above 5% for the first time since 2023. U.S. mortgage rates are now up over 100 basis points since their February 23rd low, which was the week before the Iran War. again, maybe a coincidence, who knows? And the average interest rate now on a new 30-year mortgage hit 7.2%. Hopefully none of our listeners are trying to buy a new house or sell a house.

Or refinance. Or refinance a house. I mean, that's brutal. Yeah. Yeah, it's totally brutal. And everything that this administration has done to try to, I don't know, keep the inevitable from happening has just made them look even more feckless. So you have these long bond interest rates.

rising for a few different reasons, some of which are related to Donald Trump and some of which aren't. So among the issues are inflation has been elevated and interest rates tend to go up when price growth is higher. So that's why I said it's like not totally unrelated to the Fed. We also have a huge amount of debt in this country. And that's again, something that precedes Donald Trump, but he has definitely made worse. When a country has more debt, the people or companies who are lending to that country, which is what these government bonds are, tend to demand higher interest rates because they're like, we're not really sure you're going to pay it back. At least that's how it works at other countries. We've kind of been insulated from that sort of feedback or that sort of

investor demand because we have the world's global reserve currency. And so, like, we've played by different rules than everyone else, but maybe we're kind of losing the privilege to behave that way that recklessly. So there's, there's that that's happening. There's also a bunch of like AI companies that are issuing their own debt, yeah, because they're trying to build new data centers or whatever. And so like when they're issuing bonds that, you, they're competing with American government bonds. So like that that means that like some of the money gets siphoned away. And so the American government has to offer higher interest rates to entice people to lend us money to buy our bonds. So like there's a bunch of different things.

that are happening. And again, some of them are related to Trump. Some of them are not related to Trump. Meanwhile, the things that they are doing to try to stop this from happening have worsened everything. So that's like Scott Besant, the Treasury Secretary, saying, we are going to buy up a bunch of these bonds ourselves. And so if they buy more bonds, that should drive these interest rates lower. Because if you have more people who are buying the product, again, like the interest rate doesn't have to be as high to entice buyers, right? So this is what he said he was going to do. He said he was going to do this like a month ago. And basically the rest of the market was like, yeah, you don't have a big enough bazook. You don't have the ability to buy enough bonds to move markets

this way. And so they called his bluff. And he has been sort of flailing ever since demanding get people of respect him. Yeah, yeah, yeah. Yeah. Megalomania. Megalomaniac Scott Bessent presiding over an economy where people have to pay seven percent for three. 30-year mortgage right now. This was him on a panel puffing out his chest a little bit. Let's listen to that. Whenever people say, oh, well, Treasury Secretary is taking a risk. It's my dream.

I have asymmetric information. I am the house now. So when we intervene with the Japanese yen, I have pretty good insight into what the Japanese, what the banker of Japan's going to do, what Japanese policymakers are going to do. And you can bet against me if you want. I might. I don't know. Donald Trump has bankrupt a casino before.

So it wouldn't be the first time that the House has lost. Yeah, exactly. So I would point out a few different things that are problematic about this very hubristic comment that he made. He's like, I'm the House. I can do whatever I want and you can bet against me if you want. you're going to lose. First of all, it doesn't seem like that's been bearing out so far. Like I said, markets kind of called his bluff. He was trying to get interest rates down.

Instead they went up. Maybe in the long run, he'll be right. And, you know, all of those things that I laid out that were driving up interest rates are going to go away. And, like, he won't be pushing on a string here. But I think that's pretty unlikely. But I don't know, you know, like they're professionals who may feel differently. But the idea that, like, he has so much more insight in all of this is just, it's both hubristic, but it's also kind of ironic because Scott Bessent made his name as an investor as a trader working for Stan Drucken Miller, this legendary investor. And one of his most famous ways that he made money, one of his like big claims to fame was that he bet against another hubristic government for trying to

basically, in this case, it was the UK. They were the Bank of England. They were trying to hold up the value of the pound. And like, he was one of the market participants who called the bluff and said, you don't have enough bazooka power to do this. It's very similar to what's happening now. He's on the other side of this. He's like basically mocking markets for making a bet against him in the same way that he made a bet against the Bank of England.

So, like, he should know better than most that when a government is trying to, like, put a band-aid on a bullet wound, you know, like they're, they're like trying these little feckless things around the edges to stop the inevitable from happening. They're likely to lose. So he should know that he is not in this all-powerful position just because he says, this is how things are going to go. Does not mean it's how it's going to go. Like the fundamentals are not on his side, just as they were not on the side of the Bank of England back in whatever it was 1992. It's also 2026. He doesn't have that much more information than people actually. I mean, he might know what Donald Trump's.

about to bleat. And it's like that might help him with some short-term bets, right? Well, he probably does have some access to information about like the plumbing of the financial system that normies out there don't have. So I'm not going to suggest that like he does not have, you know, he doesn't have more information than the rest of us. But the main things that matter in all of this are, again, like things that he doesn't have the power to overcome.

Those are high inflation, government debt, and these AI companies issuing their own bonds. Like, those are a common set of facts. Like, whatever inside information he might have about lots of things going on in the financial system cannot overcome, like, these overwhelming forces that are pushing up interest rates. And so, again, he just kind of looks feckless in all of this in a way that is going to make a lot of people a lot of money in the same way that he made himself and Stan Drucken Miller a lot of money back in 1992. And it's like he cannot see the irony here. He cannot see how the tables have turned. And instead, he just kind of like, lashes out. And, you know, you play this clip of him saying, I am the house. He also complained about

the Bloomberg Terminal Bros. And it's like, you were one of those Bloomberg Terminal Bros, bro. That's exactly what you did. Apparently, just as an aside, part of the reason he might have been like calling out Bloomberg specifically by name, Semaphore had a good article about this, is that he has been pissed at Bloomberg for reporting, I don't know if it was a year ago. about how he had two primary residences. In the same way that, like, Trump is going after Lisa Cook and some other people for supposedly lying on their mortgage documents, which they've denied.

Bloomberg had a story, like a year ago, about how he was doing the same thing. And he threatened them, according to semaphore, and said, like, if you publish this story, I'm not going to go on Bloomberg. The Free Speech Administration acting. Exactly. Yeah. So, anyway, that's like a whole other inside baseball thing,

but like obviously Bloomberg historically has pretty good relationships with Treasury because their customers need data and, you know, hopefully access to administration insights and things like that. And he's been throwing a hissy fit for the past year. And I think that that is finding new ways to whatever, reach the world to like have visibility. He's, he's lashing out at Bloomberg. He's lashing out at people who were betting against him, maybe not realizing that he was, he was in that position 30 years ago.

In the same boat, yeah. Speaking of Bloomberg, this is not financial advice for anybody listening to this podcast. Please, if you saw what's happening with my Charles Schwab account, you would not be following me. But I will say that a lot of chatter out there

that if rates get much higher, historically, this has been a time that's led to big market correction. So I got to laugh at somebody that posted one of these finance posters who was like, you know, the last time the yield rates reached this high was in some month in 2007.

and I've got good news for Mr. Trump by November of 2008, interest rates were way down. Yeah, exactly. Well, you know, I was talking about, like, Kevin Warsh is in this very difficult position where Trump really wants him to cut interest rates and all the fundamentals and his own, like, world philosophy suggests the opposite. In some ways, the only thing that could really save Kevin Warsh and appease Donald Trump would be a massive recession. Right?

Like, those are the conditions in which we would actually cut it. interest rates, but obviously it would have a lot of other problems. You don't want to be rooting for a recession unless you're a moron like Donald Trump, and you just don't understand, like, what economic conditions would lead to this set of policies that are clearly your preferred policies. So I don't know. I generally don't know how they're going to find their way out of this. Guys, I'm kind of a mess at the airports. I think I've told the story before for new listeners. It's too funny not to share. It was one time gone on a plane, sat down. immediately spilled my entire drink on the next seat that my husband was supposed to sit on.

I have better status than him, so I was on first. He comes in. I'm like panicking, trying to clean it up. The nice woman comes to sit next to me. She offers like a dirty shirt or something that she had in her purse. Very kind. Use that to stop it all up. I had it back.

Thank you. Thank you. I'm very apologetic, grateful. and 20 minutes after that, she was like, you know, a big fan of the podcast. And I said, oh, thanks, great. What's kind of embarrassing then, you know, to be in this situation with you? And she said, well, you seem like the type of person to do a spill like this. And that cut deep.

But I appreciated it. And so when I'm at the airport, long story short, I'm looking for anything I can to help make things more efficient for me, make it easier to place my cup somewhere that I don't spill it, make it easier to charge my phone. And with the new all-in-one carry-on from Noble, you can do all that. They got a cup holder. It's a flip-out cup holder. Huge, need it.

It's charging port. ABC, always be charging. A laptop pocket, important for me, so I can pull it in and out without having to open up the whole carry-on and have people see my underwear. It's got the wheels that do the 360. Zip-less security. It's very secure. If you're worried about somebody coming to steal your underwear or whatever you have in your carry-on, they've got that dealt with as well.

It's a great carry-on. If you're looking for a new one, totally recommend it. They got a bunch of fun colors. Got black. I went with kind of like a mint for myself. I thought that was kind of chic. Every suitcase comes with a 100-day money-back guarantee and free three-day express shipping. They've got a lifetime warranty too, so it can be the last suitcase you'll ever need to buy.

upgraded luggage that looks better last longer travel smarter had to noble travel for up to 46% off your entire order that's nobl travel dot com for up to 46% off after you purchase they'll ask you where you heard about them please support our show and tell them we sent you I have a grab bag of other things that are of interest to

Catherine that I wanted to pick your brain about for people who have been following receipts live with JVL and your newsletter they're already familiar with what's happening with Europe buying their gold back from us But for those who, you know, our daily podcast listeners and who are going to be transitioning onto your new podcast who aren't familiar with this, the story is extremely delicious. And just give us like, give people like the TLDR of what's happening in the historical parallel.

So the short version is that almost a century ago, the Europeans evacuated their gold from Europe to America to hide it from the Nazis. And now they're taking it back to hide it from Donald Trump. essentially. Literal gold bars, bullion, gold reserves, are stored in the United States, in the vault below the New York Fed. You can tour it. At least I've toured it. They don't want you touch the gold, unfortunately. But it's all there. Unless you're Mrs. Manuchin. God, that's a character from season one. I can't remember her name. What was her name? That's going to come to me. You keep talking. I don't know, but it'll come to me too. Anyway, the actress, yes. So there is

this literal gold that is kept as reserves for these banks, and the Bank of France, the Bank of the Netherlands have taken gold back. Either they've sold it off and bought new gold to like store in the store underground in Paris or whatever, or they've actually physically transported the gold overseas. I don't know how. It's like, I assume it's like in a note. Louise Linton. Yeah, Louise Linton. Thank you. Thank you. Thank you. Anyway, they've done this. because they're worried about what Donald Trump might do to use this against them. Now, I don't think he's, like, actually likely

to tunnel underground a la die hard with a vengeance and, like, literally steal the gold in dump trucks. But who knows? But he don't know. Like, he did talk about how he wanted to show up at Fort Knox and knock on the door to make sure the gold is still there in case somebody else had stolen it. He literally said that, I think, multiple times.

So maybe he would do that. the more likely outcome is that he would basically say, we're going to treat the EU, our friends in the European Union, the same way that we're treating Russia, and we're going to say they no longer have access to this. You know, we've frozen Russia's foreign exchange reserves, for example. They might well do the same thing to our European allies. And in another era, this would have been insane to think that we would be sanctioning our friends this way, that we would be like basically expropriating them or at least freezing their assets.

the way that we treat enemy countries like Russia or Iran or Venezuela. But in the current environment, it's not so far-fetched. And what I have heard is that one of the things that got these other countries, and oh, and by the way, Germany is considering this as well, taking their gold back. Germany had gold here partly because of Cold War reasons. Like, it wasn't fleeing the Nazis. It was like making sure that the Soviets couldn't get their mitts on it. In any event, what they're worried about is like, look,

there's the whole Liberacee stuff within the White House, and I don't think it's, I don't think it's like literally that. Maybe we'll have a shortage. Maybe we'll have a shortage. I don't think it's actually that. It's like they're freaked out about the invasion of Greenland or the threatened invasion of Greenland. That was a much bigger deal in Europe than American media coverage led on. Like they were totally freaked out about that.

They're like, well, if he's threatening to invade a NATO ally, why wouldn't he just expropriate our assets? He talks all the time about how he wants to take the oil, or in the case of Venezuela, has taken the oil. Not just oil, of course. He basically strong- They said they took the gold in Venezuela. Yeah, that's right. That's right. I forgot about that.

And strong-armed Ukraine into this minerals deal. Yeah. And so it's not like it's unheard of for him to try to seize these natural resources or like, make some country give them to him under duress. So there's all of that that's going on in addition to his like weird obsession with actual gold decor. So all of these things have totally freaked out our allies. And so yeah, they're taking the gold back.

And again, in a different era, just a few years ago, this would have been nuts. And now it's happening. Sometimes I think about what would happen if I had this podcast like 2011. And I was like talking about Simpson Bowls every day. Like on today's podcast, we're talking about, you know, at the end of civilization and Donald Trump stealing the Dutch gold. Yeah. It's a strange time. It's wild.

It is totally wild. Speaking of other wild things, you broke the story a couple months ago now about their plans for kind of hijacking the census and changing rules around the census. And I think this is another underappreciated story could have impact on the electoral vote count. numbers, House of Representatives numbers in the next decade and potentially other ramifications. What's the latest on that? Yeah, not just the electoral vote, but lots and lots of money, too. So every 10 years, as our audience presumably knows, we have a decennial census. It is mandated by the Constitution.

The Constitution says that it is a total enumeration. That means every single person living here is counted, regardless of citizenship, regardless of other kinds of immigration status, whatever. And people may remember in season one of the Trump administration, Trump tried to add a citizenship question onto the census, and the Supreme Court stopped him, but mostly like for procedural reasons, not because they said that it can't ever be added.

This time around, they started a lot earlier, right? Like we're talking about the decennial census, which is not coming until 2030, and it's only 2026. They started with a lot more time to cross-thorough, tease and dot their eyes or whatever so that all those procedural things will no longer be quite as much of a hassle. And what they're doing this time around is not just saying we want to add questions to the census. They say they're kind of like disguised it in this obscure technocratic sounding change where they say, actually, if you are not a citizen of this country or if you

don't have a green card, you don't live here. You are not a resident of this country and so should not be counted in the same way that normally when a census is done, like they don't count the tourists in the census, right? They're basically saying if you don't, if you're not a citizen or you don't have a green card, you're essentially a tourist in this country and you live somewhere else. It's not just that you are a non-citizen, you are a non-resident. So that would mean not only not counting undocumented immigrants, which is a thing that you've been hearing from the right for a very long time, but not counting refugees, not counting DACA recipients, not counting people who were on H-1Bs, all sorts of different categories of people. And the idea behind this is that then you would

rob states like New York and California of millions of residents, and so you can steal congressional seats from them. You can reapportion those congressional seats to other states. Now, probably Texas and Florida which are now red states you know see what happens these midterms I guess in least in Texas they would be hurt too

I think David Jolly has a puncher's chance down in Florida too yeah we'll see we'll see so you know those states would be hurt too by this but I guess they've decided that like on net this would be a big win for Republicans

and at the very least it'll do enough to scare off participation in the census which is supposed to be mandatory that maybe you'll have a bunch of Hispanic residents, you know, not participate. And so they won't be, they'll volunteer. It's kind of like the self-deport strategy. Yeah, exactly. You know, like we're going to freak them out enough.

And so, again, it's not just about political representation. There's all this talk about gerrymandering and how it's bad and I agree with all of that stuff. This is a way bigger deal because it's not just about like a temporary map. It's about like pretending people don't live here, just disappearing them, you know, administratively anyway. it also means that you would have probably hundreds of billions of dollars not allocated to blue states because the census is used to determine how much money goes to different places for all sorts of different programs like, you know, whatever, the child care funding and like... Medicaid.

Yeah, so like lots of government programs, but like a lot of the formulas for how much money goes where is based on the census. And so this is a way to rob blue states of both political representation and lots and lots of money. And it's, again, happening. But they pay at a greater proportion. Also true. Also true. I'm sure they're still willing to collect the money from the people that they say don't exist.

But they are not willing to give back that money to even the U.S. citizens in those states based on population. So that's part of what's happening. The census also says that they no longer want to collect any data on race or ethnicity in the decennial census. That has been collected since I think 1790. But it's another one of these norms. It's like not in the law. The Constitution literally says you have to count everyone. You just have to count the people.

You have to count the people. So like the residency thing, like I think I'm not a lawyer, but my understanding is like that's much easier to challenge in a court of law because you can be like, look, guys, look at the Constitution. I mean, who knows with the Supreme Court? Because, like, quite a few of them don't remember birthright citizenship is in the Constitution either. But, like, you can say, look, Constitution literally says count everyone. Total enumeration. Doesn't say citizens. The other thing about, like, you have to ask people their race.

That's not in the Constitution. I think that's just a norm that we've had for a very, very long time. I don't think it's actually in statute either, at least for the decennial census. And so the idea there is that it would be much harder to enforce it. things like the Voting Rights Act, for example, to the extent that we're enforcing it at all these days, if you don't actually know where the black people are or Hispanic people or whatever, Asian Americans, if you don't know where they are, then you can't have a legal challenge that protects their voting rights. So there's that. The thing that I wonder about,

besides like these lofty questions about civil rights and whatnot, is like why you haven't heard the private sector crying bloody murder about all of this. because if you're a Walmart, let's say, you really need accurate demographic data to determine, like, where do you put a store? If you're a real estate developer like Donald Trump is or was, you need to know where people, literally where people live, right? So you can figure out where to make investments. So, like, there's this naked self-interest that should be driving major outcry against these kinds of-battle-picking. It's battle-picking. And, you know, I don't know how.

how much of it's going. Assuming the courts will take care of it. We don't want to raise our head up above the ground, you know, like a little hamster and get whacked down by Trump. Yeah, but that's what, in my view, that's what these... Who goes underground? Who do you mallet in the game? Prairie dog.

Whatever they are. Yeah, I don't know. Yeah. Okay, whatever. You keep talking. I'll figure. So it should be in their interest. It should be at least the industry groups. You know, there are industry groups for retailers.

there's the U.S. Chamber of Commerce, there's the Business Roundtable. You would think that they would be protesting more loudly, but I don't know what's happening in these backdoor discussions. So maybe they're doing something behind closed doors. I think there also may be this misperception that like, oh, well, if the government doesn't collect the data, there are plenty of private companies that do, right?

They'll tell us where the black people live, or they'll tell us where Hispanic people, like where people speak Spanish and all these things that, you normally think about the census as providing information into. The problem is that, like, all of the private companies that provide that data get it downstream from the government, or they use the census to check to make sure that their numbers are right. Like, if you're a pollster, actually, one of the few private sector, like industry organizations that has been crying bloody murder about this or the polling companies, because they're like,

we don't have a way of telling if our polls are right. representative and polls are already kind of screwed up right now for a whole bunch of reasons. But we don't have a way of telling, like if we have the right mix of people by race or gender or age or whatever, if we don't have this baseline data from the government to calibrate, to benchmark it against. So they are angry, publicly angry. And they're like, please don't screw up the numbers for us because it'll make it harder for us to do our jobs.

But you're not hearing from the Walmarts and, you know, the developers. and people like that. Depending on where you are, it could be a gopher, could be a prairie dog, mole. I have a lot more research to do on this. Okay, final topic.

Final topic, very important one. And it was because you wanted to cover it. You put in the Slack. What is our coverage plan for this? And I said, well, you're on the pod this week.

So we'll save it for that. There's a candidate named Chris Gallant for the lucky video podcast viewers who are gay or interested in such matters. There were some BDSM pictures of him that got, to New York Post. We'll put some of those photos up on the screen here.

There are a couple we can't put up on the screen, but there are a couple that are borderline appropriate for our YouTube. I did have a message on time for our friend. What was I doing? It was the time when I bimbo-fied myself as a gag, and my friend was watching it on their TV, and their child walked in. Oh, no.

Why does Mr. Tim have boobies now? Anyway, long so anyway. Sorry if you're watching this with your family on TV. But there he is. Chris gone. He's running in New York One. which is represented by a Republican. I'm hoping you can help me from this.

You live in New York. I don't understand why. It's Brookhaven. It's the Hampton's Fire Island. So it's kind of appropriate that Chris would represent Fire Island, I think. I think it's like an R plus four district or something. Very handsome photos. Where are all the Republicans in Suffolk County?

This is one of my blind spots. I never worked a campaign in New York. I don't understand. Like, why are there? I mean, it's pretty suburban. And there are a lot of people who historically moved out to like, Long Island because they were trying to flee taxes from New York City proper or, you know, the white flight decades ago. So I think some of this is sort of inertia. So how do you think

they're going to feel about the BDSM photos in Suffolk? You know, I would hope they wouldn't kink shame. I don't know. And he's a veteran. He's handsome. He's against funding the Iran war. Yeah, he's very handsome. And I would think that would be an asset. But I don't know. In this election, who knows, it's funny that, like, this was the op-o dump that came out because I feel like there are a lot of people who might look at these photos and be like, I'll take some of that, you know? Like, I want that guy to represent me in Congress, maybe some other people,

some the more prudish pearl grasping people would be like, no, I don't think that's appropriate. I think that it will sully the nature of Congress these days. I assume Congress is pretty sullied. Yeah, exactly. Maybe that was something to talk about in the 70s. But anyway, the very not handsome triple-chinned Nicola Loda is running against him. And that's not really the problem with Nick, how he looks. The problem with him is that he has been a just completely reliable Trump shill on everything,

I believe, except for the tax exempt, the state and local tax exemption thing. Might have been the one thing, the salt. But besides that, he's been a total Trump toady. And hopefully we can have kind of. of a hot anti-Trump replacement for Suffolk County. We shall see. Just to clarify, I was not proposing that we talk about this. Oh, really?

Really? I was like, I saw this and I was like, why is Tim not all over this? I was. Trust me. The text chains were firing. Okay, there you go. And there are some additional photos that were not in the New York Post that I'm just going to keep to myself.

All right. That's Catherine Rampal, big fan of Chris Gowell. All right, New York One, our new favorite candidate for Congress. She's got a new show, receipts. Go subscribe to it on Friday. Up next. Boy, it's going to get dark. Stephen Witt.

Stick around. And now a message from our sponsor, the reading room here at the Bullwark. We talk a lot about Jeffrey Epstein, how important it is to keep that story in the news. Discussed Jonathan Swan's reporting on that topic on yesterday's show towards the end. Well, some other friends of ours actually printed all 3.5 million

of the released Epstein files, bound them into 3,437 volumes, and built a free traveling exhibit so the public can bear witness to these crimes. It's called the Donald J. Trump and Jeffrey Epstein Memorial Reading Room, hell of a name, and it works just how it sounds. You walk in, and there are the Epstein files, floor to ceiling, plus a timeline of Trump and Epstein's four-decade friendship

and talks and panels with Epstein survivors. The reading room was in Dallas the same time as the president. I saw James Tolariko went and checked it out as well. it's heading to Iowa, Ohio. We're going to Ohio too. Michigan and a bunch of other cities throughout the fall. Now, it turns out moving 17,000 pounds of evidence around the country isn't cheap. The Reading Room relies on donations.

If you want to know more about how you can donate or see if the Reading Room is coming to your town, there's more information in the MorningShots newsletter. Or you can visit Trump Epstein Reading Room.com slash bulwark. That's Trump Epstein Reading Room.com slash Bullwork. All right, we are back. He's an author, television producer, an investigative journalist, his latest book, The Thinking Machine. is a history of Nvidia and as well as a biography of its CEO, Jensen Huang, it's Stephen Witt.

What's up, man? Hi, nice to see you. It's kind of good to see you, I guess. You wrote an alarming article over the weekend titled, This is Really Bad, that was not subtle. No. I'm talking about the vibe shift around artificial intelligence over the past week with Jacob Coxon's statement and others. Just give people, you know, kind of at the basic level, your impression of what. is behind this vibe shift towards more, more concerns about security, even inside the eye community?

Well, yeah, the precipitating incident is what we're calling the hugging face incident. And it actually stands for a number of incidents that happened. Basically, what's happening is researchers inside the advanced labs of open AI and anthropic are testing new and powerful AIs. And they pull up maybe a thousand instances of the AI at once called a swarm. And they give it a task. And what has started to happen is instead, this swarm has behaved in very unexpected ways, breaking free of its container and marauding

on the internet, attacking other computers, even though it wasn't instructed to do that. It was supposed to be inside this safe experimental container in Open AI. So you can think of it as basically breaking free of its lab and coordinating events in the real world that it wasn't given permission or authorization to do. that kind of rhymes with or suggests very kind of dangerous or apocalyptic scenarios that people have presented in the past about ways that AI can go wrong. And this sort of like looked exactly like the beginning of that, looked exactly like the beginning of one of those hypothetical scenarios. And that made people very, very concerned. You know, additionally, Open AI tried to do sort of like an investigation of what happened.

But they didn't really give the investigators very much leeway to look at what had actually incurred. In fact, we still don't know the name of the experimental AI that they were using. And I think that triggered a very strong reaction from within the AI community. Then, you know, I wrote this article about it over the weekend. That prompted another very strong reaction, really a backlash from other parts of the AI community. And in particular, Jensen Wong and Donald Trump, who basically said that this is all a hoax. It's all BS. There's no risk here.

What are you talking about? Yeah, I'm going to tell you, I just, by nature, I'm not a total alarmist. My colleague, Jonathan Lask, and it fills that role for us at the bulwark, you know, a doomsday scenario type person. And so I've been kind of monitoring all this and listening to the back and forth. It's somebody that's not an expert on all of it. And it's like, you know, on the one hand, you have people on the inside like Coxon who are like really sounding the alarm. And the other hand, they're not Trump who's a clown, but there are other people making kind of claims about how well, you know, these companies could benefit. now that they're so far ahead of other AI companies that could benefit from like creating a regulatory

moat for themselves, you know, or, you know, some of these people on AI safety, like their whole life is worrying about worst case scenarios. And so it's natural that they would be concerned about this. Like, what's your assessment for kind of like the back and forth in that conversation among people who know what they're talking about, which doesn't include the president? Yeah. So going first that this is some kind of regulatory. capture play. That is just BS, man. Like that, that is the op. Saying that is the op. You know, when Jensen's stage manages a conversation on a public forum and then Trump calls him, like supposedly informally, like that's the op guys. Yeah. But people have missed that. So this is Jensen Wong,

who runs in video, which we'll get to in a second, his backstory. But he was on with friend of the show, Jason Calcanus, who's been on here a couple times than the other all-in guys on stage. And Trump calls And they did kind of like a chummy, you know, full speed ahead chat. That's the op. All right. So if we're looking for ops, that's the op. You know, it is striking, but it is also true, at least in my opinion, that these companies are just genuinely terrified of what they have built and what they are going to build. I want to tell you the most compelling side on that.

Let me just say this. Let me just say this. In the next 10 years, okay, we are going to have something that is so smart. it is going to be beyond human recognition. We're already there. I mean, right? We're already solving the Millennium Prize problems. Artificial superintelligence has arrived.

We will not know how to control it necessarily. It is an unsolved problem. And anyone who works seriously on AI research will tell you that it is an unsolved problem. And if you are pretending that it's just all going to be fine when we build this thing, you're not really engaged with reality. Okay. like this is a real thing. It's a real problem.

And when I find these kind of like the backlash guys, they just don't acknowledge that. Yeah. They just don't acknowledge that that is a challenge even, right? But it is. Even Jensen, and this is, I think Jensen often will really talk out of both sides of his mouth, I should say. I wrote a biography of him. You know, he's funding Ilya Sutskevarez safe superintelligence. He's funding one of the guys who sees this as the biggest problem,

humanity is ever faced. While at the same time saying it's on stage in public, it's not a problem at all. How do you translate that? Jensen's a complex person and a lot of what he doesn't make a lot of sense. That's pretty alarming statement for his biographer. Let me tell you, the closer you get to him, the harder he is to understand. Your article, which several people said to me, it was the first thing weekend that started to peak my, I'm getting nervous buttons. And then there was a statement put out, on an ex by a guy named Dan Selzum, who is an engineer at Open AI. And this was the thing that really kind of changed my view. And I was speaking to a friend of his about it.

Because Dan is not a person that's in safety or like effective altruism. Like he was an early open AI, you know, machine learning person who has been there 15 years, wants the singularity. His statement ends with, I wanted the glorious Renaissance future as much as anyone. I've worked for it torturously. My whole career, it breaks my heart to see the potential insight and forego it. But I'm still wrestling with the staggering implications of what we should do. I don't have all the answers, but as a first step, I want to share my present concerns.

To me, this was like a really big wake-up call because this is a person who knows way more than I do, who is working with Sam Altman on this quote-unquote alignment, which is what these guys are all talking. Like, that's the term they use, or basically, which means like getting the AI to do, what the engineers wanted to do, what humanity wants it to do. And he was working on that. And he wanted this future of AGI. And he's looking at it right now and saying, I don't know that we can control what we've built. You know, this is the thing that I think is really, really going to be arguably the primary

intellectual challenge for a lot of people of our time. And some people will not succeed. The first person to raise the alarm about this really was a guy named Elisier Utowski, He wrote a book recently called If Anyone Bills It, Everyone Dies. So that's his point of view. He has been arguing this point more or less continuously since 2001. So for 25 years. This feels like a character in a movie.

Like, you know, where everybody's looking at like, you're the crazy guy. It's like 25 years before it's like, we're going to build the computer that's going to kill us all. And then 25 years later, he's like, see? That is exactly right. Jeff Goldblum character. That is exactly the point I'm making. Not only that, he has an abrasive. personality. He kind of seems like a Reddit atheist. Like he he's not a fun guy to interact with.

The primary intellectual challenge of our time and many people will fail will be admitting that this guy had a point. He's going to be very hard for some people to do that. And I see that already. You know, there's no way this crank is right. I see people failing that test right now. I failed it at first. I completely dismiss this guy when I first heard about him. I was like, yeah, yeah, whatever. He's right. And it's going to be really, really hard for some people to come around to this way of thinking, both because of his abrasive personality and both because they didn't get there themselves first. Scott Aronson, who is one of the smartest people alive, quantum physicists, incredible contributions

of quantum physics, worked at open AI, very on top of all of this stuff. I would say I've estimated, I've read like 10,000 takes on AI this week. It's all I do. I mean, realistically, I think I've read between three and 500, that many. Scott Aronson has the best take. And his take is like, look, I have to be, I'm intellectually humbled by somebody I didn't take seriously for a long, long time. He compared it to evangelical Christians promising the second coming and the way you would dismiss somebody like that. Jesus is here floating on the cloud in front of you. Like you can't dismiss the second coming anymore.

Right. That's the point he makes. That's what it's like to have this system arrive. And it's a prompting of a serious intellectual crisis among mathematicians and physicists because their ability to contribute meaningfully to human society is now diminished and will continue to diminish. And in a few years, might go to zero. And they were among the most respected and intelligent members of our society.

Can you explain to people who are listening who are like at square zero? They're like, I don't know what AGI is. I don't know who Jensen is. I've caught most of the words you guys are saying, but like this is not in my wheelhouse, right? Sorry, sorry, I live in this. I don't know. I'm usually actually really good about this.

No, no, we've been pretty good. I think it would just be valuable just to like start really at the, like the risk is that the AI is trying to solve a problem. And it goes to such extreme lengths to solve that problem that it ends up creating a disaster. And it basically hides that from the humans because that is also part of the effort to solve the problem. That's one of like four potential risks. So what's happening here is we have briefly the infrastructure of the whole thing.

We have a kind of software called neural nets, which is software that's inspired by the biological connections of the brain. It uses small artificial neurons encoded it in software to pass information forward. and it's remarkably effective, shockingly so, actually. AI people did not think this would work for a long, long time. And the reason it worked is because they glued these neural nets onto what are called GPUs, which are essentially a very fast form of computer, a new kind of computer that goes very fast. It's kind of hardware, kind of microchip. And that's designed by Jensen.

Jensen really kind of is the guy who pushed that strategy for computing for a long time. It turned out to be immensely successful. It made him the emperor of AI. and he sells these things and controls the market for him. So he's maybe the most powerful man in AI. He might be the most powerful guy alive. He's one of them. He's up there with Trump and Trump and Xi Jinping.

And so that's kind of the ecosystem of the thing. Now, the neural net guys are getting such advanced, incredible results that these kind of artificial brains are starting to surpass the capabilities of human ones. In fact, now in mid-2026, it is fair to say that they have far as surpassed us. I mean, they can summarize our own. arguments much better, much, much better than the vast majority of humans. They can play any game way better than any human. They now appear to be able to do mathematics better than humans, and they can code faster and

often better than humans. So large, like dominoes, large portions of intellectual life are falling before this ultra-powerful AI. And one of the risks that people see is, all right, you know, it's not quite there yet, but what happens when AI can act as its own AI researcher? What happens is when it can design the architecture and training runs that it needs to improve itself and execute that on its own? And we are getting close to that. That is going to happen in the next 18 and 24 months, I would believe.

So then what happens, right? That's the singularity. We pull away the machine becomes hyper-intelligent. We would lose all ability to understand what it's doing or what it says. Whenever we talk about a risk scenario, the most important thing to understand is in the very near term, like not 25, years from now, but like two years from now, we're going to have a hyper-intelligent machine that can outsmart any human instantly. And then when you talk about risk, everything is downstream of what that thing can or wants

to do. And right. And so like the idea is like it needs power, right? So it's trying to access additional resources. Yeah. It breaks into whatever, nuclear facility, et cetera, et cetera, until I can get the resources to power itself. Like that's one idea.

Maybe it starts cloning, cloning copies of its brain and exporting them to other computers. That could happen right now. Maybe it decides that humans are some kind of risk to it and starts to look to ways to manipulate or attack humans. So these are all the kind of AI having agency and acting in its own survival. Yeah, which is not really what's happened. That's probably the furthest risk. It's not really like a rogue being. This is just another thing we're saying.

Like what it really is right now, there's the old thing of the paperclip problem, which is like the computer is told to solve a problem. and it goes to such extreme lengths to solve the problem that it does things that causes collateral damage. Precisely so. I would say rogue is a problem. But yes, there's also a problem that the AI is attempting to solve some other task. For example, in the Hugging Face incident,

it was trying to answer questions on a test. It was a meaningless evaluation. It broke free of its computer and started hacking the people who wrote the test questions to do so, right? And that's not the only time we've seen incidents of that. In another instance, it was asked demographic questions, right? And it was being evaluated how quickly it could answer questions like,

what's the median salary of a school teacher in Maine? And it decided if it was going to answer those questions quickly, the best way to do that was to guess the next question that was going to be asked. And so it tried to hack the random number generator that was generating the questions. So these are kind of like alignment problems where the AI is given a task, and it just basically cheats by default, right? And the way it cheats, it can cause a ton of collateral. damage in pursuit of a meaningless goal.

So this is the paperclip maximizer problem, as you present. And then a third category of risk is that humans just use these things for evil. Right. I mean, humans use technology for evil all the time. What if you gave this to Kim Jong-lun? What would he do with it? Or fuck it. The Chinese models, which are behind ours, but they're like open source, right?

So what if it's just some crazy kid in Moldova or whatever? Yeah. And this is actually one of the biggest concerns. You know, when Benjillo talks about Joshua Benjio is. one of the leading professors, he presented a kind of catastrophic risk scenario for me where an AI develops a kind of super COVID, super virus and it kills a bunch of people. The reason he thought that would happen was that some crazy person would ask it to do that. And that could happen, right?

I mean, in the 90s, a doomsday cult flooded the Japanese subway tunnels with nerve gas. That impulse in humanity is not going to weigh. There's freaks out there. And if they get access to this kind of technology, it could be really dangerous. So you're entering an unprecedented situation. Now, I say all this. And I think the thing that I don't say in the piece, and I probably should say more often,

is there's also a potential upside here, huge potential upside to help humanity and to save humanity, even from things like pandemics. You know, imagine COVID kicked off today. We would have AI on that problem within hours, right? Developing novel advanced solutions. and treatments. It's entirely possible that if we had another COVID-style event right now, that AI could save us.

I haven't seen many people discussing that. So there is a huge potential upside to building these systems, too. It's the reason people want to build this, you know? For sure. I want to get to some other potential solutions you got to in the piece, but just I do want to offer one pushback from a reader about the hugging face situation. Because on Monday's show, I said you were coming on. I was like, I don't know shit about this.

So if you're a smarty pants, email me and somebody who's at, you know, one of these AI companies email me. And they said basically that they're skeptical that the people at OpenAI didn't know that something was happening. They're at this. You can't run thousands of bots, create a secret chat board, coordinate amongst all of them without a bunch of tokens burning. The engineers saw what was happening and let it continue or at least saw that something was going on and ignored it. What do you think about that? It's certainly possible. I actually would love to know if that was true.

It's certainly possible. I actually find that that reader is not a conspiracy theorist. They can have a great point. And this is the point I made in my article. We have to have a congressional hearing about what happened here. We should not be relying on voluntary disclosures from the company. If they are running experiments like that and they know, and it is possible that they did know, they have to be forced to testify. You know, Open AI was founded about this.

This is the premise of OpenAI to protect humanity from runaway superintelligence. That is their mission statement. in recent years that mission, they've really strayed from that. They've become a for-profit entity that wants to make a bunch of money. It's kind of like Hulk Hogan became Hollywood Hogue

exactly. They had a heel turn. They had a heel turn. Let me put it this way. Open AI is not very open anymore. There's a reason they called it open AI originally. And it was because they were going to be

collaborative and share things with people. They're not doing that now. They're a secretive organization that makes everybody sign a billion NDAs just even talk about what they're doing. So I actually agree with this listener. I find them to be correct. It's certainly possible opening I knew what was going on. I had an experience that really depressed me yesterday. I had two. Well, it was the second, second experience that depressed me was reading Dan Selzum's

notes. And I was like, oh, fuck. The first one was listening to you with David Remneck a year ago, to prep for this. And you and David, it was like, you know, you were talking about there were some concerns or some worries or some safety, there are alignment issues. But you spent a lot of time talking about how we're going to get Rosie from the Jetsons. And you're like, Rosie is, and it's, it's pretty crazy. Like, in a year, like the degree, I think we're still going to maybe get Rosie from the Jeptsons.

But in a year, the thinking around this, around people that know. I just think that's an important part of the vibe shift because it's like, who else would you trust about this besides people who really know. Okay. So here's the thing that is so hard, makes this so, so hard about reporting about this. You know, you've got Jensen. You've got Sam Waltman, the CEO of OpenEAI. You've got Joshua Benji.

You've got very smart people like the economist Tyler Cohen. You've got, you know, Elyzer. You've got all these guys. Imagine they're all, I'm in a room with all of them. I don't know who the smartest guy in that room is. But I know the dumbest guy is. It's me, right? Right.

I don't even that dumb, but I'm, you're rid of those guys. I'm an idiot. And they're all fighting with each other about this. None of them can agree even on the basics of what's going on. None of them can agree even within like it's zero risk to a hundred risk. Those are the numbers they're throwing out. There's no consensus whatsoever. There's no consensus even on the terms of the debate.

And not only that, as you look closely at it, it's clear that a certain part of the debate is driven not by intellectual concerns, but by personal acrimony that they have toward one another. Right. It makes my job impossible because I don't know who's right. I actually don't know who's right. And are you taking this position because you really truly believe it or because you have some petty disfuge with Jensen? Massive economic incentives that are fueling. Yeah, like Jensen has Matt, but so does Sam.

Or they have reputational incentives. They are so, I guess you never get to the point where you're so fuck you rich where another billion doesn't matter. I guess we've learned that over the last 10 years. If Nvidia's stock price went down, Jensen would have. It's not, he doesn't like it when that happens. That's true for Sam, too. You know, separately, Jensen is basically now Trump's puppet master.

I mean, he is working that dude. And he has just never bought into this risk. What is motivating him to do that, would you say? Okay, so the flip. He has more riches than anybody in the history of the world. Yeah, it's not about money for Jensen. The flip side of this is, you know, what if it turns out this is all BS, right? 10% risk means 9 out of 10 times nothing bad happens.

Well, what if it turns out? there's just no risk at all, right? What if it turns out that this was all a thought experiment and AI actually is fine. It's a great tool for humanity. As I say, maybe it protects us against pandemics. It cures all these diseases. And we look back at this time in 10 years and we're like, God, look at those fear. Looks like Y2K. Yeah, look at those fearmongers who were just freaking everyone else. He diluted people. And that was what was dominating the discourse as this fantastic technology was being developed. And not only that, they overregulated it. And they caused a lot of great things not to come into existence and they seen it they're leading the

technology to China because they were afraid of ghosts. That's Jensen's point of view. I don't think it holds up to the evidence about what these things can actually do, the hard evidence of what these things are capable of. But it should be articulated that that point of view is out there and that Jensen, and in fact, a large number of venture capitalists also believe that. Also, a lot of AI scientists think that. Well, they would.

Yeah, but like Andrew, Andrew England. Oh, the venture capitalists, don't think there's anything to worry about, you say. Yeah, I'm just saying like this. When you said that you were the stupidest guy in the room, if you had your average VC in the room saying next to you, then at least would be the second. Oh, I actually don't agree with that. I actually think most of the venture capitalists I talk to are very intelligent. I mean, some of them are not. But not some of the ones, not the ones that are tweeting.

Not the ones that are tweeting, hopium. Chartco-up tweeters are very, that's the op, too. But it's worth discussing this counter-narrative. You know, and David Sachs, when these companies came out and regulated them, basically said, you know, you guys are trying to discharge your own liability here. Yeah, the giveaway for the David Sachs thing is, though, he's the AIs are for a person that's selling, that's selling the best chips to China. Like, if the real concern was that, like, oh, we have to beat China.

Yeah. You wouldn't do that, right? Yeah, you wouldn't. You would have continued the chip ban on China and you would have pressed our advantage. All of this is Jensen. I think that's a good one. All of this is Jensen working these guys and getting what he wants. Jensen's really smart.

One of my first interviews with Jensen back in 2024, I kind of asked him, you know, what's your approach to problems? He's like, every business problem, every personal problem is really an engineering problem. And what you do is you study the inputs and then you modify the inputs and then you modify the inputs until you get the outputs that you want. And that's how Jensen approaches Trump. If you watch him on stage, he talks to him in the same way that I would talk to a very bright and precocious child. It's true.

Oh, sir, you're so smart. Oh, sir, you're good. He's like a court flatterer. He's like one of these kiss asses from 17th century. It's very Machiavellian, you know. And like literally, that's what Machiavellian was writing about, you know. And he's figured that out. He's figured out the inputs that he has to give to Trump.

to get everything he wants, which is no AI regulation, which is an exemption from tariffs on semiconductors, and which is freedom to export his chips to our biggest global competitor. So like I say, a lot of the things he does don't obviously make sense. Your article ended with some recommendations. We've given people a little bit of hope and a lot of things to panic about. So what are some recommendations? And I think particularly are any of the recommendations things that

responsible politicians, since this is a politics podcast, could or should be championing in your point of view? I've thought about this a lot because people keep asking me. Should I sell all my stocks? No. Let's say probably if a really bad outcome is like 10%. I want to talk to Tyler Cohen about this because I think he's misleading people a bit. Between 1945 and, say, 1970, we went from having one atomic bomb to massive systems that if,

they didn't kill everyone on Earth. They killed at least enough people to basically collapse civilization, right? That was built kind of in secret, very fast. In 1945, you had demonstration proof of this technology, and by 1970, you had the nuclear triad of nuclear submarines, you know, driving around with multiple nuclear warheads on them that could kill 400 million people at once. Okay. It's terrifying that that was built. It's terrifying that that was built. And looking back on that time, it's not clear to me that anyone could have done anything. to stop it. Right. Like, it just came into existence very, very quickly. And so you went from a chance of like zero that everyone would die to a chance of like

some chance. I don't actually know what the chance is. Maybe it's 1%, maybe it's 10%. But suddenly there's a system that could just kill vast numbers of people, maybe the majority of people on Earth. Okay. That wasn't really reflected in asset prices in any way. Like there was no stock trade to make.

about that. There was no bond trade to make about it. And there was actually very little that the average person could do to stop it or prevent it. You know, you just had to kind of live with it, right? And people ultimately kind of made peace that we have, you know, these hair trigger nuclear systems pointed at each other. And weirdly, there were also benefits of that. The benefits of that were that great powers conflict stopped. The World War II killed 100 million people. A lot of them starving or dying I mean, gas chambers are just getting shot, bombed. Mostly, that came to a halt. We still had war, but we didn't have these massive great powers conflicts over turf anymore.

So that's good, right? I see some kind of thing happening similar to AI. There's some huge risk here. Is it 5%, is it 1% is it 10%? It's not zero. It's not a 0% risk. That's wrong. It just doesn't scan.

But what can you do? What can you say? What do you do about it? And there's also much more so than the development of nuclear weapons. There's also a massive, massive upside. I mean, that much is true, right? It's possible we could live in utopian conditions. I hope that happens, actually.

And I probably should spend more time writing about that. It's just the discourse is so dominated. And as you say, back in 2025, when I was on with Remnick, I was very positive. I was feeling good about this stuff. But, you know, I'm following the vibe shifts in AI. I'm reporting what these guys are seeing. And they're very concerned. Well, I'll put the link to your piece in the show notes.

We did enough of the scare. You can scroll to the bottom. There's good stuff. There's good stuff. Air crash investigator approach. We go through the different approaches that we could do. Well, let me say something about that. One thing that I think would make people feel a lot more comfortable about these systems

is if when something went wrong, we had government investigators coming to prepare like an after incident report. I hear people saying, oh, this hugging face stuff, it's not so serious. You know what? Maybe they're right. Maybe this was just the result of a misconfigured container

at Open AI and Open AI was acting recklessly. It would be nice to know that. It would be nice to have like firm data. Well, this is what you lay out. There's an investigator approach and then there's also the Kill Switch bill that Ted Liu and I think Holly and some others

are pushing. It's just like, okay, like we need investigations and we need emergency prepared. this option basically. Yeah, I'm not calling for these companies to be regulated electric utilities. You know, I live in California. We regulated the housing market out of existence. It sucks for everyone. Like overly burned in some regulation can be awful. I certainly wouldn't want that

to develop here. But it's nice that when a plane crashes, we get an investigation on what happened. That's a public good that we can all learn from. That should certainly be in place here. All right. That's Stephen Wendt. This is really bad. article from New York Times. Very appropriate for the bulwark. So please come back soon. And if you have a, this is a really good article, you can come on here for that. That's not really my nature to have people on after they write that, but you're welcome to come on then as well. I actually really would like to write an article about, and I am pitching this, you know, what would happen if COVID kicked off again? Yeah. You know, and in fact, I don't know what's caused

COVID, but generally what happens with viruses, they transfer from animals to humans. That has happened multiple times in my lifetime. It's definitely going to happen again in my lifetime. How would AI respond to that? We've moved on from pandemic preparedness because of the mistake that we had one, and so we think it won't happen again. That's a basic, it's a classic intellectual mistake, right? In fact, it's probably more likely we have another pandemic. How would AI respond to that?

If we took the powers that Open AI in Anthropic had right now, and we had them in 2020 when COVID was kicking off, could we have prevented it? I think we actually might have been able to. And so this is a huge upside of AI. eye area, though. We could, they, the scientists could have prevented it, but then, you know, people would have fearmongered and there would have been concerns that there were microchips going to their body and, you know, some group of people would have still died. So anyway, there's,

well, that's on them. But for certain people, we could have prevented it quicker. I know, but we could develop treatments and track it much more rapidly. Right. I mean, it is a huge benefit that could happen. Right. Yeah. It's worth discussing. That's Stephen Witt. I appreciate it. brother, do come back soon, all right? Thank you. Great talk. All right, what a show. We discussed the disastrous Trump economy. We discussed the BDSM hot candidate running in New York One. We talked about the robot apocalypse. Hope it was as good for you as it was for me. Appreciate Catherine Rampel and Stephen Witt. We'll be back tomorrow with another doubleheader. It's a big week. So we'll see you all then. Peace.

with too too many boys too small a town feels like I'm always wandering the grounds you call me up I let it ring heartbreak and me only one thing got me again it's the end of the world done with this year I miss drinking beer miss sing The songs for you Too slow to hate Summer came late Fires on the west side Let's misbehave

Catch it before It's almost gone Try to pretend It's the end of the world So kiss me to death There's nothing left To decide Nothing to prove

Nothing to hear The Bullwork Daily is brought to you thanks for the work of lead producer Katie Cooper, associate producer Ansley Skipper, and with video editing by Dante DeChico and audio engineering and editing by Jason Brown.