Jasmine, what are you tracking for the midterms next week?

Yeah, I mean, these are interesting midterms because I think it's the first election where people's feelings about AI, about data centers could plausibly have shown up.

I mean, I think the top level on it is, even though we saw a lot of discourse, a lot of anger about data centers in particular, that's been building up over the last year.

I did some reporting on it over the summer in Wisconsin and Michigan, two states where data centers were actually a really big factor in the primaries.

In the majority of states, I don't think they are yet shifting people's votes.

They're an 80-20 issue in the sense that people feel very negatively about data centers.

People feel very negatively about AI.

There's going to have to be a shift in how they approach AI regulation in Silicon Valley.

And I'd love if you could sort of talk through your experiences in terms of the people you talked to and, and sort of where the root of their opposition was coming from, in your opinion.

Yeah, for sure.

Um, so I, I totally agree with your characterization that there's an assumption, I think especially from folks in the AI industry and who haven't spent a lot of time in places where data centers are being built, that the primary thing going on is everyone's watching too many TikToks about water use, right? [chuckles] Um, and there is this pervasive idea that AI and data centers in particular require tremendous amounts of water, and that's the primary harm, and I think the reaction that a lot of folks in the industry have is: well, most of the new data centers all use closed loop cooling systems.

They actually don't use very much water.

That's a myth.

Because the water thing is a myth, therefore, the whole backlash is overblown and fake and based on misinformation.

And I think it's that last extrapolation that I found in my reporting to not be true at all.

... what, uh, what these two places are thinking about and, um, and why it doesn't sort of map out to what a lot of the rest of the country is thinking about.

Is this was this inevitable? Was this like just a happenstance of events? Is there a grand EA media comms conspiracy to make this all happen? How contingent is the past week that we've just seen? Do you think, Jasmine?

I guess I was wondering for a long time whether these two sides would collide.

Because I think there are internal debates, for example, within the EA community, whether the frontier safety people should engage with people like Tucker Carlson, who are seen as not aligned with them on a bunch of other issues, like maybe a sort of not epistemically rigorous, truth-seeking guy.

And then so in general, I think that there is a lot of hesitation from the more technocratic wings of AI safety about whether to talk to Bernie, whether to talk to Tucker, whether to try to make this a gigantic mass movement or if that would be destructive.

And I wasn't sure whether it was going to happen, honestly, because for a long time also the left and people who didn't trust AI people also didn't trust the safety people because they were seen as part of the same Silicon Valley tribe of SF and Berkeley weirdos.

And so I think it is surprising, but not out of the imagination to see that in fact, that coalition did emerge.

I think it was sort of like the left got over its distrust of AI safety and its belief that maybe AI was fake.

It's a pretty crazy

It's in spirit nice-

It's a pretty crazy level of access.

Like, they really are, like, you know, Daru's saying, and I think he plans on it, like, give the Meter people desks and Slack and code base access.

And, like, you know, I both-- o- on one hand, I'm like, yes, I support having more than just Meter because I think that Meter is, like, a very specific approach to AI safety, and it would be more credible if there were, like, a few other orgs not from Berkeley.

But at the same time, I can't imagine Anthropic being willing to let in some, like, nonprofit from, I don't know, Boston [chuckles] or whatever, um, and giving them full Slack access because we know how paranoid they are about security, and we know how little they trust most people.

And, like, the reason they trust the Meter folks is because they've, like, personally known the Meter folks for, like, five-plus years, and they, they've just decided, "Oh yeah, like, Ajeya should get Slack access." Like, we would be okay with Ajeya Cotra getting Slack access or Ryan Greenblatt getting Slack access.

But, like, I'm not sure that that transfers even to other people at Meter or other people at other orgs.

You don't come across as a lazy person.

When the tech companies come into these towns and begin talking to the city council members, when they begin talking to the community, what are they sort of promising on the one hand, right, this, this is why this will be good for you, and what are they asking for on the other?

Yeah, I mean, so these things do provide an incredible amount of tax revenue, right? And so, um, in Mount Pleasant, the, on the old Foxconn site, which ended up being bought out by Microsoft, they saw, "Oh, there's all this infrastructure already here.

Why don't we build a hyperscale data center on this unused, industrially zoned land?" Uh, Microsoft is on track to pay nineteen point six million dollars in property taxes in twenty twenty-six.

This is expected to c- to continue for many years.

Again, this is a small village of 28,000 people.

Um, and so there are really meaningful property tax benefits.

And I do think it-- I find it frustrating personally when anti-data center organizers say that the job creation is a myth because I think that 500, you know, maybe six-figure jobs for folks who go through apprenticeships but maybe don't need college degrees that last two to six years, that's enough time to build a family.

And how much of it is actually something that is more related to people's feelings about AI, which is I don't want this built here because why would I wanna pay the cost for a thing that I don't want there to even be more of in the first place?

So, um, yeah, tell me about that one.

Yeah.

And so next I went to Port Washington, which is I think it's the biggest data center project by acreage, at least.

It's still under construction, 750 acres on a bunch of farmland outside of Port Washington, which is sort of a wealthier, conservative-leaning city on the shores of Lake Michigan about an hour out from Milwaukee.

So I talked to the mayor there because he's sort of one of the most vocal data center proponents in Wisconsin.

Like every-- he was always on the news, he was always giving these pro data center quotes, so everyone kind of knew his name.

Very knowledgeable guy.

So you s- you had three sites in Wisconsin, and then the fourth one was in Michigan, right? What, what did you, what did you see in Michigan?

Why were companies using NDAs so heavily if they didn't have anything to hide? So

one of the most controversial pieces that I feel like I really underestimated until I actually went and talked to people is the NDAs really did so much to amplify distrust and break down negotiations between people, their governments, and these data center companies.

Even the most pro-data center people I talked to were like, we regret the NDAs.

It was a huge mistake.

It lost so much trust with people.

Because I think the NDAs basically are a dark money and politics story of like some big trillion dollar company is coming to town, paying off your local officials, not only not inviting public input, but actually banning the public from knowing how the negotiations are going.

And then you only tell people once the deal is already done.

Is it maybe good that these moratoriums are in place? Will that sort of force this slowdown that they claim that they kind of want?

How active, though, is that area? Like, how much interest is there in Wisconsin, Michigan in general about data centers, and what is it about those locations that, okay, it's like these could be attractive geographies for them given what they need?

Yeah.

I mean, so Michigan and Wisconsin sort of rank near the middle to the upper middle of the pack when you look at how many data centers are being built across the 50 states.

I think each of them have maybe around, like, 50, maybe to 100 projects or proposed projects in the state.

So it's not an incredible amount compared to, again, Virginia and Texas, where you have 600, 800 active data centers or, uh, data centers being built.

The thing that drew me to Michigan and Wisconsin, one is that I had heard that proposals for new data centers were ramping up, and that's because, one, Wisconsin, for example, has a sales tax exemption on the chips and on some construction costs, so that made it a somewhat attractive state.

Gretchen Whitmer, governor of Michigan, ha- who has always taken a pretty pro-growth standpoint for Michigan, how do we bring business back, was also sort of actively courting data centers.

And so even if you construct the world's most beneficial tax arrangement, you had a lot of people in your piece who were talking about, well, the company is just gonna figure out some way to kind of wriggle out of it.

Tell me a little bit about Mount Pleasant's experience with Foxconn, and how this serves a little bit like a microcosm to help us understand how these small towns' relationship with corporations in the last few years shapes their relationship with data centers going forward.

Yeah, so honestly, this was one of the most fascinating things to me, was that, um, so much of the antipathy towards data centers had been primed by very bad experiences that these folks had getting screwed over by corporations often working hand-in-hand with their local officials over the last couple decades.

And so it's even hard for me to separate, uh, separate out, like co- people don't like corporations and people don't like government.

The story I heard was much closer to an oligarchy or a dark money in politics or a, uh, sort of people are getting bribed and sold out type story, where the idea was that my local officials are colluding with or, or having the wool pulled over their eyes by big corporations to screw us over.

So to go to Mount Pleasant, this was one of the sites that I visited.

Microsoft has built two hyperscale data centers, the Fairwater sites, in this town of Mount Pleasant.

I think that has twenty-six thousand, twenty-eight thousand people in it, so it's not super big.

Um, like, what, what, what do you make of this lacuna, this, this, this gap in a local argument or in a local case for these data centers, despite the fact that certainly in the short run they clearly are job creators?

And so, um, I'm sort of curious as to, obviously on a local level, people who are contractors and doing construction and stuff are feeling the benefits, but is that, is there a sense in these communities there's a true debate over the pros and cons of the economic stimulus?

Yeah, I mean, I think the main case that these AI data center developers are making to communities is tax revenue, right? Like, we will pay a lot of property taxes.

And I think the problem here is that American culture right now, and I know Taylor's done a bunch of reporting on this kind of thing, it's like suffused with this sense of, um, people are really worried about distributional concerns, right? Like, people have a felt sense based on the past few decades of the economy that just because, you know, unemployment numbers look lo- low and GDP goes up and, like, the stock market's doing really well, that does not mean that most Americans, that does not mean that working class Americans are going to feel the benefits of that.

And so when people see things like, okay, maybe in 10 years after, like, this tax incentive district shuts down, if the data center is still around by then, it will start pouring property taxes into my city coffers, but do I also trust my city officials to spend it on the issues that matter to me? What are they going to do with that money? I think the sort of growth in the abstract, like revenue in the abstract is something that people have become a lot more skeptical of.

And especially when that tax revenue is being positioned against, like, all of these much more tangible and sort of concrete sounding harms, whether it is, like, construction is really annoying, or whether it's, uh, you know, um, what happens if this becomes a stranded asset.

But I think I felt a lot of, um, s- frustration from local officials in these communities because they're trying to figure out, how do I run the city and pay for public services like firefighters and, like, you know, uh, like traffic services and whatever, without continuing to raise taxes, right? And so everything's becoming more expensive to run.

Uh, education's expensive, like, infrastructure's expensive, but people don't wanna pay higher taxes, and so they're thinking, "Let's bring in this data center.

Um, how does, how, how do you see all of this tying into to what you wrote about the concepts of permanent underclass and how it's understood and perpetuated by Silicon Valley?

When the tech companies come into these towns and begin talking to the city council members, when they begin talking to the community, what are they sort of promising on the one hand, right? "This, this is why this will be good for you." And what are they asking for on the other

otherYeah, I mean, so these things do provide an incredible amount of tax revenue, right? And so, um, in Mount Pleasant, the, on the old Foxconn site, which ended up being bought out by Microsoft, they saw, "Oh, there's all this infrastructure already here.

Why don't we build a hyperscale data center on this unused, industrially zoned land?" Um, Microsoft is on track to pay $19.6 million in property taxes in 2026.

This is expected to c- to continue for many years.

Again, this is a small village of 28,000 people.

Um, and so they're a really meaningful property tax benefits.

And I do think it...

And how much of it is actually something that is more related to people's feelings about AI, which is I don't want this built here because why would I wanna pay the cost for a thing that I don't want there to even be more of in the first place?

[laughs]

Um, does that feel dissonant at all, you know? Um, and you know, you hear a range of answers to that.

I think that, um, for a lot of people, for some people I would say, the probably more, more common belief among researchers at the labs is yes, there are harms, yes there are risks, but if we really get this AI thing right, we are going to enter a world that is better than ever.

We are going to get the end to all disease because AGI is going to help us find the cures to diseases and humans will be immortal.

We're going to have all consumer goods and services be cheaper than ever because the space factories or whatever it is mean that like we're not gonna have to pay for a house or a car.

Maybe it'll be better when the AIs can do all the jobs and we're able to live off UBI and we can sort of engage in these lives of infinite leisure and focus on our relationships and our communities instead of having to engage in like demeaning wage labor or something like that.

And this vision I think is genuinely inspiring to a lot of people who work in AI.

Mm-hmm.

Like, what have you noticed?

I, I guess people just talk about it more.

I remember I, I'd bring up AI populism to people last year, and people would normally say, "Oh, it doesn't matter," like, the, the, "Technology always introduces some discontent," like, "People get used to it, you know.

Like, it's like the internet.

Like, people get annoyed, but they get used to it." And that was, like, the standard reaction I would hear last year.

Um, this year, it's no more of that.

This year, I mean, I think the reason that OpenAI and Anthropic have felt pressure to introduce, like, economic policy suggestions for job automation is because they're seeing people worried about it.

Yeah

What do you think people think the permanent underclass is, and what are the steps that you have to believe in order to believe that that will happen?

Yeah.

Um, I mean, so the permanent underclass is definitely not an idea I came up with.

It's sort of a meme and a term and a concept that's been going around AI circles for the last couple years.

The basic notion is something that's deeper than just short-term job loss, which is kind of what I became interested in.

It's not merely the idea that AI will replace some jobs and then people will have to find new ones, but rather that there is a limited window of time that humans can build wealth before AI and robotics are so advanced and so general that they can fully replace human labor.

Because at the point, if you were to believe that AI and robotics will become a one-to-one replacement for human labor, then Humans have no leverage from their own labor.

I would love a premium.

See complete disclosures at public.com/disclosures.

When you own your own business, you own every decision.

Now own the card that rewards you for it.

The Chase Sapphire Reserve for Business card brings the best Sapphire Reserve benefits to business owners who expect hardworking rewards.

Designed to meet the needs of business owners at scale, this pay in full card elevates your travel experience and offers premium benefits and value toward business services that will take your business to the next level.

Fuel your business and maximize rewards with 8X points on all purchases through Chase Travel, 3X points on social media and search engine advertising, annual partnership credits, and more.

Make every journey more rewarding with a $300 annual travel credit and access to a network of airport lounges, whether you're looking for pre-flight productivity or time to rest and recharge.

Hmm.

But what is that process of first starting to look maybe without judgment, like actually, how does that happen?

Yeah, I mean, I think that it I think of like, it starts with my notes app.

First of all, like I'm a big notes app person.

I don't believe in like having these complicated systems of whatever.

I just like, as soon as I have like a theme or an idea or the seed of an idea, I just start a note in my notes app.

Whenever I travel, I'll start a note that just has the place and the year.

So like China, 2026 or like New York, DC, November, 2025 or something like that.

And so I

Can you tell us what is meant by permanent underclass and are people being sincere when they talk about what is what does that idea mean in the Bay Area right now?

So the permanent underclass is this idea that I've noticed getting picked up traction over the last year or so in the Bay Area, especially in the AI world and the startup world, where a lot of folks expect that as a result of super powerful AI, artificial general intelligence, AI that can do everything a human can do and more we might end up in a scenario where you have a permanent underclass a permanent lower class of people where workers have basically lost their leverage their ability to earn income by wages because the wealthy will simply be able to rent super intelligent machines for a much lower price and at much higher quality than humans can and so the idea is that our class positions would be frozen in time far off in the future where nai is as good as people are Because workers wouldn't be able to sell themselves, rich people will use the AIs, and those who are existing owners of capital, those who already have a lot of wealth to spend on machine labor, will be able to become more and more productive and earn more and more, whereas everybody else might be sort of rendered useless and forced to live on welfare at best or doing extremely menial physical labor tasks at worst.

I will say that the permanent underclass is definitely in a lot of ways a meme and a hyperbole.

There's a lot of Twitter jokes about it.

And so I don't want to say that this is a very common, sincere belief that the majority of people in Silicon Valley have.

But the thing that I do think is very real about it are two things.

One is I realized as I was in conversations with people in the industry, startup founders, venture capitalists, that the vast majority of people I knew in the industry expected that AI would default increase inequality and decrease economic mobility.

You recently visited D.C. Can you tell us a little bit about that trip and kind of what you learned about the debate at that time?

What role does a company like OpenAI play in this, and how does their kind of like corporate philosophy feed into this idea of a permanent underclass?

Yeah.

So one of the things I was really interested in examining in the piece is like what is the role of these companies? Are they worried about it? Are they doing anything to stop their technology potentially being used to replace workers? Are they engaging in policy or sort of broader action? And OpenAI is like a particularly interesting case because I feel like they've actually been flip-flopping a lot over...

Which is, you know, maybe nothing new for OpenAI, but like they have really changed their tune on this several times over the past many years.

So like the thing about these AI companies, like OpenAI and Anthropic and DeepMind, is that they really believe they're building AGI.

They really believe that their goal is to build AI that can do everything a human can do, and their mission is to create this.

And I think like the OpenAI mission defines AGI as something that can do like almost all economically valuable tasks, which is basically, again, worker replacement.

AI does seem different, but, you know, as somebody that follows this industry more closely, like, where do you see it?

Yeah

... uh, w- you know, we, we need to do AI because China.

We need to g- reinvest in manufacturing because China.

We need to, uh, educate our kids better because China.

The, the specter of China competition and China eating America's lunch on the economy, on geopolitics, on whatever, was sort of used as an all-purpose justification in Washington, DC.

And I think that sometimes this is fair.

Like, again, I think some of the AI risks are really real.

You take super intelligence, and then not too long afterwards you get super robots

in different ways, "Hey, be a little weirder."

Some of it, but not all of it.

I mean, so I talked to, for example, James Yu, who is the co-founder of Sudowrite, which is one of the earliest creative fiction AI writing assistants.

I talked to some other folks who similarly were in the fiction writing LLM space.

And like you said, to an extent, a lot of writers are already using these a lot, are already leaning on LLMs to generate large amounts of text, and it can be very successful, and it can meet readers' needs and whatever.

Um, but, like, even these people who I was talking to, they were describing to me how freaking hard it is to undo all of the post-training that the labs have done.

Um, so they are, like, applying immense amounts of engineering effort that clearly when- in, in my conversations with them, clearly frustrates them, that it is so hard to get these models to stop being so chirpy, so sycophantic, so PG-13 and everything, uh, in, in order to get them to this sort of, like, base model state where they're able to be weird again.

Mm-hmm.

And I was wondering if you could kind of walk me through, like, how did Silicon Valley end up kind of aligning, at least from the boss perspective? with Donald Trump and get this sort of, you know, anti-woke sort of rightward ideology.

So there was a really illuminating conversation that Marc Andreessen, one of the co-founders of Andreessen Horowitz and prominent Trump supporter in the last election, he had with Ross Douthat at the Times where he sort of walks through his journey.

And I think he was actually being largely quite honest here.

During the Biden administration, there are two sort of dominant forces in the Democratic Party.

One was taking a pretty, like, corporate accountability type approach, whether that was Lena Kahn leading the FTC and pursuing antitrust action against a lot of big tech companies, or whether it was having quote-unquote AI doomers regulating AI and introducing things like the executive order to sort of enforce more civil rights and transparency requirements on these companies, or whether it was crypto regulation looking at oh, there are a lot of these crypto frauds going on.

FTX just collapsed.

We should have a lot more scrutiny of what's going on in crypto.

But I am curious, are you the last generation of writers? If not, why not?

Mm.

And rather than having the paternalistic institutions decide what the boundaries of your action are, as an individual, you should be empowered to research your own shit coins [chuckles] and to research your own peptides, and they shouldn't be making these risk trade-offs for you, you know? And so, yeah, the big pharma stuff was a big complaint that I heard from a lot of peptide users was like y- w- especially with the people who started off with the GLPs, the way they'd put it is like, "I'm spending over $1,000 a month to be on prescription GLPs, and then I find out that you can get literally the same substance for $100 to $200 a month from a gray market Chinese source." And then they start to go, "Wait, like, that doesn't make any sense.

Like, why is Eli Lilly making me pay ten times what I could be paying for the exact same thing?" And so, first of all, I think, like, a lot of anger with just, like, healthcare costs makes people go like, "There's something bad going on." Or even when you ask, like, "Why haven't most peptides been approved? Why haven't they had any clinical trials done," right? Like, "Why are there no clinical trials for BPC-157?" It's basically because it costs hundreds of millions of dollars to run clinical trials and get something FDA approved, and unless a drug is targeting a disease, and it's easily patentable, and you're gonna have a moat as a pharma company, most pharma companies are not gonna invest in extremely, extremely expensive clinical trials, and that is, like, a reality of the incentive.

So when people hear that, they sort of feel like it is rigged, and the FDA, by requiring this $100 million process, is not going to allow me to take the things that I want.

So I think that's probably my best, like, steelman of the place that people are coming from.

[pensive music]

I asked Jasmine Sun whether she'd tried peptides.

I've been tempted, but I haven't done it.

(laughs)

"They're amazing.

I gotta get you on these peptides.

I got a dealer.

I got my guy," and I'm like, "What are you on about?" And I think the first one that I heard about people using, which is the most popular that I have since heard, is Retatrutide, which is what some people call a GLP-3.

So basically, GLP-1s are through the FDA process.

People are, like something like 12% of Americans, depending on what survey you look at, are using them.

Yeah.