A (1:58)

Welcome back to the podcast, David.

C (2:00)

Thanks for having me.

A (2:02)

So now that I have everyone primed to talk about AI, I'm actually going to throw you a curveball, which is you used to cover data and elections at the Washington Post. So we're going to ask you to put on your elections cap for just a second. You're based in Los Angeles, and this week the race for California governor was shaken up in a real way because of accusations of sexual assault against Eric Swalwell, the leading Democratic candidate in the field. And it doesn't seem like Democrats have

B (2:33)

that many great options to replace him. And so, first of all, I want

A (2:38)

to get a sense of who Democrats might rally around, what. What folks are talking about in California.

C (2:46)

Yeah, so there's a whole bunch of Democrats running in this race, obviously. Cause it's a big blue state where if you go through the primary and you get nominated as the Democrat, you are very likely to win. So one of the top candidates, especially now that Eric Swalwell's out, is Tom Steyer. You might remember him from 2020 and that whole presidential primary. He is one of these kind of like progressive billionaire types, has some wacky ideas, some mainstream liberal ideas, but one of these guys who's made a lot of money and now wants to do a lot of politics. You have Katie Porter, who, you know is kind of a. Someone who was at one point sort of a shining star in progressive politics, has had some personal issues jump in some campaign missteps throughout the course of this. She's been sitting in the DDHQ average at third amongst the Democrats. You have kind of other people who are sitting and like the low single digits. You have Antonio Villar Grossa, who used to be the mayor of la. You have Matt Mahan, another mayor. You kind of just have a tier of people who you would expect to make a run for California governor, not all of whom you would expect to be nominated. But, you know, it looks like with Swalwell out, it looks like one of them will probably be in with Steyer currently leading. I mean, in some ways, you know, prior to all of these allegations and everything that's. That's unfolded there, Swalwell in some ways made as sort of the party's chosen son. He's somebody who has elected experience, has a little bit of a national profile because he had his. His 2020 presidential run as well. But, yeah, now, now that he's out, we're kind of all looking around and seeing who's likely going to be our next governor, I guess Steyer. But it's pretty open. These. These primary fields have a weird dynamic where, because there's so many Democratic voters and there's so many Democratic candidates, you have people shifting a lot faster than they would shift in, like, a general election environment.

A (4:46)

I have to ask. California is the biggest blue state with a huge Democratic donor base. You would imagine that Democratic politics there are pretty competitive. And it's ended up with this field that at the top of the field was an accused sexual assaulter. Then you have Katie Porter, who is accused of mistreating her husband and has all of these different allegations against her for how she's treated staff. She was on camera earlier during the campaign, kind of berating a local reporter for asking her pretty mundane follow up questions. You have a weird billionaire in Tom Steyer who was like the butt of a bunch of jokes up until now, like why can't California get its shit together when it comes to local politics?

C (5:35)

It's a great question. I think there's, there's a couple answers. One is that the state's biggest shining stars are running for or seem to be looking towards and waggling their eyebrows at running for president. Right. So like your two biggest names in politics in the state are Gavin Newsom, obviously, and Kamala Harris, the former vice president. Some of our sort of like biggest names either, you know, termed out exporting to national politics, what have you. That's one reason. Another reason is that, you know, California is a big state. I like living here. It's, there's a lot of cool stuff going on. But it's also a state that when it comes to politics, I think a lot of Democrats will even like admit that the last decade or so has not gone the way that they might have wanted it to. Housing is really expensive here. You have some numbers over the last decade of like in migration versus out migration that have not been fantastic. You have cost of living issues. You have a lot of things where if you're kind of a high ranking Democrat and Democrats have been in charge for a long time and you sort of admit that there's all these issues, it's a little bit hard to sort of look good in that light. It's a little hard to produce all those shining stars. And then I think that some of the other people who are potential governors, who you might think would be potential governors, we have a lot of like House members who seem to be like solidly in their House district, that's what they're doing. Maybe they're interested, maybe they're not, but it's not, it's not quite a fit. So there's just kind of a confluence of things where you'd expect more from California potentially. But yeah, it's just kind of not there.

A (7:18)

This gets us a little bit into my, I think final question about California politics before we get into AI, which is what are California voters motivated by in this election? I mean, maybe that helps us get to a place of having a sense of who's liable to win the primary, at least advance to the general election in November. Since of course, California has a top two primary. It seems like Democrats have not done the best job delivering for voters. Do they feel that way, are they mad at Democrats? What's the vibe?

C (7:56)

The vibe is, I mean, it's an overwhelmingly Democratic state. In an overwhelmingly Democratic state year, it was a state that Harris won by 20 points while losing the presidency. It's, it's barring some, I think, unlikely scenario where two Republicans advance through the, the runoff system or whatever. Barring that. I'm not like watching the horse race super closely here. It's been interesting because a lot of the polling on this race has been from partisan outlets. But one of the nonpartisan polls, Survey usa, gave some texture to this. So what they found is that Democrats in general and like some primary voters, if you look at like these, those cross tabs, they're interested in housing affordability, people want costs to come down. They're interested in sort of fighting Trump type of stance. This is something that Gavin Newsom has done that I think has worked politically very well, which is to take the fight to Trump, so to speak, be willing to be confrontational. I think that there's also some energy around like not doing crazy levels of immigration crackdown. This is one where, you know, it goes against the grain of what we might have been talking about a year ago with the way that public opinion was swinging and so on and so forth. So you kind of want lower prices and someone to be a sort of aggressive Democrat. And I think that that makes a lot of sense in a state as big and as varied as California with like all the crazy different interests. Those are kind of the things that like a gigantic number of Democratic voters seem to be able to agree on.

A (9:27)

Let's talk about artificial intelligence, which is also a pretty California based topic, but we'll talk about it nationally. What has changed over the past five months or so that's gotten us to the point where I think we can unquestionably say that AI has entered the political mainstream and is not some sort of abstract thing that we were talking about last November.

C (9:52)

A lot has changed and it's such an interesting issue because on a lot of political issues you have the rhetoric from politicians signaling to the voters what's important and what's going to matter and you know, what positions they should take. There's a whole political science literature about that. I think something very different is happening here. I think bottom up, societal changes, economic changes, worries that people have are shifting public opinion sort of in that foundation first way. So a few things that have happened. One, the job market continues to be tough. You ask any young person whether it's easy or hard to find A job, they're probably going to say hard. This is something that is just sort of known. There's a lot of, you know, people call it job hugging going on. It's just tough to find a job. At the same time, you have a lot of companies talking about AI when there are layoffs. Oftentimes people cite AI. In some cases these companies really might be replacing people with AI. In some cases, companies might be sort of planning for an AI future where they're leaner. In some cases, you know, a company might of overhired during the pandemic and they're just saying AI to justify layoffs. But the point is, is you have news cycle after news cycle after news cycle of essentially this tough job market and some type of layoff or some type of event at a big company related to or explained by AI in one way or another. That's what the CEOs oftentimes are saying. You have data centers, everyone's talking about data centers, like needing to be constructed in order for AI to advance. Time has passed. People are trying to do more data centers in order to meet AI needs. Usage is up. You know, if you look at the trends for weekly usage, it's something like half, I want to say, of Americans adults are using AI weekly. That's a massive sea change considering nobody was using this in, you know, 2022. And then the other thing that I think is so those are all, you know, important. And I think the other thing that is maybe a little hard to capture, but I think is underrated, is that for a long, long time, a couple of years, people were talking about how AI is going to displace jobs. AI is going to, you know, have this effect where all these tasks are automated. And then what you had over the last several months, however you want to draw the timeline, is an uptick in use of these agentic coding systems, your Things like your OpenAI codecs or your cloud code and your other products from other companies where essentially you have a sector of the economy, software engineers, that's well respected, that's widely seen as like, and correctly seen as like a hub of very smart people kind of en masse, suddenly saying, oh, this can do a lot like AI can do a lot of work. And they have different views, different software engineers about, you know, how many jobs it'll displace or what's going to be the impact or effect. But I think essentially what's happened is people have started using these things more, feeling the economic pinch, listening to what other people are saying. Whether they're kind of the software Engineer folks or CEOs talking about why they're changing the structure of their company and they're saying, oh, this is real. This isn't sci fi. I've touched the technology. I've read a million news stories about it. Something is happening. And that, I think, is the foundation for the change in public opinion we're seeing.

A (13:12)

Yeah, it's interesting in the sense that oftentimes something seems scarier in the abstract than it is once it becomes real. So just a year ago, according to Quinnipiac, just 37% of Americans said they'd used AI for research. In the most recent polling out within the past month, a straight majority said they'd used AI for research. That's a big shift in just a year's time. Yeah, but Americans haven't said, oh, now that a majority of us are using this technology and we find it useful, we're more optimistic about it. They're saying, we're using this technology and we are more pessimistic about it. It seems as though the more AI enters the culture, the workplace, whatever, the worse people feel about it, what's going on.

C (14:07)

So I have a theory about this, and I don't have the exact constellation of data points that I would need to fully prove this theory. So I'm going to present as. As a theory. It's something I wrote about on NRG substack. You should subscribe. It's called Real Research on Artificial Intelligence. But a lot of people are using AI, and they're using AI frequently, but they're basically using it like a search engine. They're using it to look up stuff. They're using it to, you know, learn something new. Maybe they, like, find a product on it, what have you. They're using it to look up stuff, and, and they're using it like a search engine. And they're also kind of understanding it a little bit like a search engine. So I asked this question in a recent poll. It was essentially a multiple choice question that said, you know, when an AI chatbot is doing what it does.

A (14:54)

This is not my exact question wording,

C (14:56)

but when an AI chatbot is, you know, generating answers, what is it actually doing? And I gave them four responses. And the most commonly selected response was that AI is looking up something in a giant database. A lot of people think that what they're doing when they use AI is getting like a really great search engine. So you have that on one side as the benefits for people, and then what you have on the other side is everything that I was talking about before, all the news stories that people are hearing about, you know, the potential for job displacement, you know, the potential for the end of the human race. If you're, you know, reading the Eliezer Yakowski book, you, you have all these other like harms that people are hearing about over and over and over again. My theory is that like there's something happening with sort of your average person that's a cost benefit analysis that's not quite computing in the positive range for, for everyone, you know, and you can, you can argue this a couple different ways. You can take that, that idea a couple different ways. You could say, hey, this is something where as usage becomes more sophisticated, people will see the benefits. And that really it's not just baseline usage, it's sophistication of use that's going to like be the unlock for people. There's also people who would say, like, you know, you just got to take the voters for what they say at their word. They're saying they feel bad about it, they feel bad about it the more they use it. So you can kind of argue different things from that point. But if I had to square the circle of why usage is up and why, you know, pessimism or bad feelings it's up is it's because of that sort of perceived cost benefit benefit based on what people are getting and what people are seeing, if that makes sense.

A (16:37)

Well, in some ways, and you've, you've referenced this already, but in some ways the CEOs are their own worst PR. Like the people who head up these large organizations like ChatGPT and Anthropic are in the press pretty regularly telling us that our whole world is going to be turned upside down if these people obviously don't need to be elected, but if they were politicians, the comms folks would be like, you gotta stop, you're scaring the shit out of people. So in some ways it's no surprise that when the people who are leading the organizations tell us it's going to basically destroy our lives as we know it, that people would say, well, that doesn't make me happy.

C (17:23)

Yeah, I mean I, I think that's very real. I think a lot of people who are sort of in high level positions believe that AI is like super powerful, is going to, you know, do all these things that disrupt the way that we do business and create all this kind of change. And you know, I think that people hear that and they're like, oh boy, am I going to be the one that gets disrupted here. Like, I, you know, there's a lot of ordinary sort of people who are hearing these messages and feel like they might not be on the winning end of that trade. And I, I do. This is a qualitative feel. I haven't done, like, social listening or something like that. You, you hear positive messages from people of like, oh, everyone's going to have, you know, their own startup and everyone's going to be like, making discoveries. And you hear stories like, I don't know if you saw this one about the guy whose dog was dying and managed to basically, like, vibe code a therapy that like, managed to give his dog extra years of life. You hear stuff like that. But I think there's just like a large volume of news on the other side which says, like, hey, look at this super intelligence which can do these, all of these crazy things and buckle up is, is the news that people hear. And yeah, I can, I can very much imagine how that would be influencing current polling numbers.

A (18:41)

And I just want to put some solid numbers to this. When Quinnipiac asked about excitement versus concern for AI, just over a third of Americans said they are very excited or somewhat excited, while about 2/3 of Americans said they are not so excited or not excited at all. So pretty clear majority is on the side of not excited when it comes to concern. 80% of Americans say that they are very or somewhat concerned about AI, while 18% say they're not so concerned or not concerned at all. And this is an area where seemingly the people closest to the technology in some ways, as we've already discussed, as you get younger, people express more concern about AI, not more optimism. We're speaking a little bit in the abstract here, but there are some pretty specific political questions that are live right now about AI, and I want to go through them. I'll bucket them into several different categories, which is one regulation, like, who gets to say what happens? Is it the local politicians? Is it state politicians, Is it federal politicians? And how strict should that regulation be? There's data centers in general, which has taken on a political life of its own. There's national security. And then perhaps lower down on the list, I wouldn't necessarily say that this is in the mainstream, but there's copyright issues and there's child safety, which I guess child safety is in the mainstream, but it's a broader child safety as pertains to technology, not just child safety as pertains to AI.

C (20:24)

Right.

A (20:24)

So when it comes to regulation, where are we right now in the Fight over the federal government versus the state, light touch versus more stringent intervention.

C (20:36)

All throughout this, the White House seemed to have made pretty clear that they want AI deregulated. They want to build, baby, build, grow, like, make the technology happen, take the restrictions off the table. And you actually see when you have even red states, I think it was Tennessee, if memory serves, I think Utah had something on child safety. But there are even some red states who kind of started to test drive or think about some regulations. And you've seen the White House actively try to push back and say, no, we don't want regulations. We don't want, like, more transparency laws. You've seen a kind of friction here. I think that part of the reason you're getting a little bit more of these attempts in the states is because, you know, we have, you know, Republicans in control of the White House who are, you know, very pro AI, and they don't want to sort of do federal AI regulation in the way that some legislators on the state level would want to. I think that part of it is also that Trump is getting a little bit less popular over time. And as Trump gets less popular, I think people test out the boundaries a little bit of, can I push something the White House might like? How are they going to react? What's the polling going to look like? You get a little bit less lockstep as sort of the leader of your party gets a little bit less popular. So you have some regulation going there and happening on the state level. Obviously, Democrats want to do regulation on AI. That's something that has become a priority for a lot of Democrats. It looks different for different contenders, but they don't really have the power to do that right now. So it's sort of a question of what are people willing to try to sort of get through the administration, which is, you know, I would say, fairly anti regulation.

A (22:28)

Let's talk about data centers. The White House put out a national blueprint for AI, essentially proposing federal legislation that would preempt any state laws. We're recording this on Wednesday. So it is actually today that Maine passed the first ever statewide moratorium on data centers. That's sort of above a certain size, and it's through the end of 2027. So it's not just like a blanket ban on all data center buildouts, but that's a pretty significant political development from where we were last fall. I'm curious, first of all, just for where public opinion opinion is on data centers. It seems like it's become. It's taken on a life of its own.

C (23:10)

In some ways it has. So I pulled up some data from Pew and it says it's about that's

B (23:17)

the end of today's preview. Head over to GDPolitics.com to become a

A (23:21)

paid subscriber and catch the full episode.

B (23:23)

We chatted for about an hour and got into some of the specific ways AI is shaping our politics, how popular some of the proposals are, and how

A (23:31)

ambitious politicians are positioning themselves on the issue.

B (23:35)

We also got to a good data, bad data or not data question on synthetic polling, basically asking AI to be poll respondents. It led to a recent op ed

A (23:44)

in the New York Times titled this is what will Ruin Public Opinion Polling for Good. We get into it.

B (23:50)

Like I said, head over to GDPolitics.com to become a paid subscriber and catch the whole thing. It's 8 bucks a month or 80 bucks a year.

A (23:57)

Paid subscribers get about twice the number

B (23:59)

of episodes, can join in the paid subscriber chat, and most importantly, ensure that we can continue as an independent podcast, prioritizing curiosity, rigor and a sense of humor. When you become a paid subscriber, you can connect your account to wherever you listen to podcasts so you'll never miss an episode. There's a link in the show notes explaining how. Again, head over to gdpolitics. Com.

A (24:19)

Hope to see you there.