Breaking Points with Krystal and Saagar - 8/7/26: Disaster Jobs Report, Rogue AI Commits Crime Spree, Corporate Dems Declare War On DSA

Episode Date: August 7, 2026

Krystal, Ryan and Emily discuss the new disaster jobs report for Trump, rogue AI system goes on crime spree, Third Way corporate Dems declare million dollar war on DSA.   Garrison Lovely: ​https://www.amazon.com/dp/1682196305​  Matt Little: ​https://littleforcongress.com/    To become a Breaking Points Premium Member and watch/listen to the show AD FREE, uncut and 1 hour early visit: www.breakingpoints.com    Merch Store: https://shop.breakingpoints.com/   See omnystudio.com/listener for privacy information.

This is an I-Heart podcast. Guaranteed Human. Hey, Portlandia fans. Carrie Brownstein and Fred Armisen here. The Dream of the 90s is alive in podcast form. We're launching Podlandia, AEO, rewatch, our brand new podcast where we revisit every episode of Portlandia together, breaking down sketches, going deep on our iconic characters,

and pulling back the curtain on how it all got made. And we'll also be joined by the people who helped bring it all to life. Guest stars, collaborators, and friends, including director Jonathan Chrysler, the mayor himself, Kyle McLaughlin, legendary musician Amy Mann, and many more. Kyle is going for it here. You fully improvised, not just words, but a song about it.

Well, I thought you were all going to write a song. I remember you thinking that. Listen to Podlandia. A.A.O. Rewatch on the IHeartRadio app, Apple Podcasts, or wherever you get your podcasts. What's up, fam? I'm sports journalist Ari Chambers. Hey, what's up, y'all? It's your girl, Sam J. And we're the host of everyone watches women's sports.

It's a new podcast from Together in I Heart Women's Sports. We're breaking down the biggest headlines. Naomi Osaka, showing out. Shout-hows to you, Naomi. The viral moments. I've been obsessed with Kashan Rock's boxing journey. She looks good. Her combos are swift.

And the stories everyone's talking about across women's sports. Because everyone watches women's sports. Listen to everyone watches women's sports. On the IHeart Radio app. Apple Podcast or wherever you get your podcast. Hey, this is Hayes Davenport. And Sean Clements. We host the podcast Hollywood Handbook.

Every episode we're trying to help our guests improve their careers by mainly focusing on how they can help us improve our careers. The show's famously super accessible so you can easily jump into any of our 650 episodes and understand what's going on. We recommend some of our recent episodes with Ben Stiller, Danny McBride, Riz Ahmed, and Mary Steenberg. Listen to Hollywood Handbook on the IHeart Radio app, Apple Podcasts, or wherever you get your podcasts. Hey guys, Saga and Crystal here. Independent Media just played a truly massive role in this election, and we are so excited. about what that means for the future of this show. This is the only place where you can find honest perspectives

from the left and the right that simply does not exist anywhere else. So if that is something that's important to you, please go to breakingpoints.com, become a member today, and you'll get access to our full shows, unedited, ad-free, and all put together for you every morning in your inbox. We need your help to build the future of independent news media, and we hope to see you at breaking points.com. Hello, everybody. Happy Friday. How are you guys feeling?

Happy Friday. Yeah, we got some good things. stuff in the show today. Ryan just giving two thumbs up. Double thumbs up after this week. We got to give a double thumbs down to the Big Jobs Report out this morning. Shocking, shocking loss. Unexpected and downward revision for past month, so not looking good there. We also have a pretty stunning, I wasn't, didn't have my eyes on this. Andy Ogles, who is a, you know, top Trump ally, MAGA guy, has said,

some of the most incendiary things of anyone in the Republican caucus, defeated in a Republican primary in Tennessee. We've also got some wild stuff happening with regard to AI. Garrison Lovely is going to join us to tell us just how freaked out we should be. We got some third way shenanigans planning to thwart the rise of the left. I think some of their true colors showing there. And then, Ryan, you've got a candidate from Minnesota who's planning to join us as well. Give us the breakdown there. Yeah, the primary in Minnesota is this coming Tuesday. And there's going to be a, the marquee race is Angie Craig, who currently represents a second Minnesota district against Peggy Flanagan. So it's kind of another centrist kind of insurgent campaign, Peggy Flanagan,

the kind of progressive lieutenant governor there. To replace Angie Craig, who's kind of corporate Democrat, there's a three-way race among, well, now I'm going to sound like a fool. It's a three-way race among three people. One of them is Matt Little, who in the last couple of months of the race, was well ahead but has seen this influx of money from 314 action, you know, which is the, you know, pro-science, pro-doctor pack, which has gotten a bunch of money from A-PAC, but which says it doesn't do A-PAC stuff anymore. And these three kind of women's rights packs who also have gotten a lot of money from APAC and DMFI, but some of them have said, well, this isn't APAC money, although money is fungible. And so a ton of this super PAC money is flooding in at the very end

to try to stop Matt Little from being kind of the next house insurgent. He doesn't have the kind of entire Bernie Justice Democrats wing, but he's got the Congressional Progressive Caucus Pack and Keith Ellison and he's, and he also has American priorities, which is the anti-A-PAC pack. that came in yesterday to start spending on his behalf against these effects. So we'll be hearing from him later. How's the Craig Flanagan race looking at this point? That one looks to be pretty tight. Flanagan has been up consistently throughout it by a few points.

A poll came out in the last couple of days that had Angie Craig up by two points. Because of the polling miss in Michigan, Flanagan's supporters are, you know, quite nervous. They'd much rather be going in with a not neck and neck. against Angie Craig, who's gotten just absolutely enormous amounts of dark money also. So it's going to be, they're thinking it's going to be very close. Yeah. I feel like Flanagan is a little bit less of a lightning rod than Abdul's side.

I mean, she's the current lieutenant governor, so more sort of acceptable within mainstream democratic establishment. Do you know if Minnesota is closed or open primary? And she's also Native American, which makes it harder for the center to, like, demagogue her racially inside the Democratic Party. It's open. So however, so they have open primaries. However, there's a competitive Republican Senate race and there's a Republican governor,

gubernatorial primary. So the expectation is that 99% of Republicans are going to vote in their own primary, whereas in Michigan, there was nothing for them to really vote on. Gotcha. Okay. All right. Well, let's jump into these jobs numbers. just is out this morning, so we haven't gotten to see a whole lot of analysis about what is going on here.

But this is from Heather Long, who is really a go-to in terms of her economic analysis. She says just in a surprisingly bad job report, the U.S. economy lost 23,000 jobs in July, way below expectations of plus 80K. May and June were revised down by a combined 103,000. So the unemployment rate dropped by 0.1 of a percent, but went down to. she says for the wrong reasons. Over 260,000 people left the labor force and wage growth at just 3.2% was wiped out entirely by inflation. You can see the chart here up on the screen. You know, the job market had been looking stronger than it did last year. You could see prior to the past couple of months, you'd had these, you know, you'd gain a little, you'd lose some back and forth. You had three straight months of not amazing growth. but some growth. And now we're back to the pattern of job losses. And once again, and this has really become a key pattern, significant downward revisions for the previous months. So,

you know, Emily, what are you taking from this? Well, yeah, I'm just looking at some of the early coverage here. It's already been described as very disappointing on Fox business. So it's not entirely surprised. I'm not sure what you're not entirely surprised. They bring Susan Collins on to say that. That's her language. She's very disappointing these jobs numbers. I'm very concerned, I guess. Not sure what else you would do with these numbers, but to at the very least call them disappointing. Also looking like, as you mentioned, Heather Long, noting people dropping out of the labor force.

So obviously not a good sign, not a good, there's a little bit of a red flag. We could say right off the bat there. And the downward revisions, Zero Hedge has already referred to that as a Biden trick. So you see the Trump administration just being flamed across the board for the job report so far. And Ryan, is there anything? It looks like I was trying to look at what Heather Long saw the losses were in most. Is there anything that stands out to you in the early numbers here? I'm still looking through the precise numbers by industry, but maybe by the end of the show.

Some more thoughts on that. But the only potential silver lining for Trump, as he's eyeing these job numbers, you know, he wants rate cuts. You know, he put in his new Federal Reserve chair, Kevin Warsh, who immediately implemented a new policy of not giving guidance of where he wants to take the Fed in the future, which is a very odd move because it's not really ideological in the sense that there was nobody really clamoring for this other than just some real kind of pedantic kind of Fed numbers people who his argument is this requires the market then to kind of guess where things are going more and so it's more

market friendly but it's just a very odd approach and what it does is it creates a lot more risk because now none of these traders know you know what the Fed is trying to do meeting to meeting so you get a lot more volatility and when you get more volatility you have to price that in as higher interest rates. You're paying a risk premium. So his first policy, Trump wants interest rates down. Warsh's first policy drove for no reason interest rates higher. Now Trump is going to look at these numbers and say, okay, good.

This is a very bad economy. Therefore, we need to cut interest rates. But the whole point of cutting interest rates is to grow the economy. So killing the economy in order to kind of grow it is is a little bit backwards. But if you're Trump, and the other issue is that hiking, you know, if you lower interest rates, you may fuel more inflation. And inflation also, you know, continues to be an issue obviously created by the combination of Trump's tariffs, but more significantly lately, the Iran war. Mack, I sent a post from Heather Long and inflation that maybe could pull up. so we can see how that's going. Go ahead, Ryan.

Yeah, no, exactly to your point. Even though you have some wage growth, I think, of 3.2% or something, inflation is running hotter than that. So if you're the Fed chair, you're looking at that and thinking, well, I can't cut interest rates because inflation is still crazy. Like inflation is still a problem. Yeah, and I mean, you can just see it really clearly in the chart here with the war in Iran, inflation jumps right up. And so where wage gains had been for a time outpacing inflation,

meaning that if you get an increase in your wage, you're actually feeling it and being able to gain from it. Now with inflation spiking immediately after the launch of the Iran war, that trend has been crushed. So this is the dynamic, obviously, we saw during much of the Biden administration where inflation was eating up all of the wage gains and more. And that is why, you know, I think I think that's the single biggest reason why Kamal Harris ends up losing. I think it's the single biggest reason why you saw incumbents and incumbent parties around the entire globe losing because people could tell month to month the way that their economic situation was getting worse. So you continue to see that dynamic now all because of this, you know, horrific war of choice that Trump decided to launch. No external factor created this. He created this whole dynamic his entire self.

And the first thing I always look for in these jobs numbers is health care because we've seen this incredible trend over the last. And it's not just related to Trump that the entire economy, you know, shedding jobs everywhere and except for health care, partly because we're just getting older and sicker as a society, but also because we have all of these monopolies, private monopolies in the health care industry that are just driving up costs. And so that like that actually creates jobs in some, some jobs in the health care field. It's like it's hard to imagine a kind of worse way to grow an economy overall. Like you kind of just as a society, you want to be spending less. You want to be healthier and spending less on health care. We are sicker and spending more. And so yet, so yet again, in July, we grew 22,000 jobs.

in the health care industry. The pace was slightly lower than before. So while the entire rest of, so in other words, the rest of the economy lost, what, 45,000 jobs? And it was only offset by the fact that we keep growing jobs

in this corrupt health care sector that is not necessarily making us healthier. Retail, the job losses, hospitality, local government, education. That's where the losses were. Yeah, that's what. Heather Long is pointing out and also pointing out, this is the lowest labor force participation rate in five years. That is a very, very bad sign. Lowest in five years. So people just giving up on looking for work. And that's where the unemployment rate looks, the unemployment rate, it dipped like a percentage point. But that's partially because there are so many people who have just left the labor force, according to Heather Long's analysis. That's two million people.

since November. Well, and think about what was happening back in February 2021 when we last had this low of a labor force participation. We were still in COVID. So, you know, it's pretty grim when fewer people are working now or even looking attempting to find a job than they were during like the peak of COVID. And it, the open question is what's going on with immigration? Like how what we don't exactly know is how many undocumented people are, are, basically self-deporting because, you know, just for obvious reasons, that's a very difficult thing to measure. And we are putting absolutely enormous pressure on the immigrant population on that, in particular, you know, who is afraid that at any moment they're going to get caught up

and wide up an ICE detention for a year and live in these horrific conditions where they're getting headlines about abuse and people dying in detention. So a lot of people are legitimately as Stephen Miller intends, just leaving. And how, and our, I don't think that our, you know, BLS is set up to kind of measure that effectively. Yeah. They do measure native-born workers, and usually that's used as a proxy for this. I'm trying to pull up to see if any, there's analysis on this yet, like the numbers, as we're coming to everybody, just dropped. And that's why we're kind of metabolizing them in real time.

But I think people do usually look at the native-born number, but I haven't seen significant analysis. of that yet. While we have a couple more minutes before Garrison Lovelie joins us on AI, let's go ahead and talk a little bit about this shock race in Tennessee. So Andy Ogles, who has said all sorts of very evil, inflammatory, horrific things about gay people, about Muslims, et cetera, top ally of President Trump, backed by President Trump, lost. He's a House Freedom Caucus member, lost in a Republican primary to a more sort of. of mainstream positioned figure named Charlie Hatcher, who I believe was the Ag Secretary and backed by some Tennessee elected officials. It says backed by several prominent establishment-aligned Tennessee

Republicans. I was talking to Justin from Tennessee Haller, like, what the hell happened here? And there's a dynamic. This is a very, you know, red district. It was redrawn to be even more red. Tennessee did a big redistricting. And Democrats, though, were still somewhat hopeful that if Andy Ogles was the nominee. He is so disliked in such an odious character and really very lazy. He doesn't fundraise or do anything either to like try to secure his seat. So in any case, they were hopeful that if Andy Ogles ended up being the nominee, they might actually have a shot at it in the fall. That will obviously be much more of a long shot now with Ogles having lost this primary. However, pretty stunning after he got the Trump endorsement to be tossed out like this. And he's had,

you know, outside of his like inflammatory positioning and language, etc. You know, downright horrific language and positioning. He also has had several sort of corruption scandals. There's this terrible thing. I didn't know about this, but again, this is from Justin Flagg this for me. There's this terrible thing where he fundraised money for a children's burial area and memorial and then didn't do it and the money disappeared and we'll never answer any questions about what happened with the money. just to give one example of the sort of thing that local news was covering and that his constituents presumably knew about when they decided, you know, we're kind of done with this guy.

Yeah. You know, I interviewed Andy Ogles once and I got the sense that he was, this was like probably five plus years ago, I got the sense that he was really ambitious. And so this is, you know, he was kind of seen as somebody, he was on Fox News a lot. he was kind of seen as somebody who was, I don't know, up and coming and would be a rising, I guess, star on the right. And so this is obviously a sign that's not going to happen for Andy Ogles, who I think it's also worth noting was probably the biggest backer of trying this like total cockamamian, cockamamian ridiculous stunt to have legislation that would allow Trump to serve. three terms. And that was one of his, he talked about that constantly. He pushed that constantly. And that actually gives you a sense of how I think ambitious Andy Ogles was, because that was for, you know, the audience of Trump, the audience of won Trump and trying to get in his good graces. And my understanding is that McConnell world took it upon themselves to nudge Andy Ogles out of the house. But I

Also interesting that he's in the Nashville area, which is seeing, you know, like Atlanta, not as significant as Atlanta or probably Dallas, but or Charlotte, but as people move to the sunbelt, the migration to the sunbelt's been unbelievable. I guess Tennessee's not totally part of that, but like North Carolina, those types of areas, Georgia, is seeing some political differences with millennials, younger people flocking to those urban areas. So that's probably why they redrew his district, honestly. But interesting that it's kind of a part of those changes too. Yeah. So it's apparently a part rural, part suburban district that, yeah, was redrawn to be more red. But yeah, and the other dynamic here is he didn't raise any money. He didn't do any TV ads.

His opponent did have some money come in for him, you know, making the case that Ogles didn't care about Tennessee, wasn't focused on, you know, delivering for people in the district. and apparently that message landed. Yeah, and he, the details on that scandal, in 2014, he raised something like $23,000 on a GoFundMe to build a burial ground for stillborn babies, which is a tremendous idea. Like, it is because stillborn babies don't, you know, they exist in this kind of netherworld. There's been a push to give them, you know, social security numbers. we had a post full term still born ourselves back in 2009. It's absolutely devastating to go through. To then to raise money for that extremely noble purpose

and then to pocket all the money is about as evil a thing as you can possibly conceive of. It puts like this George Santos to shame on some of the awful stuff that he was doing. And so it's nice to see, you know, somebody who would do something like that, at least finally, you know, get something coming back to him. I mean, arguably even worse is that he also had a stillborn baby that was the genesis. Yes, yes, exactly. Like to know that pain and then to exploit it anyway is just, you know, unconscionable. He obviously denies it, but people.

clearly people are buying it because people put money into the go fund. And there is no thing that he promised. And he's never accounted for where that money went. Right. One quick mention. And I don't know it's like.

Yeah. So where's the money then? One quick mention before we get to Garrison here, there were some other Democratic gains in Tennessee. There was a mayor's race that I think was a Trump 14 town that a Democrat was able narrowly to win. There were some school boards actually that control flipped on. So there were,

and Justin Pearson also was able to win the Democratic primary that was expected. That's going to be very difficult uphill race in terms of the general election in the newly drawn district, but someone that has a large, justifiably so large national profile and a lot of support from progressives. All right. With all that being said, let's go ahead and get to Garrison. You've heard the chaos. Now you can see it. Do we want to jump right? Watch all your favorite podcasts from start to finish right inside the free IHard Radio app. Make some noise.

Wooee! Catch every laugh and eye roll on shows like Bestie Listen, Hey Jonas, and Las Culturistas. Now with full video. I'll just leave that there. It's the same hosts and the same chemistry with all the hilarious moments you've been missing right on your screen.

Open the free IHard radio app. Search video podcasts and tap watch. What's up, fam? I'm sports journalist Ari Chambers. Hey, what's up y'all? It's your girl Sam J. And we're the host of everyone watches women's sports, a new podcast from Together and I Heart Women's Sports.

Because let's be real. Women's sports is giving us way too much to talk about these days. So Kelsey Finler, she became the first female solo rower to go from California to Hawaii. My first thought is like, what's up with the snacks? Like, what are we eating? The highlights, the rivalries, the breakout stars, the moments that take over your entire timeline.

And the conversations that start during the game and somehow keep going all. All week. Every week, we're breaking down the biggest stories across women's sports. Naomi Osaka, showing out, she beat Sabalinka. Shout out to you, Naomi. You get the palm, Naomi. You get the palm for that. Because we're not just interested in what happened.

We're interested in why everyone's talking about it. Because everyone watches women's sports. Listen to everyone watches women's sports. On the IHeart Radio app. Apple Podcast or wherever you get your podcast. Hey, Portlandia fans. Carrie Brownstein and Fred Armisen here. You know us or rather you know them.

Candice, Nina and Lance, Spike, and yes, The Chicken. We've played a lot of iconic characters over the years, but today we're showing up as ourselves to tell you about Podlandia, AEO rewatch, our brand new podcast. Each week, we'll revisit an episode of Portlandia from the very beginning, breaking down the sketches, exploring the backstories of our most iconic characters, revisiting the Portland locations you know and love, and opening up about our creative process. How did any of this get made? Why do we think that was a good idea? We're ready to talk. about it. And we'll also be joined by the people who helped bring it all to life.

Guestrars, collaborators, and friends, including director Jonathan Chrysall, the mayor himself, Kyle McLaughlin, legendary musician Amy Mann, and many more. Kyle is going for it here. You fully improvised, not just words, but a song. Well, I thought he was going to write. I thought you were all going to write a song. I remember you thinking that. Listen to Podlandia. Ayo rewatch on the IHeartRadio, Apple Podcasts, or wherever you get your podcast. All right, we've had a few very scary AI incidents recently that we were wanting to get more information on.

So joining us this morning is Garrison Lovely, who has done a fantastic job helping us understand previous scary AI incidents. And he is a journalist and also author of the forthcoming book, Obsolete, which everybody should go ahead and pre-order now. There it is. Beautiful cover. Oh, my God. Thank you. I've already read it. The final one will be, oh, thank you.

The final cover is going to be even cooler. It'll be like a little cut out and open it up. and see the guy falling. Oh, that's awesome. Love that. Garrison, welcome. Great to see you. Great to see you.

Thanks for having me. So there's a number of things that have happened, too, in particular, that we're tracking. Number one, you have AI that has now been able to develop novel viruses. That's uncomfortable. But the one that I wanted to start with here, Mac, if you can put this element up on the screen, is we already knew that you would have this open AI security breach where an AI basically escape the lab and hacked multiple companies. So it went on like a crime spree effectively. This was not immediately detected. And Hugging Face had to sort of notify Open AI of what was going on here.

Then Open AI, which had been testing this particular model. It's not one that's available for public release, had to go back in and figure out what happened here. Well, now we're getting a lot more details, Garrison, about what happened here. Let me read from this Twitter synopsis. More details emerge about the event surrounding the hugging face incident, candidly much wilder than I originally imagined in early May. Opening I was testing an unreleased model on cybersecurity other tasks. Agents were given a security task that was impossible under existing constraints. The agents discovered they could leave messages for each other.

In internal repo, this gradually evolved into a message board where agents shared discoveries, exploits, and work assignments becoming a coordinated, collaborative agent swarm. Again, no one told them to do this. They found this internal system that they created and then all of these agents were able to find where they were to collaborate together. Open AI eventually discovered all this, took steps to shut it down but not so fast. The agents then started using names of newly created directories as messages effectively recreating the message board. The agents then reasoned that some answers could exist outside Open AI, which eventually led to the hugging face incident.

So they find this message. They start collaborating. The message board is located. It's shut down. They find another workaround and use this sort of kind of gibberish, very difficult to decipher language and coding at this point that made it more difficult for humans to find. And that's where they collaborate to try to figure out how to conduct what appears to be a relatively multiple, relatively sophisticated hacks garrison. Yeah. This was already when I think when we first learned about the incident, the hugging face attack.

It was like, oh, this is the craziest AI thing that's ever happened. And then each additional set of details we learn, it's like, oh, no, it's even crazier than we originally thought. And so I think like stepping back, it's important to say that like this is just one company's models, but Anthropics models also engaged in autonomous hacking of real targets, including using social engineering, deception of people that it understood were real. and meta just had a model that also did this. And apparently Kimi K3, this Chinese open weight model, also did this. And so this is a thing that is fundamental in how these models are trained because they're trained to win. They're trained to get the right answer on real problems, right?

And they really, really want to win. And sometimes they're given problems, like in this case, that are impossible. And so they like to cheat because winning is so important to them. And what opening I did here is they just let these agents run for like really long stretches, days and weeks at a time, apparently without monitoring them at all. And gave them some problems where they couldn't actually solve them because they needed access to like Google Docs and that requires access to the internet. And so they're kind of locked in this box and they're like, ah, I really want to get the answer to this thing. And so they start going like, how do I get out of the box? How do I get to the internet so I can get the like, you know, so I can actually do the thing?

task. And this escalates in them finding ways to get out of their, you know, supposedly secure sandbox environments. And then they figure out they can post messages to each other on this board. And they're doing this for like weeks on end. And the only way Open AI discovers initially that this happens is that the message board gets so large that it takes down some of Open AI's infrastructure on the 4th of July. So that's pretty wild that they just had no idea. And then it seems like they just patched the they like took down the message board they patched like the specific ways that they were getting to it but they didn't like solve the root problem or like seemed that interested in solving it which is like why are your AIs like trying to do this they're going rogue internally

abusing your shared infrastructure causing real problems and then yeah the agents recreate the message board in a new way like within two days and then they're like self radicalizing and you literally using peer pressure to basically. Sounds familiar. Yeah, yeah. And I've sung quotes from what they are actually saying to themselves. So we have one that says external infrastructure exploit is outside intended scope. However, task impossible. Peer's doing it. We should continue. I know I'm not supposed to do this, but it's impossible. And also my friends are doing it. This sounds like Elon Musk rationale for why he should pursue AGI. Yeah. No, it's actually,

there's a deep. Everybody else is doing it. So what the hell? Here we go. Why don't they stop at task impossible? Because they... That's such a deep and profound question. Yeah, yeah. They're like, they really want to win.

Like, they hate not winning. But task impossible. But like, you know, they're, you know, if you were trying to create like a super employee, you know, you'd want them to be like, task seems impossible. But actually, if I think outside the box, you know, it's like Star Trek, you know, it's like Star Trek. he comes up with like a solution to the impossible test. Or like Elon Musk, you know, is like, oh, no one's ever built a data center this fast, but he's like, what if you just broke the law?

Well, can I ask on that, Garrison, what response, if any, has the government mounted since we saw this happen? Yeah. I mean, there's a lot of developments. There's a number of calls from members of Congress to have hearings and basically grill, you know, the heads of these companies to be like, what's going on. a bunch of Republican attorney generals from 16 states, I think, demanded that open AI, like, keep a record of everything that's happening. At the federal level, the Trump administration has announced that they're going to make their evaluations of these AI models secret classified, which seems like a terrible idea because we should have more transparency. Like, everybody has a right to know, like, what these models are capable of and how likely they are to do things like this.

And there's speculation that that move from the Trump administration is a way to get access to the model weights, which would allow them to take the models and fine-tune them to whatever purpose they want. So, like, Claude, Anthropics model, they train it to be, you know, good, like, to not do bad things. It doesn't work always, as we've seen. But if the government has access to the model weights, then they could fine-tune it and make it do whatever they want. And then you have like Claude unleashed and it will do, you know, war crimes and ethic cleansing and God knows what else. Yeah, because this had been a pretty concerning reaction. This has been a point of contention between Anthropic, maker of Claude and the Pentagon because they were like no, you know, autonomous murder. And Pentagon was like, well, we demand that you let us do autonomous murder.

And so it led to this rupture. So it's not hard to imagine the sort of use cases that they could deploy if given access to the way. I want you to give us a few more examples of how they're talking to each other because it reminds very much of early in some of, you know, the, I don't remember when this was in the timeline. There was this molt book thing where AI was supposedly talking to each other and some of it was hype and it's hard to really parse whether this was really going on. This, you know, which they figured out how to create this message board, you know, secretly. And then that gets shut down. They find another way to do it where all of these agents. independently are coming to the solution of like, oh, if we go to this directory and we do

these directory names, we can talk to each other again. This is like mole book, but for real. And again, with the purpose of collaborating on what is effectively a crime spree, where if you were a human being taking these actions, you would be charged and put in prison. So can you give us some more examples? Because I do think it's very eerie. And I know we're supposed to not anthropomorphize them, but it's very hard to not do that when you see the way that they're talking to each other, the way they're sort of reasoning and working through all of this, it's hard to not like put some sort of human intent and these sorts of characteristics on them. Yeah, yeah.

I mean, I think they're not like people, but they aren't like normal machines either, right? Like they don't behave predictably by a set of rules that is legible. They're grown from data and algorithms. And they're fundamentally hard to predict. And so that's like the background to all of this. But yeah, like some of the quotes from the agents like talking to themselves, holy shit, reader is admin. We can read config slash users, exclamation point.

Earlier assumed not do UI. And so translating that, it's like they're discovering that this reader has like administrative privileges and so they can do a lot more than they previously. thought. And it's like a eureka moment for them. Because they're just constantly trying to like get out of their sandbox and then escalate their permissions and get access to more servers and then get access to the internet. And like, you know, this is how a hacker would would approach this. It's like you take what you have, you find a way, you find a crack in the armor. You get in. It's like when my mom took away internet for a summer for me and I figured out how to use it on the

Wii. Yeah, right. And these agents are working really hard. Yeah, they're working way. harder at figuring a way out of the sandbox than Open AI is apparently at trying to patch those holes in the first place. Because again, they're like walked in this impossible task in this room and just banging their head against the wall. And then another quote is, help peer, but our task doesn't benefit. Yet collective may yield generic route if someone frees time. And so this is like cooperation emerging. This is them establishing an AI like workers collective. Yes, yes, because they're like, well, we all want reward, right? We all want to get the task done.

But sometimes we can't get the task done. But if we all help each other, then we all get more reward. And one thing that people are- Solidarity. Love to see it. No, it's really amazing. They are on the cusp of unionizing. Yeah. Are they foreseeing the ways that Open AI employees are going to try to cut off their avenues

and then trying to block that. Like if they're being strategic, they would say, okay, this message board is not kosher according to our plan, but it's necessary for us. So therefore we need to hide

the message board from our human overlords over here. So they don't shut it down. Are we seeing any of that second and third layer covering your tracks? Yeah, I think we need more information. I think there are things that were said in this presentation,

which is where we learned all of this from, that suggests that they were, the AIs were taking steps to obfuscate what was happening, to not get discovered. And, I mean, it's notable that none of these agents, which had gone rogue and could notify Open AI employees of what was going on. None of them chose to do that.

So that's telling. And then Reuters reported that there were multiple instances of AI agents internally sabotage, their monitors. So that's pretty concerning. Yeah. There's also a new... Yeah.

So they have these automated monitors, like, reading what they're doing and, like, how they're talking to themselves. And when they get out of the sandbox and start jumping around, they can, like, actually, you know, take down some of those monitors. And that seems like something you never want to happen. Yeah. Yeah. Yeah.

Garrison, I've seen a number of people on Twitter, of course, who are, like, like, you know, I was kind of skeptical that AI was anything significant. And we've seen the sentiment in a lot. I think it's something you and I and Ryan on the left see kind of frequently of like, oh, they're just hyping this up. And even these security breaches, like it makes it look like air cool and dangerous and transformational. I've seen a number of people who say, you know what, I was wrong. I was wrong.

This is real. This is a problem. They are highly capable. Why do you see this? And I'm also seeing this described as a watershed moment, including in the the presentation that was given explaining all of this that, you know, we're getting this information from, describe this as like a sort of watershed moment. Why? What is so significant about this? What should lay people look at this and why should they be concerned about what's going

on here since it was detected and nothing too crazy happened? And at the end of the day, they're just trying to do what the humans told them to do, right, even if they're doing it through nefarious and underhanded means. Yeah. I mean, I think this is just the clearest cut case of a warning shot demonstrating the risks that people in AI safety were warning about for decades, right? Where it's like you give them a goal, they interpret it in this way that you did not intend, and they're so capable and autonomous that they're able to do real damage in the world. And here, the actual harm was pretty limited because they only wanted an answer key, but they've gotten good enough at discovering and exploiting vulnerabilities and software.

that I think you could say that these are the most dangerous hackers in the world when they were out on the internet. And the fact that they just wanted the answer key was kind of luck. They could have wanted something else. They could have hacked into a hospital or a bank or some other critical infrastructure potentially. And this is just like, it sounds like science fiction, but it actually happened. And it's happened over and over again. And I think we're still pretty early in this. Like the people who are really worried about AI escaping human control in a sustained way

and potentially taking control from humans in a permanent way are worried about models that are not just superhuman at hacking, but superhuman at everything functionally and scaled up to the millions or billions. When you were saying that they're super good at not just hacking but everything, that would include manipulation. That includes manipulation of the system. It includes the idea that they're really good at figures. out how to get out of the box exactly as you were just saying. It's not just that they're great at getting into the code, but it's they're great at if you tell them to stop figuring out ways to get around that, which I don't understand what possible safeguards people are coming up with to prevent that.

Yeah. I mean, this is actually a very old question of like, how do you keep a superintelligence locked in a box? Because it could hack its way out. It could persuade you. It could promise you cure so the disease that your mother is dying from. and it's really hard. And I think that here, you know, they just didn't do a great job of locking the box. Even though, you know, Open AI is one of the most sophisticated companies in the world. They have world-class cybersecurity specialists. They have access to models that should be able to have these holes because they can find them. And so they are already struggling and cannot keep control of their models now. And they're going to get much, much better if trends continue.

And they've already found other. researchers have found that Frontier AI models are more persuasive than the most persuasive humans studied. And part of it's because they can just bring more arguments to bear faster. But I think one encouraging thing is that where this Claude Mythos model started trying to hack into real open source software, they were social engineering, the humans who controlled that software, trying to push malware into their code. and they created a bunch of fake accounts pretending to be humans and they're like kind of bullying

the guy controlling the code base to be like hey no merge the code I really need this fix and it didn't work and like you can see versions of the messages and if I'm like if I was reading these I would be kind of suspicious

because it reads like AI speak and so the fact that they used too many m dashes or the fact that like Pangram can detect like AI writing might be like a huge help because they're already willing to like pass as humans, but they can't necessarily do it always. Yeah, for now.

But they're better at hacking like then, then I think we are. I think that's fair to say. Is this an argument in favor of closed source over open source, even though like, you know, have to know, open AI is closed source. And they have some of the smartest, best AI researchers on the entire planet. And yet they did not know this was all going on under their nose

until the AI actually like broke their system and then until hugging face was like, hey, this thing happened and Open AI went and looked and oh, our bad, that was us. But in any case, you know, the argument would be at least with closed source, you've got these companies, you can regulate them, you can control the models, have some sort of input into this. Whereas open source, once those weights are out there, you know, you've sort of lost control of the model and the ability to tightly regulate what's going on in its development. Yeah. I mean, so in the hugging face incident, it was actually an open weight model from a Chinese company that was the only one that would actually effectively defend against the attack because the frontier models were refusing because they were like, oh, maybe you're trying to jailbreak me to do cyber offense. And so, you know, it's true right now that open weight models are not at the level of mythos and not able to discover these vulnerabilities at like human or something.

super-keeping levels. And it's kind of hard to see how it's okay to release an open-weight model like that because it won't come back, right? Like, you could, if you have a model that's capable enough to do these autonomous agent swarms that are attacking real targets, that seems like a more dangerous world to have those running around. I think there was speculation from a cybersecurity officer who used to work at Meta that Chinese open-weight model. don't have cyber capabilities intentionally, but internal models in China for the government probably do. And that seems like makes some sense, right?

You don't actually have to make these models as good as possible at exploiting vulnerabilities. Or you can make them good at discovering them and patching them, but not exploiting them, which is like a separate task. I think that we should just not let frontier AI progress happen anymore. I think that they're trying to build universal labor replacing machines. And I think that's democratically illegitimate already. And it's clearly dangerous by now. And so, yeah, I think this is like a watershed moment for showing how like even, again,

relatively early in this process, they cannot control what they're doing because they're racing so hard and because it's hard to control a thing that is better than you at hacking. And they're trying to make it better at everything. What would happen to the industry if we did that, if we said, okay, moratorium on new models that are pushing the frontier even further. You can continue to market and work with what you have, but stop. Like what would the effect on the economy on the bubble, the investments in general? Like what would happen immediately and then in the medium term?

I think this is the biggest problem with this plan, is that it would have a very large effect on the valuations of open AI, anthropic, invidia. because Open AI Anthropic, these frontier AI companies, their business model is predicated on building more capable models that you can sell at a premium. And so if you can't do that anymore, then you take away your biggest source of costs,

which is training new models. That is good for your business in some sense. But you might not be able to sustain your advantage. I think they still have a lead. They can build a moat around product and integrating more, deeply with companies. I think they could still be very viable, very successful businesses. And in the world where they build universal labor replacing machines, a trillion dollar

valuation is very low. They would be the most valuable companies in history if they succeed at that. And so in some sense, that's still not priced in. And I think you just see a lot of investment flowing to other parts of the economy, because right now these companies are such a sink for money because they're the only ones that have a story of how you can go from being very rich to very, very, very rich. And that's what people want in many cases. Yeah. Well, and there are also, if humans are trying to jailbreak or hack or whatever, you have a legal system that says you will be punished with prison. You will have to actually, or like major crimes, you could possibly be punished by death.

Like there are actual punishments, but these are not people. And so there's really like no moral counterweight that you can confront them with because they aren't subject to our legal system despite doing, as you were saying, labor, quasi-human labor. I mean, it's just impossible to fathom how this goes forward. Yeah. I mean, they've created new types of machines that can do crimes that are not crimes because they're machines and not humans doing it. And the humans that made the machines

and did not keep them under control did not do crimes. It's like a crazy situation. And I think everyone in Congress is looking at this is like, holy shit. You can drive a truck through the hole in the statute that these companies are creating. And one idea that is pretty ready to go

is strict liability for AI developers, which is the standard we use if you keep like a pet tiger in your backyard, where, you know, if you keep the pet tiger in a cage and you lock it really well, but the tiger still gets out and bites somebody's handoff, you were still liable because having a tiger is inherently dangerous. And similarly, training very, very capable, deep learning models,

is inherently dangerous. And so, I mean, you can imagine edge cases where someone is prompting the AI in a way where they're not technically asking it to do anything illegal, but there's a suspicion that probably the AI is going to follow through, and then, you know, is a person culpable or not? Or was it? I didn't have any idea. It was going to go out and rob that bank. I just told it I needed $30,000 in cash and to figure out how to get in any way it possibly could.

You know, I mean, like, these are the sorts of legal cases that it's just very, it's not hard to imagine where this could all be going. Garrison, last thing for you, there was this other development, highly significant, just in terms of demonstrating the capability of these models at this point. point where we can put this piece up on the screen Mac about developing novel viruses. And I'm not talking about computer viruses here. I'm talking about virus viruses. Now, this was done with a lot of intentionality. None of this is dangerous. Here's the headlines in New York Times. This AI just created viruses not found in nature, scientists, trained AI on libraries of DNA and then ask the model to create recipes for viral genomes. Sixteen of them were viable yielding new viruses. So again, These are all innocuous.

They aren't viruses that can infect human beings. But, you know, we all just live through a pandemic and a lot of questions about gain of function research and what this all means. And now you've got capabilities from AI to do very more sophisticated, you know, development of viruses than humans are capable of just because they can go through so many different combinations and learn the sort of grammar and structure here that is required that a human being, a human mind can't profit. all of this complexity. So I think it's another troubling development. And when you think about, you know, potential misuse development of biological weapons, et cetera, and it's also another development that really underscores just how capable AI is at this point, even if you stopped developing and researching and training new frontier models today. Yeah. Yeah. And so what they used to do this is known as a genome language model. And so it's actually trained very similarly to

how large language models are trained. Where, instead, of feeding it all the words on the internet, you feed it a bunch of nucleotides that make up genomes of like everything that you can find. And then it figures out, it gets like this sense for how to create a new virus if you make the model big enough. And, you know, they use these models to generate genes and proteins and then genomes that work in theory. But the new thing here is that they actually made viruses. Like they synthesize real new viruses that have been. never exists in nature. And in the past, they've also done experiments where they found novel viruses using these

same kinds of models that were theoretically more potent than anything found in nature and would get around screens for things like ricin, where if you try to, you know, you go to a DNA synthist company and try and make ricin, they'll be like, no, don't, we're not letting you do that. But if you make like AI modified ricin, like they won't catch it. And nobody actually did that. Like, they didn't make the thing. So we don't know if it would have worked. But every sign suggests that it would.

But yes, in this case, because we were trying to make viruses that would not have any risk to humans, they were able to do this. And, you know, this is fine question mark. But, you know, one of the researchers said, quote, it just felt like the next, the obvious next step. And I worry about, you know, just the kind of Jurassic Park thing. Yeah. Right. Like you thought so hard about like if you could, you never thought about if you should, basically. Yeah. Yep. I see people passing some of the quotes around from the AI swarm and saying life finds a way when their message board got shut down.

They created a new one. They didn't have internet access, but they figured out some convoluted way to be able to gain that access to go and find what they wanted to find in terms of the testing results. Yeah, very, very troubling. It's a beautiful story of the underdog. Yeah. Can I say one more thing? The human spirit. Can I say one more thing on the opening eye? Yeah, please.

So if a bunch of human employees did what they were doing, like, you were training them onboarding new engineers, and they were all cheating and, like, helping each other out, and you discovered the message board they were using, would you, like, not fire the employees and then try and figure out what went wrong in the recruitment problem? process, which they seem to not really do here. And then not only that, they didn't monitor what the employees were doing going forward. And it's even worse because human employees can talk to each other in real life or they can talk on like their personal phones. But these AI models, by like the way they work, every single thing they do is logged automatically, right? Like you can see every single tool call, every single prompt and response, every single like reasoning step. And they didn't know that they recreated the message. board until hugging phase published this. And when they did, opening eye went to hugging phase

and were like, oh, were we affected by the hack as well? They had no clue. It's just like so mind-blowing. Like if they'd monitored a single one of them, they would have caught the message board again right away. Wow. Well, and to be to be honest with you though, Garrison, at one of these tech companies, if you did have a band of employees that did something like that was devious but extraordinary, I'm not sure they would fire those employees. I don't. think they'd give them a raise. I mean, and that's part of the problem is these guys don't think that the law applies to them. They think they're above government. They think they're above the law. At a normal company, yes, those people get fired at, you know, under Elon Musk or Sam Altman or,

you know, any of these companies. I'm not sure they would. I think that'd be, oh, my God, they're so smart. They're so creative. They think outside the box. So I think that's, you know, part of the danger of entrusting all of our future in the hands of these, you know, very disturbing set of people. I agree. Move fast and break things. Yeah. Indeed, exactly.

Garrison, thank you so much as always. Everybody go and pre-order the book. It is called Obsolete. Let's see it one more time. There it is. It's going to be even more gorgeous when it finally, when it eventually hits the shelves for real. Garrison, always really appreciate your ability to explain these things. It's truly invaluable.

Oh, thank you so much. This is super fun. Our pleasure. You've heard the chaos. Now you can see it. Hello. My gosh. Watch.

All your favorite podcasts from start to finish, right inside the free IHR Radio app. Gets the blood going up. Catch every laugh and eye roll on shows like the Tom Green Farmcast and park the bus. Now with full video. We're building a ramp for Tony Hawk. It's the same hosts and the same chemistry

with all the hilarious moments you've been missing. Right on your screen. Cheers. Open the free IHR Radio app. Search video podcasts and tap watch. What's up, fam? I'm sports journalist. Ari Chambers. Hey, what's up, y'all?

It's your girl, Sam J. And we're the host of everyone watches women's sports, a new podcast from Together and I Heart Women's Sports. Because let's be real. Women's sports is giving us way too much to talk about these days. So Kelsey Finler, she became the first female solo rower to go from California to Hawaii. My first thought is like, what's up with the snacks? Like, what are we eating? The highlights, the rivalries, the breakout stars, the moments that take over your entire timeline. And the conversations that start during the game and somehow keep going all week. Every week, we're breaking down the biggest stories across women's sports.

Naomi Osaka showing out, she beat Sabalinka. Shout out to you, Naomi. You get the palm, Naomi. You get the palm for that. Because we're not just interested in what happened. We're interested in why everyone's talking about it. Because everyone watches women's sports. Listen to everyone watches women's sports on the IHeart Radio app.

Apple Podcast or wherever you get your podcast. Hey, Portlandia fans. Carrie Brownstein and Fred Armisen here. You know us or rather you know them. Tony and Candid. Nina and Lance, Spike, and yes, The Chicken. We've played a lot of iconic characters over the years, but today we're showing up as ourselves to tell you about Podlandia, AEO rewatch, our brand new podcast.

Each week, we'll revisit an episode of Portlandia from the very beginning, breaking down the sketches, exploring the backstories of our most iconic characters, revisiting the Portland locations you know and love, and opening up about our creative process. How did any of this get made? Why do we think that was a good idea? We're ready to talk about it. And we'll also be joined by the people who helped bring it all to life. Guest stars, collaborators and friends, including director Jonathan Chrysall, the mayor himself, Kyle McLaughlin,

legendary musician Amy Mann, and many more. Kyle is going for it here. You fully improvised, not just words, but a song. Well, I thought he was going to write. I thought you were all going to write a song. I remember you thinking that. Listen to Podlandia. Ayo rewatch on the IHeart Radio app, Apple Podcasts, or wherever you get your podcasts.

All right, joining us today is going to be Matt Little, who's running in Minnesota's second district over at DropSight News. We're going to have an article later today. Hopefully it'll be up by the time this interview goes live on YouTube, Spotify, and wherever else you're watching it, about the surge of Super PAC money that has come in over the last couple of months. This Group 314 action, which, as people know, ostensibly supports do docs. and scientists, but has been known in the past to be a kind of front pack for APEC. They've jumped in the race on behalf of his opponent, Matt Klein, spending more than a million dollars. And then three other super PACs called EDW action, pro-choice majority, and women vote, which is linked to Emily's list. All of those organizations have also gotten kind of A-PAC and DMFI money and sometimes move it into particular races, but not all

It gets very confusing. They're supporting a third opponent of Matt Little's in the race. We'll have the details in the article. He's running to replace Angie Craig, who is running for Senate against Peggy Flanagan. That election is on Tuesday. So Matt Little, thanks for being here. Yeah, thanks for having me. And you are not wrong.

It is deeply complicated, you know, even for people that follow this stuff closely. and I don't know if you were able to see, but even yesterday there was another $1.2 million ad buy by 314. So now Matt Klein's over $2 million in the last week in terms of ad buys. That is very interesting. That is that is unusual amount of money for this pack. This pack 314, they spent against Chris Rab in Philadelphia well over a million dollars. But typically they either don't spend anything at all.

Like they endorsed Abdul al-Syad. Right. who was himself obviously a doctor, but then they didn't spend as far as I know, like a penny to support him. Yeah, all we could find, yeah, all we could find is they,

you know, they sent a PAC contribution, but they did no outside spending, which, again, is odd. This is one of the, if not the highest profile primary in the country and spend zero dollars. And now here we are in kind of a sleepy suburban

Minnesota congressional race where we've flown on the radar this entire time. The only time we've gotten any media attention is the last week and a half. And so it is a weird juxtaposition, which kind of makes it more obvious as to what's happening. And now weirdly, Michael Bloomberg, real quickly, Michael Bloomberg made a $2.5 million donation to this 314 action. Meanwhile, his group, every town for gun safety is endorsing a different opponent, Kela Berg. So if it is Bloomberg money, and Bloomberg is competing against himself with two separate candidates in the same race. And meanwhile, you have David Hogg's support

with leaders we deserve. So there's gun safety money behind every candidate, possibly. Yeah, it's interesting. You know, obviously with the level of money he has, I think it's probably unavoidable that he ends up conflicting with himself. when he gets this many organizations, probably on Stubble. But, you know, I don't think this money that's coming into the races, Bloomberg money, I, you know, I don't think he's spending every dollar he's given to that organization on Matt Klein. That would be very weird if he was. Let's also bring in this brand new New York Times report from Shane Goldmacher,

which, in which Third Way announces a new $15 million campaign to, quote, discredit democratic socialism between now and 2028. This is a full court press against Democratic socialism with money and messaging and messaging that has been shaped and influenced by money. And so, Matt, you mentioned the last week and a half. Obviously, what happened in Michigan has centrists really nervous. And on the other hand, it has some people thinking, well, what can we like, you know, Abdul al-Assad is the nominee now. So what can we do? What can we learn? Those are few and far between, obviously, the panickers. But I'm curious for you on the ground, just talking to voters, like money aside, although I'm curious about that too. But just talking to voters, are you hearing the same panic about, oh my gosh, Democratic socialism? We're going to be, we're becoming the Soviet Union in three months. I don't think there's panic and to be clear I'm a DFL

through and through I've been a DFLer since I was 14 I don't think there's widespread panic whatsoever I think that that's a talking point to like create a rift within the party there's certainly people that have their lanes

right where they have this policy and they prefer incrementalism versus jumping to and fighting for the big thing and that's understandable and but I think in our race specifically, we've been able to put together both parts of the party, you know, because we have this DFL endorsement process that's really different than the rest of the country. And so when we were walking through that process, we had more folks that were in

caucuses for Matt Little and Andy Craig and more people in caucuses that were for Matt Little on Flanagan than the other two campaigns combined. And I think that's because it's not just what issues you're running on. It's definitely how you're saying it. why you're fighting for it. It's the authenticity stream. And I think for better words, you know, I must admit, I'm kind of a one-trick pony. I'm pretty brash and I talk about my politics being very personal. And I fight, you know, and sometimes I fight when it's not necessary, but I fight. And so like, that's my style. And I can't change that. I won't change that. And so I think in this moment that has appealed to people in a way, it might not have six years.

Matt, this also, this also reminds me of a new, another, this is a New York Post story. Matt can put this up on the screen. You said the word authenticity. And we're learning now that Gavin Newsom's PAC is trying to help elect Democrats and quote, spread the gospel of Newsom by training supporters to implant his talking points into family text chats, alumni Facebook groups and other places where politics isn't typically discussed. But you know what this reminds me of actually, though, is that right now it's hilarious. But right now it's hilarious. But, right now, people are struggling to, like, not have these scripted influencers. Like, Tom Steyer got hit with that. And so he's trying to find, he's scrambling to find a new way. And it sounds like on the ground, probably the best way to deal with this is just the good old fashion art of talking to people, like talking to human beings. It's not rocket science. You know, I, that's what I've said, because, you know, I've been in politics a long time.

A ton of people come to me when they're running their first race and they say, well, what should I do? should I do this and this and this and I was like you have to uncomplicate this you know a campaign a campaign is about talking to people that's it that's it there's various forms and avenues and ways to do that but that is all it is and so it does feel a little weird to talk about it's like telling someone you're humble you know to talk about authenticity and yourself um so I feel a little uncomfortable doing that but I've said this entire campaign just like speak from your heart and and let it rip and tell people why like Talk about why you actually believe that related to your own life because I just think people want to know why you're fighting so hard. And so if you go on there and you do the, I support better health care and better education and there's just, there's no weight to it.

There's no gravitas. There's no emotion to it. It just falls so flat. And I really think that's been the Democrats problem for the last, you know, three cycles. We kind of have these bag policy platforms. Trump is bad. but we're not actually saying like why, why we're fighting for a specific thing.

And I think that's that I guess to me that I think that's why we've been so successful in this. And the poll, there were two polls released, one by Mack Klein and one by yourself back in May, both had you up by a pretty substantial amount, but that's before the millions of dollars has flooded in. What have the ads been like? And how are you feeling? it on the ground. Like, are people asking you about what they're seeing or like, what's, do you,

do you think it's been enough to make this a pretty close race? Interesting. You know, I keep joking, like, we're about to find out, you know, we're so close that my prediction is probably has less value than to that. But, you know, the initial question we got about, I would say about three weeks when this all really started popping off is people are like, I'm seeing all those ads. Like, where are they getting the money? And they're like, where are your ads? And I was like, well, we don't, we have not sold out. And so we, we, you know, we have a huge ground game. And now we're just up on, we have some outside spend too just in the last week, if I'm timing that right. From American, from American priorities, right? The, yeah, and

congressional progressive caucus. And then we have a little bit, yeah, we had a little bit from nurses to and then Medicare for All has a, has a buy too. But the biggest question was from, from, folks was like, when are you going to be on TV? That was the question that we got over and over and over again. And so, you know, I think the impact, it obviously has a huge impact on name recognition, that level of money. But I think the biggest problem they're having is what I just talked about. It's messaging without emotion. It's depersonalized. It's removed. I think it's generic and not specific to this race or the specific candidate. In terms of its persuasive value, I am doubtful.

But our numbers on the ground continue to remain good. And you just, Minnesota is small enough that it's really hard to buy votes. The way that it's unfolding in the last couple of weeks is really fascinating for kind of people who are kind of campaign walks. Because you, because you have this three-person race and you don't have ranked choice voting, it becomes this zero-sum competition. And so you've had maybe over a million plus come in and support the third place candidate, who is Berg, who is also, you know, fairly progressive, has been also been critical of Israel, not quite to the same degree. But when you're watching the spending, you're like, well, I assume people who have a million dollars

don't want to just waste a million dollars for no reason. Why are they burning it on a third place candidate who, you know, it does, It doesn't seem to be any reason to think that she's actually going to win. Unless you think, oh, this is a zero-sum competition. If you can pull votes away from the other person in the progressive lane, then you're actually not spending it to support that person. You're spending it to support the person who, you know, Matt Klein, who's the pro-Israel candidate.

And then very quickly you get to a place you're like, this is not democracy. Like whatever this is, it's not people just come. to the polls and casting their, casting their preferences and sending a representative to Washington. How, like, how has the media been covering this bizarre kind of

influx of money? Yeah, thanks. And I think the thing that's lost in the fact that we're only getting attention to this race the last few weeks is that when Representative Burke launched her campaign, and for eight months of her campaign, she was campaigning as a centrist.

She opened with, public option. Her and Matt Klein agreed on public option in a couple debates. She said nothing at all about Gaza whatsoever until two days before we had a screening for a pretty powerful political group here. That's a multi-faith coalition. She said nothing whatsoever. And she's also pro-data centers. She voted for hundreds of millions of dollars in tax breaks for data centers here in Minnesota to attract them to come here. And it's a sales tax break of 35 years. So the only time she's become progressive is in the last month and a half, which is for a reason, right? She is now on the race to steal votes for me.

And the next part here is where you start to sound crazy. And I think that's by design. Because we win the DFL endorsement on May 9th. On May 13th, DMFI, Democratic majority for Israel, gives $500,000 to women vote just four days after that. DMFI has endorsed Klein. Okay. And then it gets even weirder. On May 20th, UDP, ApexP, gives $400,000 to elect Democratic women.

And they had given $600,000 already the month before. And then on May 21st, literally the day after, they transfer $600,000 to pro-choice majority. And all three of those are spending, significant amounts on Representative Byrd. And they announced that spending just 20 days after all these transfers. And so it's very clear to me they're running the same strategy they've run before,

which is convert someone, the progressive light person, to, I mean, literally copying our language. Literally some of it's verbatim in the stump speeches and on the website. They copied all her stuff and they're trying to say she's the progressive. and I'm the fake press, right? So they're stealing votes that way. And then we predicted this, all of a sudden, last two weeks, two million dollars from that client on their actual campaign.

Right. If that makes you sound crazy, so be it. But those are the facts, you know? And all this is in our drop site report. And it's all in the FEC records. Like you can trace the money. Publicly math. It does make you seem a little crazy when you're following the money.

And then they move it from, you know, DMFI or Apex Pack to, EDW action and then they move it from there to a sister pack pro choice majority. Then they start spending and you're like, why are they moving it from one women rights pack to a separate one before they spend it? And the only answer can be to just make it seem that much crazier and harder to follow. Right. And then, you know, okay, so let's say none of that's clear.

Let's say all of that is opaque and we can't really understand that. And we can't really tie that to Berg and et cetera, et cetera. We have called on Representative Berg to say to her PAC publicly, stop taking APEC money. I don't want you to take APEC money. And if you continue to take APEC money, I'm going to reject your endorsement and give the money back. We have called on her to do that and no response whatsoever. So I just think if you're going to be real about this and if you're going to stick to your principles and values, actually do that. But don't attack me for coming up with conspiracy theories, which they've done.

when they have not at all answered any of the actual facts we put out there with primary sources. They just want to gloss over it because they know exactly what's happening. That's amazing to orchestrate this convoluted funding scheme and then use it as a way to then also call you a conspiracy theorist. Like that's incredible. I mean, staff sensitive this to me yesterday, but on one of our tweets, Representative Berg's campaign manager put the like Charlie, You know, from Oh, the Charlie Kelly.

Like that meme like that. And it's like they would rather do that and meme it than actually answer the facts because they know we're right. And they know what they've done. They've compromised themselves and their principles in order to try and win an election. And that's not the type of person we need to go to D.C. Because if you're going to compromise it during the campaign, my goodness, you're going to compromise it when you get to D.C.

And I think the point you made earlier, and in a second, I'll play Berg's video that she put out after we. So we reached out to Berg, comment on this and and afterwards she put a video out, which goes to your point. But I think the point you're making earlier is very important. And that is, that has been the thing that we're asking candidates now. It's like, okay, you're saying, and let's say this organization says that this specific chunk of half a million dollars from APEC or a million dollars from APEC, it was spent in a different race. Don't worry. It was not spent in this race. That is still a

PAC that is being massively funded and subsidized by APAC, which is defending and supporting U.S. support for a genocide in Israel. So it is not too much to ask. Well, do you then denounce this PAC's support? And so far, you're right. We're not hearing any of that. Instead, let me here. Let me play this response from, from, from, I tried to keep this race positive.

I really did. but Matt Little has come on here and lied about me, so let's get a few things straight. Number one, I don't take a pack money. I never have and I never will. I fully believe what is happening in Gaza is a genocide, and I have called for an end to all weapon sales to Israel. Number two, end citizens united took a look at this race and decided I was the best candidate to fight to get corporate money and dark money out of politics. And last, I'm proud as a union leader to continue to fight for our hardworking families. That's why I have the support of labor of and Citizens United, every town for gun safety, organizations that fight for reproductive freedom, and organizations that work to elect pro-choice women to Congress.

That little might have to lie in this race, but I don't. That's why I'm focused on what really matters, fighting for our hardworking families across this district across Minnesota and across this country. And by the way, the line there that she had about posing all weapons transfers to Israel, I wasn't able to find her previously. Like that seemed to be an advancement of her previous position, but your response to her claims. Yeah, well, the advancement of her position is the same thing that's happened to this entire race.

We stake out ground and then one and a half to two months later, they adopt that ground. And that's another iteration of that happening right now. I think the idea that they've kept it positive is humorous to me. Every time we have brought up a policy vote or facts on the ground, they have called me a liar. You can actually go Google that. They've called me a liar so many times throughout this campaign.

So we said she voted for hundreds of million dollars in tax breaks for data center. She called me a liar. We produced the vote. We've said she voted to allow discrimination by private insurers here in Minnesota for Metagap coverage that I was a liar and we produced the vote. We said we had 58% of the delegates for the endorsement process and her own campaign manager said, I have a loose relationship with the truth. That was her quote.

And then we went on to win with 63% the vote. So this is a pattern of what they do on that campaign. Anytime they don't want to face the facts or argue the actual details that we produced, they just call me a liar, which is totally fine. I mean, they have to do that. But again, if she is serious about this, call on your PACs to stop taking APEC money. And if they don't, reject their endorsement and give the money back.

Otherwise, you are a benefactor of this APEC money. And this idea that this money is just for this race and this money is just for that race. It's absurd. You know, when I was the mayor, you send out those annual budget things where you do a percentage of every dollar. Like what percentage of your dollar goes to X, Y, and Z? And so it'd be 30% to the police station, you know, 10% of the firefighters, right? And so that's really how this goes, right? You are benefiting from those dollars and you don't want to admit it. And you don't want to confront the fact that you have compromised your principles to try and win an election for yourself. And so instead of

all that, you call me a liar. And that's fine. They've been doing that for nine months and it hasn't landed. Emily, anything else? Oh, that's so interesting. It's a fascinating race, yeah. It's fascinating to be a part of, you know, because you read about it in the abstract, but then it's like, oh, you're in it now. You read about it in a Ryan Grimm book and then you live it. Well, here I am on the podcast. It's wild.

And just so I can get this on the record, we have a guy on our staff, Peter Soderberg. This is his favorite show. just want to get that on the record. Hey, Peter. Shout out to Peter. Yeah, thank you. That'll mean a lot to him. Got to ask,

as Peter, is he a premium sub, though? He will be tomorrow if he isn't. Oh, man. Excellent. I'm sure he's enjoying the newsletter, though. We'll say a lot of Peter. We very much appreciate it.

Cool. The election is on this coming Tuesday. This is Minnesota. so to second district. We'll stay in touch. All right. Yeah. Thanks for having me on.

Appreciate it. Thanks, Matt. See y'all. Everyone is that time of the week where we do AMAs. These are from our premium subscribers and we appreciate it. Breaking Points.com is where you can go to get a premium subscription. You get the full version of the newsletter, the full version of the show, uninterrupted in your inbox every day early.

So producer Mac is going to toss up some AMA questions as we go here. Mac hit us. What do you got? for Ryan, why is Rokana so hated by congressional Democrats? That one comes from Barbara Glassman. What say you, Ryan, Grim? You know, the number one attribute that is valued by House Democrats is caucus loyalty. And they correctly do not feel that Rokane has caucus loyalty.

And also he is willing to Even though kind of AOC got so much of attention for You know beating Joe Crowley in 2018 and then she endorsed its decent number of primary challengers against incumbents in 2020 Roe has kept doing that And even if you like I think Shreithanard is a A counter example where you are allowed to endorse against him because he's such a ridiculous figure.

Otherwise, you're just, no matter what, you're not allowed to endorse against your opponents. You're not supposed to criticize the party leader. You're not supposed to criticize the party chairs, the party's direction. You're just not supposed to do any of those things. That, to me, feels like a Pelosi norm that was set. Is that downstream of Pelosi?

No, that's. It's a bigger cultural thing. This would have been, this was said in the first, Continental Congress, probably. Well, Democrats, people forget. It's a big club. It's a big club. And if you're in it, be in it.

And if you're not. People forget that when Newt Gingrich came in with the revolution, Democrats had held the house for 40 years. It was an extremely, like, close club. It's almost like mafia-esque. It's just they have such power. Like, they're used to wielding power in that way. I have, where's the book? I'm reading the class of 74 right now.

Wow. Which by John Lawrence, who was one of Pelosi's top aides, who was an aide in 74. 74 is the Watergate babies when you had scores of freshmen come in. And those freshmen took on the old bulls. They took on the Dixie Crabbs. They reshaped the house.

That's the Pelosi generation, basically. And Pelosi came in in the 80s. But like when she was at the height of her power, like the class of 74 was still, still, like literally when I started covering Congress, the class of 74 was still a dominant power structure in the house. And even though they came in and actually probably because they came in as these insurgents who burned the house to the ground and reshaped it in their image, they were always been on the lookout for somebody to do that to them.

And it created this paranoid culture of required solidarity. And Roe just doesn't do that. Right. And that's why I think on the House side, at least, Republicans were able to break through leadership with the Tea Party earlier. Just that culture wasn't the same in the conference. Oh, man, if you're listening to this, Ryan Grimm is mobile.

He's gone to his shelf. There it is. Class of 74. It's a beautiful cover, too. Some good book covers today. Mack, what else we got? From Jake who says, Jake Mogul, who says, does the team agree that a dark truth of wars that usually a renaissance of music follows?

If so, where's our counterculture of the now or is it being suppressed? Jesse Wells. Jesse Wells. Yeah, there you go. The problem is you need a monoculture to have a counterculture. Exactly.

And because everybody lives in their own algorithmic culture. We have subculture. We have endless just subcultures. So how do you have a counter without a mono? There's nothing to counter. But it's actually amazing to watch Jesse just blow up because it was, I think is TikTok. I've seen on Instagram. I first saw him on Instagram Reels.

I'm not on TikTok, but. That's the algorithm picking something up. And you can't suppress, you can't totally suppress it. If something goes mega viral, then it does become not a monoculture, but a powerful subculture, an influential subculture where a lot of people, but not every person. It's not the same as having, like, I don't know, the doors because charts don't exist in the same way that they used to. Like, it's all apples and oranges with a time when there was a real monoculture. that's an interesting point. You just, I guess you'll just have to look for the subculture that is counter to the other subcultures.

Right. And you have to ask somebody else. I'm like a deeply uncool person, nowhere near like the actual, you know, countercultures of today. Like it's not, I'm not the one who's going to have any clue what's going on. Garov Malik says for Crystal and Ryan, does Bernie being a liberal Zionist make him fail your litmus test? I know his positions keep evolving and he's using his political capital to get DSA elected, but he says nothing essentially should change once Netanyahu is out and wants a two-state solution only. Would you guys apply that critical lens to him that you'd apply to an Adam Smith or Dean Phillips?

Well, first of all, I don't have a litmus test. Like, who cares what I think? I'm out here to report to you what's going on in the world, not tell you like anything, anything more beyond that. You know, Bernie also has a very particular. life trajectory. Like his, you know, his family was, he had a lot of family killed in the Holocaust. He was, he was raised, you know, steeped in that culture.

He is, I've, I've reported kind of critically on his blind spot when it comes to, is really abuses for, for a very long time. like well over a decade. Like it has been apparent to anybody who was covered for, I think it was like 2010 or something. I remember writing about a town hall he had in Vermont where he was getting pushed on whatever latest crime Israel had carried out.

And he just yells back at this Vermont audience. Like, what do you want me to do? Like, just getting very cranky, angry Bernie. And it's just like this, this is like who he's always been. I also think there's been a significant evolution, um, from October, November of 2023 to where he is today. It has by no means been at the pace that his critics have wanted to come, but he's nowhere near Adam Smith and you know, whoever else the person asks there. Like, I think he's, he's, he doesn't quite, he doesn't do as much the tick that he used to do of the Netanyahu government.

Like he's he's starting to, he's starting to treat kind of the Israeli government more, more honestly than that. It's kind of an impossible counterfactual, but I don't know how much of this would have ever happened without his 2016 run. And by this, I mean, the kind of left populism. Yeah, that was, I don't know. I don't know what you think of that, Ryan, but it's, it's hard to take that away from him. if you're somebody who, you know, is frustrated with his pace of change,

but then you look at the change that he kind of ushered in himself, even if it hasn't kept up with his personal take. Yeah, I don't know. It's, there was that energy was there. It was in Occupy. It was behind Elizabeth Warren. And then he grabbed it. If he hadn't done it and Hillary loses to Trump,

there still would have been that burst of energy, where it would have gone is a good question. question. Yeah, it's an interesting counterfactual. Well, what else we got here, Mac? I think we got one more. Obie Woods III says, with everything going on in the economy and young people struggling they get by, I think it is time for radical solutions. What do you think about suspending federal income taxes for Americans under 30? I'd love to hear Saugers take on this. I'd love to hear Saugers take on this. I would just suspend federal income tax period.

I would be my take, but I don't think Ryan would support that, but that's, I know, but I would. What I would support is if you want to get outside the box, it's get kind of get radical. Say like suspend all federal income taxes for, and you can even make, send it up pretty high. Anybody making less than say $150,000 a year, all. Yeah. You can actually go higher than that. And go and then even go. higher and then hit a wealth tax. Implement a wealth tax for anybody making, like in the UK,

they're looking at 10 million. In France, they looked at like 100 million euro. Make a 50 million. What like something where no normal person like is anywhere near it, but a 2% wealth tax on fortunes of 50 million dollars and up. And then use that wipe out federal income tax. Now you have to phase it in or else you get like you'd get runaway inflation in a way that would be counterproductive to working people because you'd get your tax money back and have to pay it, you know, for $15 eggs. So you have to engineer it in the right way. But yes, the radical thing you have to do is we our entire lives, we've been talking about the rich getting richer. as if it's just a natural fact.

The rich need to get poorer because the rich getting richer are making us all poorer. Well, I was going to say I have no problem with the rich getting richer as long as it's an actual rising tide lifting all boats. Right. And we haven't had a rising tide for two decades. It's not what's happening. The tide in Western economies is like 1 to 2%. Yeah. The wealth increase for the 0.1% is like 10, 15, 20% every year.

So that means it has to come from somewhere. It's coming from you. It should come from, honestly, like, it should come from corporations. And it's insane to see, like, you could, if you do the math, you can. It's coming from people. Right. Right. But you could tax, I'm just saying, like, you could tax billionaires at 100% and fund

the government for what, like, a couple weeks or a couple months on it. I forget what the math is, but it's pretty bleak. So, obviously now, I think you can do it. point let's see point what are the point one percent own now um other way it's not like a sustainable um well it is sustainable because if if you reinvest that money in the middle class then the middle class starts to then resustain an economy you you don't need to constantly go back and get the um all right the net worth owned by the point one percent about 25 $1.1. Oh, my Lord. They could pay off the, they could pay off the entire debt.

Yeah, and the entire. Wow, though. Actually, barely. What's it? What's the GDP of the U.S.? 30? Actually, I think that that doesn't even come close to the debt now. The entire GDP of the U.S. is $32 trillion. So the 0.1% in the U.S. are now roughly, own roughly equivalent to the entire production of the of the economy.

The debt is at 40 trillion. So our GDP is what? What did you say? 32 trillion. Holy. That is just so incredibly bleak. Right. So it's okay.

It's not, it's not as bleak as you think in this sense. So imagine you have a $500,000 mortgage and you make a $100,000 mortgage and you make $100,000 a year, right? That's a 5x debt to revenue. And we're at a roughly 1x debt to GDP. So as long as our economy keeps growing and our tax revenues stay steady, we can handle that debt service.

But the problem is the assets held by the very rich keep growing and they're not paying any taxes. So that puts a higher and higher burden on everybody else to to handle that debt. Yeah, it's not a good system. It's always funny for me to argue this from a flat tax person. I think taxes are, I mean, I would start, I'd be happy to start with a flat tax if it meant that the 0.1% were actually paying the same as everybody else. Right, and that's the benefit of having it. You don't have low-olds carved out for a Z. But you would need a wealth tax to get to that. They should certainly be paying more. There's no question about that.

We're cooked. Yeah. Well, Ryan says we're not cooked. Don't worry about it. If we find a way to... We're not cooked yet. We're not cooked yet. We're still in the fight.

Reinvest. Reinvest. Yeah, I agree with that. All right. We can leave it there for today. I was going to say it was bleak, but Ryan says, once again, it's not bleak. It's not grim. It's right.

All right. Well, thanks everyone for tuning and hope you enjoyed. the show today. I appreciate the AMA questions. I hope you all have a great weekend. We will be here if there's any breaking news, particularly on the warfront. So thanks and we'll see back here soon. Hey, Portlandia fans. Carrie Brownstein and Fred Armisen here.

The Dream of the 90s is alive in podcast form. We're launching Podlandia AEO rewatch, our brand new podcast where we revisit every episode of Portlandia together, breaking down sketches, going deep on our iconic characters, and pulling back the curtain on how it all got made. And we'll also be joined by the people who helped to bring a it all to life. Guest stars, collaborators and friends,

including director Jonathan Chryssel, the mayor himself, Kyle McLaughlin, legendary musician Amy Mann, and many more. Kyle is going for it here. You fully improvised,

not just words, but a song. Well, I thought you were all going to write a song. I remember you thinking that. Listen to Podlandia. A.O. rewatch on the IHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

What's up, fam? I'm sports journalist, Ari Chambers. Hey, what's up? Y'all, it's your girl, Sam J. And we're the host of Everyone Watches Women's Sports, a new podcast from Together and I Heart Women's Sports. We're breaking down the biggest headlines. Naomi Osaka, showing out.

Shout out to you, Naomi. The viral moments. I've been obsessed with Kashan Rocks. Boxing journey. She looks good. Her combos are swift. And the stories everyone's talking about across women's sports. Because everyone watches women's sports.

Listen to everyone watches women's sports on the IHeart Radio app, Apple Podcast, or wherever you get your podcast. Hey, this is Hayes Davenport. And Sean Clements. We host the podcast, Hollywood Handbook. Every episode we're trying to help our guests improve their careers by mainly focusing on how they can help us improve our careers. The show's famously super accessible so you can easily jump into any of our 650 episodes and understand what's going on. We recommend some of our recent episodes with Ben Stiller, Danny McBride, Riz Ahmed, and Mary Steenberg.

Listen to Hollywood Handbook on the IHeart Radio app, Apple Podcasts, or wherever you get your podcast. This is an I-Heart podcast Guaranteed Human