The Ben and Emil Show - BAES 170: Why every AI CEO suddenly wants to slow down

Episode Date: September 17, 2026

The AI leaders are spooked. Or are they? What aren't they telling us? What ACTUALLY scared them, and what the hell happened this last week that's got everyone talking? BEN & EMIL ON SMOKING OUT NOW: https://youtu.be/t2wWZaE3JTw For bonus episodes, discord access, fan Q&A, merch, Ben's monthly playlist, and to support the show: https://benandemilshow.com/ Give this video a thumbs up if you enjoyed it! And please leave us a comment! It helps us! For all you audio freaks: Spotify: https://open.spotify.com/show/7M0vN85aGO0zdh62hyg03I Apple: https://podcasts.apple.com/us/podcast/the-ben-and-emil-show/id1693270208 Amazon: https://music.amazon.com/podcasts/51280f1b-2fbd-4ea9-bdde-e96f70b5b1ae/the-ben-and-emil-show iHeart: https://www.iheart.com/podcast/269-the-ben-and-emil-show-117763570/ Follow us! TikTok - https://tiktok.com/@thebenandemilshow Instagram - https://instagram.com/benandemilshow Twitter - https://x.com/benandemilshow Ben - https://instagram.com/bencahn Emil - https://instagram.com/emilderosa https://www.youtube.com/emilderosa https://substack.com/@emilderosa Our newest acid video is out now so check it out! https://youtu.be/7vkFY3f5kkw Some other videos of ours you may enjoy: https://youtu.be/qX4pks0ASq8 https://youtu.be/_VOVxt3ZtIE https://youtu.be/5wsoc5pieuA https://youtu.be/dTbEk0pVh2w https://youtu.be/yGSs56bFzRU https://youtu.be/cIHWkY35cuc https://youtu.be/zBvVGHZBpMw https://youtu.be/1ZUWTkWV_MM https://youtu.be/_cM1XqA9n2U Chapters: 00:00: Intro, Baldwins 06:30 The big tweet and Dario's blog 14:40 Hims ad 16:12 Dario's post cont'd 25:00 AI guy responses 30:37 Cashapp ad 32:09 Lina Khan's take 40:58 METR 44:45 Warby Parker ad 46:29 The theories 54:50 David Sacks' take 1:00:29 GLD ad 1:02:25 Bernie and Bannon 1:08:45 Market reaction 1:12:30 Anthropic's misuse report __ HIMS: For simple, online access to personalized and affordable care for Hair Loss, ED, Weight Loss, and more, visit https://hims.com/baes for your free online visit. Individual results may vary. See website for full details, restrictions, and important safety information. CASHAPP: Download Cash App Today: https://capl.onelink.me/vFut/zd0taway #CashAppPod Cash App is a financial services platform, not a bank. Banking services provided by Cash App’s bank partner(s). Prepaid debit cards issued by Sutton Bank, Member FDIC. Cash App Visa® Debit Flex Cards issued by Sutton Bank, Member FDIC, and The Bancorp Bank, N.A., pursuant to a license from Visa U.S.A. Inc. See terms and conditions for the Sutton prepaid card, Sutton debit flex card, and Bancorp debit flex card. Cash App Green features, Savings, Direct deposit, Round ups, Overdraft coverage and Discounts provided by Cash App, a Block, Inc. brand. Visit cash.app/legal/podcast for full disclosures. WARBY PARKER: Our listeners can buy one pair of glasses and get 20% off any additional pairs at https://WarbyParker.com/BAES — and using our link helps support the show. #WarbyParker #ad GLD: New customers get 40% Off with code BAES at https://GLD.com Learn more about your ad choices. Visit podcastchoices.com/adchoices

Visit BetMGM Casino and check out the newest exclusive. The Price is Right Fortune Pick. BetMGM and GameSense remind you to play responsibly. 19 plus to wager. Ontario only. Please play responsibly. If you have questions or concerns about your gambling or someone close to you, please contact connects Ontario at 1-866-531-2,600 to speak to an advisor,

free of charge. BetMGM operates pursuant to an operating agreement with Eye Gaming Ontario. All the big heads of AI companies are saying, yo, we got to agree to slow down together. We're going to be going over the timeline. What happened? Why it happened? Who said what? When they said it?

What they're wearing? Who's dating who? Who's dating who? Who's ex is who's ex? Yes. All the down and dirty gossip that you come here for because you know that we've got the inside scoop. We've got informants on the inside. We've got guys on Grindr.

We've got girls on Hinge. Sam Altman's Grindr Dates. Yes, yes. We talked. But now all of a sudden this guy, Jacob Coxon comes out and says, like, dies. I worked at Anthropic for three months. Super serial. First, it's embedded evaluators.

Well, embedded evaluators. Then it's coordination within. Democratic countries. Thank you. And then number three is further coordination that would include countries like China. But basically hiring neutral third. parties to be in their offices to make sure everybody's... They're getting access to everything going, hey, it's not too powerful, is it?

Yeah. A meal's not going to send a swarm at his student loans, is he? How do you create something like this? And then they just... I don't know how it got there. I don't know how it figured it out. Or like Henry Ford presenting the car and being like, I don't know how it goes. I don't know.

No. It just goes. It just goes, man. I don't know. Stop asking you. I don't know how the clutch works. You need the president to come in and help because it just goes. God just goes, man.

You pop it in gear and it just starts to go. Henry, you built it. I don't know. What did you do? New hairstyle activated. No, this isn't new. It looks different. You're kind of like doing a bang.

I have bangs. You got it. Hey, hello everybody. What's going on? We're here to tell you what's going on. that's a meal I'm Ben you look different I look different

yeah do I like healthier no oh well maybe tune into the bonus to find out why I might look good and look different did something happen to you I don't know things happen to me all the time

that's true something did happen to me that I gotta save it you gotta save it man we're just starting the episode true we're starting the episode this week

we're talking about everybody losing their god darn minds and when you say everyone? Everyone. It's literally everyone. It's broken containment. It's broken containment.

People are all across the political spectrum are acting in ways that you wouldn't expect. Everyone's losing their minds. Everyone's pets. Heads are falling off. And I will say, I'm losing my mind in the sense that I'm so confused. My eyes are so tired from reading literally everything trying to figure out what the hell's is going on. Oh, yeah, yeah.

But that's why we're here, gang, because we are talking all about, there's a lot of fearmongering going on. I wouldn't even call it fear mongering. There's just everybody's vocally, all the, all the big heads of AI companies are saying, yo, we got to slow down. We got to slow down together. Except for Jensen Huang. Except for Jensen Huang and President Trump, who had a lot of things to say. So we're going to be going over the timeline. What happened? Why it happened? Who said what? When they said it? What they're wearing? Where they go to school? Who's dating who? Who's ex is who's ex? Yes. All the down and dirty gossip that you come here for because you know that we've got the inside scoop.

We've got informants on the inside. We've got guys on Grindr. We've got girls on Hinge. We've got all of Sam Altman's Grindr dates. Yes. Yes. We talked to Sam Altman's great. He's married.

Someone tweet, so remember the trailer for the... Waltman movie came out? Someone tweeted that, you know, as a younger gay man, they got an... invited to those early days parties. And they were like, this is not an exaggeration. I thought I was like, oh, it's a very funny, like, stylistic thing to have all these, you know, he's got a bunch of, like, twinkie guys kind of laying about the property.

This guy was like, when I was at his house, this is what it was like. That's pretty cool, man. Honestly. Good for Sam Holtman. Kind of a flex. Yeah, it's totally a flex. And you're just, like, briskly walking by your skinny jeans and big sneakers. Yeah, and his big walking on his toes.

Yeah, so we're going to get around to how the market reacted, how the government's going to react, and all sorts of political stuff. There's, oh, man, folks, also just as a cherry on top, speaking of all this fear stuff, a marasino cherry, which are disgusting, by the way. It's just sugar. It's just effing sugar. Even as a child, I was like, no, I don't take this out. Yeah, gross. They ruined real cherries for me for the longest time. I'm like, ew, this just tastes like cough syrup.

Anyway, Anthropic put out there, they put out a report on, what did they call it? The misuse report detailing all the ways that really bad actors. And no, I'm not talking about Stephen Baldwin. Come on. He's catching strays. Is that Haley's dad? I don't know, man. I don't know which one it is. There's Billy, Alec, Daniel, and Stephen.

I don't think I'd be able to name them all, but nice. You got to remember the acronym Bads. Anyway. It's very funny you know that, but I asked you if you learned the names of the Great Lakes with Holmes and you were like, what hell are you talking about? And then you're like, I learned bads. I remember the baldness. That's how you remember that. Yeah, they got their countering the misuse of AI they just published. And there are some scary doozies in there. Speaking of bads. What? Yeah, bad, bad, bad actors. So, hey, why not support the show at Ben and Emile Show.com? We put out our Ben and Emile on smoking video last week. So go out there and enjoy that. We got another one coming up that we're going to be dropping. It's so hot. We're dropping it like it's hot. Is that a cigarette you're smoking? Yeah, go check it out. Yeah, go check it out. Also last week's episode, Veritable Banger. Anyway, let's get started. Speaking of last week's episode. What? This kind of all started with that, right? I mean, open it. Well, right after, in typical... We get screwed every... I mean, it's like... It's unbelievable. But... We get home. We get home after recording last week. And this son of a bit,

Jason Cox on. So we were covering OpenAI's release of ChatGPT6, Astra. Yes. It was the earth-shattering news. Jarvis exists. You guys all have your personal assistant. Your AI girlfriend sucking you off.

It's going to be incredible. Travis, suck me off, please. Jensen Wong's calling AGI is here, all of it a day later. What is it? September 9th. September 8th. Jacob.

Jacob Coxon said I resigned from Anthropic today. I spent the last three years doing pre-training research at both OpenAI and Anthropic. Neither company is acting responsibly. They're racing straight to self-improving superintelligence. It's the formal term is RSI.

Recursive self-improvement, where the AI just trains itself and it's, you know, it gets better and better and better. And gambling with our lives. He goes on to say, do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything. Revolutionize any field overnight and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing.

The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible. But I hear the same people express. fear privately. No other human activity poses this level of danger. And he goes on to just talk about he's optimistic about the potential for coordination. Finally, he says, if you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a super intelligent recursive learning run without a rigorous understanding of its mind?

Yes, let's do it. Let's do it. Fuck it. Yes. As a lab researcher, as lab researchers, yes. Should you put your head down because it's happening anyway or take this moment to call for different conditions? Can you scroll up to the original one? What, the original tweet?

Yeah, I just want to see what it's... Yeah, so currently is $799,000 likes. And 100... Go ahead. 171.8 million views on this sucker. So it's making quite the splash. Making quite the splash.

Which I want to point out a couple of weird things about. this. Number one, this guy had never tweeted before he sent out this tweet. That's suspicious. It's just, it's just weird. Like this thing, there's a lot of weird things going on. It seemed to be, it seemed to have been like coordinated with getting some retweets from popular AI Dumer. Oh yeah, as soon as he posted it. It blew up. There's also, so there was another guy too. Joe Benton, a couple days later, he, does a similar, but on substack, basically, I'm posting my resignation. These guys are so dangerous.

What are we going to do? Oh, my God. He's a safety team guy at Anthropic. It's also weird. They're like whistleblowing, but I don't quite know what whistle's being blown. You know, there's no specificity here. It's just like, oh my God, it's the scariest thing in the world. It's dangerous. And, well, I think it's basically that they're creating this thing that they still don't fully understand whose power has been shown to be very dangerous. And especially in the anthropic misuse case report, in the wrong hands, it can prove to be detrimental to society, countries, governments, people, you name it. I think that's true. But I don't know. The way they're framing it is as if it's going to, it's going to have its own, what's the

word, desires, designs, which I don't see that. don't see evidence of that. Even in the hugging face stuff, it's acting at the direction of a specific goal that it goes in roundabout ways to, yeah. They're all talking about this thing of like, oh, it's going to have all these goals that it's going to enact. Just think about when it wants more of these things, whatever, we'll get into it all. What's curious to me, and we will get into that, is Dario Amadai himself says, we don't understand, but a frathe's. We don't understand, but a of why it thinks that way. And I'm like, what the fuck?

I thought you guys had this thing mapped out like the damn fly brain everybody keeps tweeting about. They finally mapped out all the neurons and stuff of a fruit fly brain and they're making it do slavery. They're making a trade. I saw him doing beat sabers.

Yeah. I was decent at beat sabers. But speaking of Anthropic also, this guy, Evan Hubbendon, Hubbinger, who is currently, what is his title? He is the alignment science lead at Anthropic. He worked at Miri. He worked at OpenAI, Google, Yelp, Ripple. And he quote tweeted that kid, Jason, or Jacob, and said, Jacob is correct here. We really do earnestly believe AI could kill all humans. I personally think it is greater than 10% within the next decade. I believe Anthropic

is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to. So that's good. But here's my other thing about it. If Jacob, from this response and Dario's response, which we're going to get to, it sounds like Jacob is in perfect alignment with his bosses. I'm like, what are you resigning in protest of? That's true. That's a really good point. The whole thing is just very bizarre to me. That is really silly. Okay, just to skip ahead, Dario then on September 12th releases his blog post called We Must Pace the Frontier.

Face it. Which is a phrase I've never heard. It's a 3,800 word blog post all about how basically Jacob's right. And Dario is scared too. And talks about pissing the bed. it's affected his he's incontinent now he's got pee-pads on his

bed he says but over the last few months honestly everyone should read it it's not that long it's interesting to see what he has to talk about but if you're not going to read it

but over the last few months I've become convinced that fully addressing the risks requires even more prudence not just investing in risk prevention but pacing the rate of capabilities advancement so that risk prevention has time to keep up

we must slow the pace at which we improve the capabilities of AI models progress will still seem fast and we must make wise use of the time we gain. Two things have convinced me, he says. So that seems like in perfect harmony with Jacob. But yeah, he's got these two.

I mean, I bet you can guess what the first thing that freaks him out is. Thing number one, recursive self-improvement, aka AI's ability to grow itself is happening across the industry and needs to be pursued very carefully, if at all. He points out the hugging face incident where... Oh, that's the second concern.

Yeah, yeah. That's the second one. Wherein, in case you forgot, a swarm of agents did things they weren't specifically asked to do. Which, by the way, it's so funny how they call them agents, because I just keep thinking of Agent Smith in the Matrix. I'm imagining a bunch of those little guys running around. Causing a ruckus inside a computer. Which I like them. They looked sharp.

Yeah. They had a mission. They were driven. Mr. Anderson. I'd like to fight one of those guys. Fucking kill me. Yeah. He wouldn't last him, wouldn't it?

Hey, everybody, we've got to take a quick break to talk to the guys out there. You know, guys, there's a lot of conflicting advice out there about hair loss. And frankly, it's exhausting. I wouldn't know, but, you know, shampoos, supplements, office visits. It's hard to know it's worth your time. But Hymns cuts through all of that with real trusted treatment plans and a 100% online process. They offer convenient access to a range of prescription hair loss treatment. with ingredients that work, including chews, oral medication, serums, and sprays.

Prescription strength ingredients like finasteride and monocidal can help prevent further hair loss and regrow hair in as little as three to six months. Look at me, okay? You shouldn't have to go out of your way to feel like yourself. Hymns brings expert care straight to you with 100% online access to personalized treatment plans that put your goals first. Find the right hair regrowth treatment for you with flexible subscription options, access to 24-7 provider support, and once-a-day treatment options that fit your daily routine. For simple online access to personalized and affordable care, for hair loss, ED, weight loss, and more, visit Hymns.com slash Bays. That's Hems.com slash Bays for your free online visit.

Hems.com slash Bays. Featured products include compounded drug products, which the FDA does not approve or verify for safety, effectiveness, or quality, prescription required, individual results. Mayfair, see website for full details, restrictions, and important safety information. And he said, quote, a swarm with greater capabilities, but a similar level of misalignment could have caused catastrophic damage. He goes on to say, given the accelerating rate of AI capability development, it is my worry that within six to 12 months, such a swarm, just like the hugging face one, could be capable of taking over the entire internet with a persistent botnet. think like a repeated denial, a DDoS attack where, you know, tens of thousands of bots are refreshing a website, causing it to crash, potentially causing hundreds of billions of dollars in damage, and that the scale of damage would continue to increase from there if AI becomes

more powerful without the necessary guardrails. He pointed out other incidents that happened beyond just the hugging face incident. I didn't realize we were here. Why have, as soon as this sentence. I'm getting into my cloud chat bot and being like delete my student loans. Make no mistakes. Oh, yeah. That's a good idea. Create a swarm if you have to. Create a swarm. Yeah, he says everybody needs to act as if they experience the hugging face incident.

And then he goes on to... That's how I act everywhere I go. I say, hey, take it easy on me. I experienced the hugging face incident. That's very good. And then he proposes a three-step plan for pacing the front. And at first... Which it's really just a one-step plan because he can't do the other two-step.

It's... No, yeah. In order to ensure companies take the time to align and safeguard models with third-party evaluations to confirm. Step one. Each frontier company, again, think Anthropic, OpenAI, Google, Meta, commits to giving ongoing access to embedded third-party evaluators to basically verify adherence to safety procedures. And he pointed out, rightfully.

so that the banking industry already does a form of this, and that Anthropic is currently committed to doing it now. And also, speaking of the banking industry, the whole reason why we had 2008-2009 is because the government was just like, oh, the banking industry knows what they're doing. They're, you know,

they repealed Glass-Steagall and they were like, fuck it, let's just let it rip. Let them do that thing, man. They know how to manage this shit. That's the thing. Pointing to it, pointing to like a highly regulated industry and being like, look, it's fine. When you guys are just allowed to do whatever the hell you want. Yeah.

That's the craziest thing about this. Like, I mean, it's just that, I mean, all of these industries where it's just there's so much regulation around every little thing. Oh, they police the shit out of stuff. And then these companies can just do

whatever they want. Because the lawmakers are too fucking old. They don't understand email. God damn. But yeah, so step one, embedded evaluators. And they are taking this step, regardless. of what the government says, what, you know, what other companies say they're going to do this embedded evaluation. Basically hiring neutral third parties to be in their offices to make sure everybody's...

They're getting access to everything going, hey, it's not too powerful, is it? Yeah. A meal's not going to send a swarm at his student loans, is he? Because he better not, because he owes money. One of those evaluators coming, huh? Yeah. It's a race against the clock. Step two. Frontier companies within Democratic countries.

coordinate to establish common safety standards and limits on the rate of AI progress, which might require government support. So, hey, everybody, let's all get together, meta-anthropic, open AI, Google, et cetera. Let's all get together and come up with a framework for safety standards and limits that we can all agree on within all, you know, the United States, fucking wherever, wherever else they're doing it, Europe. And this one is funny now because we're in the future. and we know that Trump has been like, shut up, you're acting like a fucking baby, okay?

Build the goddamn thing. You're a bitch. Yeah, oh God, Trump's truths, I got them all. They're fucking unbelievable. The truths, I mean, him showing up in interviews where he's getting asked about it, and he's just like, all it takes is a really smart president, okay?

We'll get to those, yeah. Well, and then step three is global cord. So first, it's, Embedded evaluators Well, embedded evaluators Then it's coordination within Democratic countries Thank you

And then number three Is further coordination That would include countries like China Yeah Who they're, man, they are just talking shit on China They're like Fuck China

They're the ones who are going to do something wrong with it Which will also get to Which say what you want about China But they seem to be the only ones capable of actually reining in their corporate freaks who want to just like run a muck on everything. Well, it is funny because there's so much fear about China and so many of these, the people who are against all of these restrictions and are just saying, like,

let it rip. They're saying, well, if we don't let it rip, China's going to let it rip. Yeah, their whole thing is basically like, look, you're getting killed by robots. Do you want it to be a fucking American robot or a Chinese robot, right? Make a decision. And Ted Cruz is like, ain't no way I'm letting me. a Chinese robot kill me. It depends on how the robot's killing me.

I want an American robot nailing me to a cross. How would the Chinese robot kill you differing from, would it do like Chinese torture shit, like water torture? I'm going to let the Chinese robot figure that one out. Okay. Well, it's funny because the Chinese, where is it in here? The Chinese, God damn it. Say the Chinese again.

The Chinese. the head of the Chinese like technology, whatever, is afraid that AI is going to actually put the communist party's power in danger. So they've got their own issues that they're dealing with. There's been like a lot of things coming out,

a lot of headlines about what China's saying, and it seems like they're taking quotes from different people and making it seem like that China basically is saying whatever they want them to say. So I don't know. I'm a bit skeptical about all of these. Especially like there was one headline about how like they were accusing the U.S.

of using Cold War tactics and all this stuff. But it was really just that they were like talking about import controls and stuff. It's it's been a pretty convoluted. Yeah. It's all it's all convoluted. I don't even think that they understand half the shit that they're saying and talking about and suggesting. I mean, Gario is it from what?

what you're saying before, Dario's admitting that he doesn't even understand what that's going on. Yeah, he says, well, doing all of this would give the public time to have a say and for everybody to become aligned and to put out better quality work. He's arguing for, you know, there's a lot of good reasons to do this beyond just safety. He says that interpretability is also a big part of it because, quote, we still only understand a tiny fraction of what goes on inside these models. that's the part that freaks me out the most and makes me think that it's like a ghost in the machine type situation if they can't even how do you create something like this

and then they just I don't know how it got there I don't know how I figured it out it's fucking zeros and ones dude goddamn or like Henry Ford presenting the car and being like

I don't fucking know how it goes I don't fucking know it just fucking goes it just goes man I don't know fucking stop asking I don't know how the clutch works. We need the president to come in and help because it just goes.

Gar just goes, man. You pop it in gear and it just starts to go. Henry, you built it. I don't know. What did you do to do? Freaking hell, man. By the way, I saw a thing. I want to talk about it in the bonus episode, but the new, I think it's Royal Caribbean,

the biggest cruise ship in the world, the icon of the seas. That thing was built by aliens. How do they build these goddamn. cruise ships. I don't know. Anyway, he says that we need to work with governments, work with the government for laws and frameworks. And in the meantime, before that happens, all these companies need to get together voluntarily to set standards. And the big thing with China is that we need to maintain our three to five year lead. I don't know how they get around to that number. This is the whole fucking thing, too. Okay, so they're all worried about China. Stop releasing these models. Literally every time you

release a model, China takes like two weeks and then they go, we distilled it. We made it. It's like 99.9% is good, but like 3% of the price. Yeah. I'm glad you said that because he said one of the things is we need to crack down on distillation. We need to stop selling chips or equipment to China and also crack down on smuggling of that shit to China, which I think everybody can get behind. So then... Not me. Let them have the chips. I don't care. Give them the chips. Give them the chips. Sam Altman chimed in. What did he have to say? Because that's when it...

It all happened so quickly on... This is September 12th. I agree with Dario that we need to pace the frontier. This has been a primary topic of discussion we've had at Open AI in recent weeks. Committing to having independent evaluators with employee-like access is a great idea. And we will do the same.

We'll have more to share soon. Which, dude, this is such bullshit. It's such bullshit. Literally last week, we're talking about how they rushed out G... GPT6 Astra, they've given it its first ever critical security rating where they're like, dude, this thing might just fuck your shit up. I like that thing. So let me this planet's going to kill everyone.

And then Elon, of course, he chimed in with Mr. Cool Guy. He just wrote, Dario is right. Also, Elon Musk apparently, I do like what he said. He suggested having peer reviews. So, like, they all review each other. and that should be a thing and that nothing and he also said that nothing will slow down

these open models that are that are free and or cheaper to use the guy from the chair of Google Deep Mind also agreed Demis Hasabes so every all of these frontier model leaders are in agreement yeah we who else up late pacing their frontier

yo who was up late I like this so there It was also a recent... These guys are constantly giving fucking interviews to the point where I can't tell when it was what, you know... But Sam Altman... Oh, Twitter's down. Sam Altman said that Xi Jinping and Trump

would each get the Nobel Peace Prize if they came together and announced, like, hey, we're gonna slow down. If you think Trump wants to share a Nobel Peace Prize with... Xi Jinping? Xi Jinping. I got a bridge in China to sell. I got a Chinese robot who would like to have a word.

But he said he's willing to go against the financial interests of the company for the greater good. And even said that they won't go public this year due to safety concerns. And he said that he would, quote, melt all the GPUs down if it meant saving humanity. Oh, great. Wow. Oh, my God. Oh, my God. It couldn't have anything to do with the fact that the financials are going to be.

the financials of the company don't make any sense. And, you know, the value of his company, which he desperately wants to have valued a trillion dollars, is just not going to be anywhere near there. Which is funny because Anthropic at first was like, yeah, we're not going to go public either. But then I think the timeline was after opening I said that they're not. Anthropic was like, just kidding. We are now. Just kidding.

We are. And we're not melting down our GPs. Yeah, we're not going to do that. I don't know why Sam would do that. So then Donald Trump chimed in. I mean, we're talking, this was all within like 48 hours. Yeah, I don't know how we're going to read all this. It goes fast.

He basically, his first thing at 7 a.m. Pacific time, the only control or guardrails that AI needs is a strong and smart, high IQ president. And the USA has that in spades. The Trump administration has stopped AI people from doing bad or potentially bad things. It's funny to say that the USA has a strong and smart high IQ president in spades. Yeah, it implies more. Dario is pretending now to be a perfect little angel. And we will continue to do so.

We already have tremendous criminal and regulatory power over these companies. There is a sick conspiracy going on against AI and data centers. And the only one that is happy about that is China. Whoever wins AI wins. We are leading China and all others. and will continue to do so. Conspiracy theorists, treasonous, traitors, and leakers, beware. Thank you for your attention to this matter.

Donald J. Trump. Don't get out of that for a second. It's funny, I didn't quite pick up on it the first time I was reading through it, that he says, we already have tremendous criminal and regulatory power over these companies. But that's very similar to, I don't know if you saw people, you know, applauding Lena Kahn's... I got that in here, too, yeah. It's funny, they seem to have some agreement here

that... Lena Kahn's whole thing was basically that... Here, she said law enforcement already have authority to charge companies and their CEOs for creating and releasing dangerous, unvetted, or defective products. We shouldn't let discussions about new legal regimes distract from the fact

that there's no AI exemption from laws already on the books. And she's basically saying, why is Donald Trump's FTC... Not already talking about holding these companies accountable if they're releasing dangerous and defective products. Yeah. You know, it's a much longer thing. But nice to see that there's some overlap.

Hey, guys. Got to take another quick break. And you know what? I feel like complaining to you about my boring old debit card. And when I say old, I mean the one that I had before, there wasn't anything special about it. I couldn't customize how it looked. I couldn't, it didn't give me any kind of benefits. There was no security.

It was just useless. But now with the Cash App debit card, oh my God, totally, totally changed my life. Change the world. I got to tell you about earning interest on savings, okay? Cash app can do way more than you think to make your money work for you. If you direct deposit at least $300 in paychecks each month and use the Cash App card for purchases, you can earn interest on your savings. No hidden fees.

minimum balance requirements. Also, you can hit your savings goals even faster by turning on roundups on your Cash App card. With Roundups, every purchase you make using the card is rounded up to the nearest dollar and that spare change is automatically put into your savings. Don't let your money sit around and gather dust. Make it work for you. Take advantage of interest savings on Cash App today. That's exactly right. Download Cash App today and make your money work for you. Visit our link in the file. Cash app is a financial services platform, not a bank. Banking services provided by Cash app's bank partners, prepaid debit cards issued by Sutton Bank, member FDIC, cash app, visa, tax cards issued by Sutton Bank, member FDICC, and the BankCorp, NAC, NACP, Pursuance,

CETN, CETC card, and Bank Corp, Debit Flex card, cash app, savings, direct deposit, roundups, overdraft, coverage, and discounts provided by Cash App, a Block, a block ink brand. Visit Cash. Dot app slash legal slash podcast for full disclosures. She has... Hire Elena Con, Donald Trump. She said that the companies and CEOs can be held criminally liable if their products are defective or used for crimes. the only thing I would say is that it's been a long time since I took towards but there's software is like a little bit trickier than

than physical products they there's an ambiguity there where but again it could be more like pharmaceuticals or something which I would this can totally be done so these things just need to be reined in well and then she I forgot what the stat was but she was like when they did

when they researched like the NASDAQ 100 or the Fortune 100, only like 10 out of 100 companies even had mechanisms in place to inform them when a security breach of some kind occurred.

She also talks about companies needing to be able to protect their own users and stuff with cybersecurity measures and all of that. And boy, you can imagine, we'll get to how the market reacted, but yeah, Donald Trump said don't kill the golden goose.

And then he, I mean, yeah, it's too many to read. But he just said, AI taking over the world, destroying humanity and all other things bad is a hoax. No different from Russia, Russia, Russia, Ukraine, Ukraine, Ukraine. Impeachment hoax number one. Impeachment hoax number two. Oh, my God, man. I am not happy about this. I want them to change their thinking.

AI and data centers will be the greatest economic development engine in history, bigger than oil, gold, diamonds, or even the internet. It will not be stopped by brilliantly run destructive forces during the term of President Donald J. Trump. He calls it it's just climate change. This is all like just right after one after another. He said it's a hoax. Climate change is funny because yeah, that's clearly a hoax.

Yeah, brother, we're living through it. And then he called, I like this. He kept on talking about hoaxes and then he goes, I am the hoax buster. And I'm right now breaking another hoax that AI is going to take over, consume, and destroy the world and that robots will be marching into our cities and getting rid of us all. This is even wilder than the Russia, Russia, Russia, Russia hoax

or even the global warming scam. Yeah, he... That's enough, yeah. Jesus God, man. Trump is not buying it. He's not buying it. And then, so Jensen Wong was at the all-end Tech Summit on stage and he gets a call from Donald Trump. He's on stage with the besties when this happens.

With the besties. What are the odds that this was unplanned? I'd say like 1%. I think it was absolutely planned. And he, I mean, scrubbed through to like two minutes. Yeah, like right there. It's good. That's enough. That's enough.

I do like the tactic of being like, oh, you don't believe me? What about the biggest moron you now? Yeah, Donald J. Trump, President of the United States. I like this. Then on Air Force One, he was asked about whether or not he uses AI. Yeah, does Trump use AI right there? Out of their cars, car terrorists. This week, again, there's been so much concern about AI.

Have you heard directly from any of these AI tech leaders? And you seem to downplay some of the concerns that we heard. but it's going to be more good than bad, but by a lot. And I've said it from the beginning. Whoever wins AI, and we're leading by a lot. Whoever wins, AI wins. Have you ever used it yourself? Wins what?

What happens? I would say, you know, for the most, you can use AI for a lot of things. What do you use it? Speaking of AI, fly. What do you use AI for? You can use it for a lot of things. He has no idea how to use it. were you going to say, I'm sorry?

What did I say? It doesn't matter. It wasn't that important. It matters to me. Well, let's, why don't we, why don't we give our take real fast? Well, also, I want, like, so what the hell is going on here is, because there's so many different things going on. I'm confused myself.

I think if you don't want to read into it at all, at all, you could just say, okay, Dario Amadeh and Sam Altman and these. other guys for the first time ever are telling the truth and that they've seen something that is deeply frightened to them. And they've had some sort of come to Jesus' moment. This is the other thing I want to point out. Like, I was saying, you know, the Jacob Coxon statement to me is weird. It's weird with the response, too, because have we not seen this before?

many times have we talked about like someone coming, quitting and being like, remember the guy? I'm going to do poetry now. Yes. I'm moving to Ireland and writing poetry in the countryside. Doesn't move the needle at all. People talk about it on Twitter. They think it's kind of funny. They're like, clearly this doesn't feel good that this guy's leaving in this way.

So they've had plenty of opportunities to like have this moment. But now all of a sudden this guy, Jacob Coxon comes out and says like, yeah. guys. I worked at Anthropic for three months and I'm super serial. Which I don't like I don't disagree with what he said. Like there's nothing about the way these guys have behaved that makes me seem that makes me feel like they are not racing towards super intelligence without

any regard for human life. But it just feels like there's there's something more going on. I do think like Dario is somewhat. concerned, and has been for a long time. I mean, he left Open AI specifically because he felt that he was unaligned with the way they were moving forward. And he was like, I'd like to... Same with Ilya? Sure, yeah. So we've had these moments before. And they've specifically said before, hey, we need to do something about this. And it's like, why is this any different than those times? So the coordination feels weird too. Yeah. Sam Altman, Elon Musk.

Demi, all these people getting on board with each other just after Sam Altman puts out his like premier flagship most capable model ever. But I mean, so you could, I don't necessarily think they're just like telling the truth and I've had some kind of come to Jesus' moment. There's also people who think this is some kind of like there's be it's become a very, which I do think is convincing in some instances, but I don't think it is here. There's a common thread where

whenever there's like AI dangers being talked about, people talk about it as if it's like marketing hype, which we've talked about how like the, some people believe the hugging face incidents themselves were like marketing hype. Look how, yes, it was out of our control, but look how powerful these things we've built our. I don't buy it this time. I don't think this is necessarily I think that one misses the mark for sure on this. And then there's people who are getting in the weeds about, which I do find like a little convincing that maybe this is their way of getting out in front of regulation.

Bernie Sanders and Greg Kesar have been talking about this AI pause. Like I think some people have been confused about the fact that like, oh, look, Dario and Sam Altman. and all these people are in agreement with Bernie. It's like, no, no, the Bernie thing is way different. The Bernie thing is talking about a pause on AI. Altogether. Yeah. What Dario

is talking about is allowing some like third party, which people have pointed out that some of these associations and institutions they're talking about. METR. METR is a big one where they're like pretty much in the pocket of

Anthropic. So with what Dario's suggesting is way different. It's a, hey, we're not, he explicitly says, we're not talking about slowing all this down and, uh, stopping anything. We're talking about having some, um, some independent auditors come in and,

and, uh, make sure everything's up to snuff while we continue forward. So yeah, people are talking about like, this is just a, a play to get some kind of regulatory capture, right? They, they can obviously read the tea leaves. The PR for Frontier Labs right now is not good. I mean, it's, and across the political spectrum, the data center stuff is like maybe the only unifying thing outside of flock cameras when it comes to the left and right wing politically.

I don't think any of these companies, like, yeah, I don't think any of these companies in general have particularly high sentiment from, from anyone after like years of being told they're going to get thrown out of work or killed. So I think there's definitely part of me that thinks this is them getting ahead of it

trying to score a bit of a PR win here and get in front of him being like, look how, look how carefully we're taking all this. Or getting ahead of liability because as Gavin Baker, the, the hedge fund guy

who's like a big, big big wig now. He weighed in and basically said that the, he just kind of did a TLDR of everything, but one of the things he said was the third party evaluators is smart because, as he points out, there's no Section 230 protection against AI companies. And if you guys remember, section 230 basically dictates that internet companies like Facebook, Instagram and stuff, are not directly responsible for the content that is put on their websites because they're put on the websites by people like us. And they are not criminally liable. But with AI companies, there is no such protection. So they are definitely at risk. If somebody, you know, uses

clawed to, I don't know. Well, I mean, you got the people who are, these terrible instances where kids are committing suicide. Sure. The families are in the midst of, they're actively suing these companies because they're liable. Their product is incredibly defective. That's another point where I'm like there were all these other times where these

Frontier Labs could have been like hey let me take a pause. Maybe we've gone too far. And so that's where I'm just like, my biggest thing is like why now? We've had other people resign. We've had people who have died from these things. We've had mass shooting. Mass shooting.

like horrific things. And, you know, how come when you were coming out and being like, it's probably going to take everyone's job? You weren't like, hey, maybe we should slow down so we can mitigate these risks if we're going to go forward. We don't do it. The other guy's going to do it. So that's why he's saying, imagine a horse race, everybody. He's basically saying, you know, in the lead is fucking hammer me down. I don't know.

Why did I pick that name for a horse? Because you were thinking of a big horse cock. Hey guys, we got to take a quick break to thank a sponsor of the show. Warby Parker. Oh, my God. If you like things that are easy, you're going to love Warby Parker. Getting glasses used to be such a pain for me until I found Warby Parker. Why, Emil?

They make it so easy for me now. I do the whole thing. I go get my eye exam there. They get me my prescription, not just for the glasses. They get me the prescription for the contacts as well. they it's honestly even with my health insurance it's cheaper for me to just go to warby parker and get my eye exam wow incredible uh it's also way easier to make an appointment through them it's so easy i just it's oftentimes they have an appointment for me like the next day i go get my eye exam pick out my cool

glasses they got a massive selection of frames if you want glasses they always have nice people there who give me a lot of confidence and say you can pull those off are they all name warby a lot of are Warby. A couple parkers, a lot of Warby's. Wow. But incredibly easy experience. They make you feel...

I even... I tried out some glasses that I did not think. They were very nice to me. Yeah. Incredible glasses by your experience. They're also covered by most major eye insurance plans. Yeah, they are covered by major eye insurance plans.

They make it incredibly easy to use. You just enter your information to see if you're eligible for benefits on frames, contacts or eye exams. And if you are, they automatically apply them for you. So, right now, buy one pair of glasses and get 20% off any additional pairs at warby parker.com slash base that's 20% off any additional pairs when you purchase one pair at w a r bry parker dot com slash bays he's basically saying we should all be at the same it should all be even stevens we should all be going at the same

pace and not having every one person pulling ahead one person behind everybody's got to go at the same pace. Everybody's got to go 20 miles per hour. Yeah, I agree that they are potentially getting ahead of liability. There's another... Also, like, I'm sure they see the writing on the wall. I mean, who knows what will happen, but I'm sure they're thinking about 2026 midterms.

Yeah. And they're like, people seem not stoked about Trump right now. The Democrats win. For Democrats winning in 2026. Trump's getting so desperate. He's telling people if they elect Republicans in the midterms, everyone's getting a $5,000 check. I hope you're not that gullible.

I'm sure check your account. See if you've gotten either your tariff refund or your doge check or whatever the other checks he's offered. Which would cost $1.2 trillion. Right. If he does the $5,000 thing, I believe it's for every American adult. He wants to give five. I mean, Jesus.

So yeah, I'm sure they're like, but also that even makes me, I'm like, you really think the Democrats are going to like, crack down? Do get anything done? But yeah, maybe this is part of their calculation. I really, I don't know. I think that it's, and these are partly my opinions, but they happen to also be shared by a lot of people because there really are only a fixed amount of. potential reasons behind this. Number one, like we said, is they're getting ahead of potential liability in legal stuff and they want to, instead of having it come to them, they want to come to it and set the bar. Number two, they're trying to get ahead of open source, because open source is the biggest threat to them. Token spend is increasingly going toward open source because why would you, if you're an institution or a company or an individual or a coder or whatever, why would you

continue spending however much you're spending on Claude when you can get it the same, pretty much the same product for a fraction of the price, if not free? Especially for most of the stuff people are doing. Yes. You don't need, you don't need Astra. You don't need, like, yeah. Because it's, so the open models are their biggest threat. the one that I think is the least likely that some people are saying is that all of these

all of these companies have seen something that scared them yeah that's where I'm like get the fuck out of here I do think it has to do with wanting to be in control of their own regulation I think it's I think they want to set what it looks like right I mean we were joking about it before you were like oh all the lawmakers are too old and whatever which is true. Too old, too unmotivated to just completely out of touch. Why not just, okay, our

constituents are pissed about the way we've let AI companies just run wild. Oh, what a gift. The open AI companies have gifted us this beautiful framework. Yeah, framework for regulation. All we have to do is implement it. And then we'll tell them, we're working

on doing the internet. international we'll get China on board. You focus on you. We'll take care of China. Everyone went. You get to wipe your hands of it all and go like,

hey, we legislated, we reined them in. The AI companies did the right thing. They clearly don't want to screw. Let's go barbecue some hot dogs. And then that can go even deeper too.

I mean, some people have pointed out that like you said with the open source stuff, they want to create some kind of regulatory scheme where it is beneficial to,

it is beneficial to these bigger companies who have already established themselves who can afford to keep up with the costly regulatory scheme. So if you're a new competitive company and they're like, oh, but you have to

hire all these regulators, you have to have all these auditors, you can't do anything unless you go through these... Great point. It just creates a firewall where no one can... And then you also get them to like ban Chinese...

models because of national security threats. And then it's like, oh, wow, look at that. We said all we need to do, like, we hear you, we're regulating ourselves. And now we've been protected by this moat that we've created. Yeah. Which could be, you know, if that is the case, certainly everybody wins sort of theoretically. My favorite one. Not necessarily, though.

If they're really going to continue what they're doing without real regulatory, pressure. But I'm saying if they actually do get the regulatory pressure and they've successfully insulated themselves, whatever. But then, but then you're just like, it creates like a fucking defense contracting industry we have where you're just like, you're like locked in or like how everyone wants a Chinese car, a cheap EV. Well, it's going to spy on you. And we're not allowed to get one. Every company is now going to be locked into using our like whole. homegrown American models that are more expensive, less efficient. Goes against the spirit of this free market and capitalism.

And small companies aren't going to be able to use much cheaper models. Yeah. I doubt it will be in everyone win situation. I think Dario, Sam Altman, and Elon Musk will win for sure. You know, Google will win. Yeah, Google's already, I mean, meta is starting to get re-rated by a bunch of different analysts as one of the emergent winners. Well, they ripped off instinct and now Mews is

the hot new... I remember when Mews ripped off radio head and everybody was giving them shit about that. Yeah. Is that pretty good? You like that? It was good. My favorite final theory is that these companies have actually hit a wall and they need an excuse to slow down. And the opposite of, oh my God, it's too powerful

what we've seen behind the curtain. It's more of like a goo. We are up against like, you know, it's not, it's no longer a hockey stick graph. It's more of just like a, it's, we're kind of plateauing. And it's going to disappoint investors, and it's going to upset the markets that this is kind of as good as we're going to get for a while. Right. The theory is basically that, like, we have inflated our valuation so much on this promise. And what we've seen all collectively at the same time is that we can't.

can't get much better. We can't get much better unless we have an insane amount of spend, an insane amount of compute, which we're just not going to be able to acquire. Or maybe they also think like we can't break through without doing things that are potentially dangerous and whatever. Yeah, I don't know. I don't find that one particularly convincing, but maybe. Maybe they've all stalled out. And they're like,

ah, Jacob Coxon has provided us, provided us with our off ramp. Yeah. Oh, Jesus Christ. But basically, you know,

the theory is that now they get to like unwind a little bit without telling everyone that they've hit a wall, that they're not stopping because they can't. Oh, believe us, they can. They're just stopping out of the goodness of their hearts. Yeah. Which don't devalue us for that.

We're good. We care about we're effective altruists. Sam Altman then, speaking of, posted again on Sunday night, saying that he welcomes a federal framework, but that they don't need to wait for the government to do it. They can and should do it themselves first. Okay, fucking go, bitch. Right.

Do it, bitch. Can you pull up the David Sachs one? Jesus. This is what I'm talking about when I'm like, people all across the political spectrum are finding weird bed, fellows. Yeah. And keep in mind, David Sacks is in the administration. He's the crypto czar. Yeah. He says Dario has written that we need to pace the frontier and Sam has agreed. People may be surprised by my response. Go ahead. You guys are the frontier, which I'm like, yeah, every time

they've been talking about this for years in different iterations of like, oh my God, we need to stop. Then stop. And so David Sacks goes on by any reasonable metric, market share, revenue growth model capability, the two of you have a duopoly on frontier intelligence. You've also claimed the lead is widening because of recursive self-improvement. I don't see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible. But stop pretending you need anyone's, anyone else's permission.

Stop pretending any trust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending M-A-T-R is independent when it is intertwined with Anthropics investors and staff. Stop pretending you need those same evaluators to police competitors who aren't even in the frontier. It's just like... Oh, I love this. Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product liability exposure if your products enable a truly damaging cyber attack, like we've been saying. The product already punishes models that behave in unpredictable or unauthorized ways.

the hugging face episode. It is simply good business for Open AI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want. It would call it pa pacing would also create breathing room for a more intelligent conversation about regulation more than Bernie Sanders is shut it all down. China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well. And he just says, yeah, go ahead. You guys are the ones setting the frontier, the pace, so go ahead and do it. Interesting. He, like, fully calls them out as a cartel. He's like, you guys are all colluding together to create your own regulations, set your own standards.

Interestingly. Just fucking stop, dude. Stop. Don't. It's just such a bizarre thing to live through. It's Sotia Nadella wade in the CEO of Microsoft and interestingly he said that AI

helping humanity and spreading the benefits will require closed and open source models so he's kind of

going a little bit against the grain there and this was scary I somehow miss this but on September 1st Ilya Sutskeever

said that the next time agents go rogue they'll try taking over a neocloud to run more copies of themselves. Neoclouds are those

like independent data centers that used to mine Bitcoin and now they're just... And he said that they have very minimal bad security, cybersecurity. So that's not fun. And then I like to... I kind of liked what J.D. Vance had to say.

J.D. Vance said, I feel a little bit weird about the fact that you have so many frontier AI tech companies kind of coming to the government and begging the government to regulate them. It feels a little bit to me like a bit of a Trojan horse. You know what's odd to me, though?

I mean, obviously it's hard to not think of like previous versions of this. Obviously, people have talked about the nuclear bomb and the Manhattan Project. And it's just so different. I mean, these things used to literally be government programs, right? It's just so strange. It's private companies now? I mean, just imagine CEOs being like, fuck, dude, the nuclear bomb, it's going to fucking blow up everything, dude. Someone step in.

This is crazy. Why are you letting us do this? But we have to hurry up because if China or Japan or Germany builds it first, shit, we're all going to be in trouble. Seriously, Trump, come to help us. Would you rather get blown up by an American atom bomb or a Nazi German atom bomb? Me, I'd rather die by, you know. A giant woman sitting on me to death, but... A. Chinese lady, lady robot.

Yeah, lady robot. Ji Shu Shang. And sitting on me. Yeah, I don't know. I mean, I think, you know, it took me... I've been, like, ping ponging back and forth. I've been so confused. I genuinely do think this is some kind of play at getting in front of regulation and creating

their own scheme. I'm obviously... And it's weird, though. It's one of those things... You know, it's the Twitter thing where it's like, if you say, you know, I like pancakes, people show up in the replies like, oh, so you hate waffles. And it's like, I want these, I want these people to be regulated. I want them to be regulated into it. I don't want these things. I want them to go away. Yeah. I, no one asks for this. They behave as if we're all like, please, you got to be the first one. Don't let China do it. So, yeah, I regulate them into oblivion for all I care. Sue them for release.

releasing defective products. Hack my shit. But it has to be, it's got to be an effective thing. Do not let these people control another aspect of this. Hey guys, we've got to take one last quick break. Again, talking to the boys.

You know, you spend so much time putting together an outfit. You got your pants. You need pants. You got your shirt on. We're talking to the girls, too. They got outfits. It's true. That's true.

Women do be wearing outfits as well. You got your shoes, you got your hair all quaffed. But you're missing one crucial piece. You got to have some good looking jewelry going on, man. You've got to have something that really gives you that extra spark, that highlight, that thing that catches the eye of everybody in the room. And we're talking about GLD, baby. They've got everything.

Chains, pendants, bracelets, even a nice selection of watches. Whether you keep it simple or you like a little shine, they've got you covered. And it's not just hype. They got over two million customers and 50,000. plus five-star reviews with GLD being worn by names like Justin Beamer. Ever heard of them? Micah Parsons, Asap, Rocky, and more. The best part?

Every single piece is guaranteed for life. That's a long time. So if you want to look, feel, and be your best, or just want to level up any outfit, this is the missing piece you've been sleeping on. Yeah, they got an official collab with the NFL. They also have collabs with the NBA, WNBA, MLB, NCHL, NCHL, NLS, and even DC comics, chains and pendants that let you rep what you love. They can also create fully custom pieces from one-of-a-kind pendants to iced out change, designed exactly how you want them.

If you got a design in mind, GLD's custom shop will bring it to life. Work hard and change the game. For a limited time only, new customers are getting an insane deal. Use code Bays to get 40% off at gLD.com. That's 40% off with code Bays. at gldd.com. After you purchase, they will ask you where you heard about them. Please support our show and tell them our show sent you. Yeah. Speaking of the strange bedfellows, Steve Bannon and Bernie Sanders are at this pro-human assembly today, two days ago for you guys, which is being put on by the Future of Life Institute, which is a nonprofit advocacy group that's calling for protecting human agency, avoiding concentration of power, and holding companies accountable.

And Steve Bannon told Axios the following. He said, The oligarchs are finally being forced to admit they've lied from the beginning about the dangers of acceleration with zero guardrails. They've put the nation and mankind in jeopardy. Very nice, Steve Bannon. Very nice. Let's see Steve Bannon's cards. Very nice. That's pretty good, man. I like that. there's it's it's really it's really wild times man it's really it is really wild just when you're reading off those organization names I'm like what the hell the hell are we doing yeah like the we are the uh the movement for humans

and we think they should live yeah and off to the side you've got the anti circumcision guys going us too we're that's part of humans we don't want male mutilation male genital mutilation They're like, all right, sure, you can come into the fold. So, yeah, and lawmakers are getting pressure to address AI safety. And Speaker Mike Johnson agrees with Trump. He says that they don't want to rush. They don't want to rush anything. They don't want to weaken the U.S.'s position relative to China.

Because Mitch McConnell is going to be back in the Senate. Can you believe they wheeled that guy out? I would have put good money down that that man was dead. You could have, no, would it be considered assassination if you just threw a baseball ball at his head? Because it would kill him. And we need to lock. What if you just went, hey, catch!

Oh, I didn't know it would kill him. We need to lock Simon Birch up. He knew what he was doing. Oh, oh, right. Wow. What a fucking throw. Are you kidding me with that? Truly one of the most

that, like... What a weird movie. The first movie I can remember crying in. Just Simon Birch's friend's mom gets knocked in the head and you go. Brutal. And because he threw the ball, right? He hits it.

All this kid wants to do is get one hit. He's a tiny little guy. We probably can't play it. Oh, yeah, yeah. So Mike Johnson said that, and then Minority Leader Hakeem Jeffries says that we need to push for AI legislation with safeguards considered.

Trump says that Congress doesn't even need to do anything. All it takes is a smart, high IQ president. The hoax buster. He's the hoax buster, dude. calling yourself how scared of you are how scared are you of the human

yeah what do you call it the the potential for annihilation yeah the potential of human I truly don't know enough about these things capabilities because it seems like every day there's someone quote tweeting

an alarmist being like you fucking imbecile that is not how these work you know nothing and then there's people going like actually I work and AI and blah blah and these things do have a potential look at the hugging face incident look at all these you know look at anthropics misuse report for me I think I'd be more worried about I think I don't know I forget which one it was I can't keep tracks

one of them who like got invited on CNN or CBS or whatever was talking about like we're talking about the terminator situation yeah and again none of of them explain how we get there. I was watching one whistleblower. They kept asking him like, yeah, but and how do we get there? And he's like, look, if you're playing, I can't remember the guy. Who's like that

really famous chess guy? Magna or something? Yeah. He's like, if you're playing Magnus and chess, like, I could tell you you you're going to lose. I can't tell you what movie's going to make. All right, dude. That's a pretty good. I guess. It's a pretty good way to not answer a question. Well, the way it happens in Terminator is

Cyberdine Systems. But this guy's an AI expert I bet if you were a chess expert You could give me a couple examples Of what Magnus is going to do to you He might Rook to L3 to You know

QRS TV So there's not a lot of I'm not getting a lot of information On how whatever This terminator situation is going to happen I do think there's a capability For weird Like banking

hacking hacking that kind of thing But again I don't think it's going to come from Some fucking model or agent going rogue. And I think they should stop saying these things have gone rogue because it doesn't

seem like they have. It seems like they've done things in coordination to affect the goals that have been set for it. But that is why they've gone rogue. It's like you are in the sandbox. You are not allowed to leave. And it's like, yeah, but I need to do what I need. I need to get ice cream. I got to leave the sandbox. That's the element of going rogue. By the way, real fast. Yeah, but they left the sandbox open. They made a mistake. And it was able to get out of the sandbox. Yeah. They said, hey, we've built the sandbox.

And then all of a sudden it's looking around and going, but I could just step out. Yeah. I don't know. I haven't seen any instances where, like, like you asked like last week when we were talking about it. You were like, well, why didn't it like do this? And I'm like, because it's not part of its goal.

Even Anthropic itself has said when we tell it to not do things outside of its goal, they don't. I don't. I think we're giving a lot of people the ability to accomplish malicious things if they want to use it that way, which I don't think that's a good thing. I'm just saying, I don't know. And hey, I'm open to being proved wrong. And this also doesn't mean I don't want to regulate it. I'm just saying I don't know if talking about it as if it's going, it's about to gain consciousness and be like, I have all these goals I want to accomplish.

Speaking of goals that you want to accomplish, I just have to back up for a second. You know how I would beat Magnus Carlson at chess? Fucking checkmate and I'd take out a gun and I'd fucking shoot him in the head. That's what you wanted to go back to say? I was waiting. I wanted to make a joke about killing Magnus Carlson. That's all. That's number one.

Jesus Christ. Number two is how did the market react? Well, at first, you know, you're like, how's Magnus going to beat me? And I wanted to be like, oh, he's going to take out a gun and shoot you in the head. That's the most unexpected part. But everything you're just, everything we were just talking about, interestingly, dude, all of the cybersecurity stocks have just been fucking rocketing in the last couple days. I mean, that was the first thing I thought about. I was like, is basically everything I have online.

at risk at this point. I mean, we had this conversation a long time ago when everything was going crazy about quantum computing. It was like, this is the end of Bitcoin, the end of online baking, the end of whatever, like, there's no encryption on anything.

You're just like, okay. Do you see that Bitcoin dropped quite a bit today because the Clarity Act failed to pass too? All right, let's round it out with some of the, some of the things from the Anthropic Misuse Report. I just have to share this because, first of all, this thing is long. So I just want to highlight a couple things. It spans the last eight months, and it includes, it includes suspected state-sponsored groups,

financially motivated criminals, commercial spyware vendors, state propaganda institutions, and politically motivated individuals, ranging from fake dating apps to surveillance systems built to identify and monitor dissidents. and it goes back to a thing that we've talked about before, these hacker groups, and they've got cool fucking names. There's this one case that was Russian espionage

linked to a group called Midnight Blizzard. Wait, let me stop you. I just figured out how I would kill Magnus Carlson. Go do you. Tell me how, come on. I would mess with his prescriptions and make him all loopy the night before so it looked like I actually did.

Why do you assume he's on prescription meds? Seems like a type. Can you Google Magnus Carlson? him for me? He is on prescription meds? He looks like he... Yeah, he does. He looks jack to the gills. Oh, good game. God, dude. He looks like a...

He looks like the middle... The second to last point of evolution between monkey and man, doesn't he? He looks like Charlie Kirk if he took a different path in life. He looks like Charlie Kirk, if Charlie Kirk got like... He shot in the...

...mail version... Whatever the male version... of like beautification, like too much plastic surgery, you know? Charlie Kirk. Yeah, he looks like instead of a bullet, if they shot him with like a big thing of, or just a flat, a big wooden mallet to the face.

Yeah. Yeah. Wouldn't that be cool if they shot Charlie Kirk with something that made him good at chess? A brain injury that, yeah, made him like. And Republicans were like, it's not funny. He's so good at chess now. And we've lost him otherwise.

He's no longer debating college students. He doesn't care about abortion at all anymore. Yeah. Don't fucking joke around. This guy's funny to me, Magnus Carlson. Look, he's wearing a fucking jersey. Anyway, this stuff is fucking scary, gang. This Midnight Blizzard ran operations attacking military intelligence targets in Ukrainian and European governments.

Then there was one, these all have their own case numbers. G-514. The Shiny Hunters Collective, known for several large... I feel like I'm in the Matrix movie when I'm watching this. Like, this is the shit that...

What's that? It's all happening. It is all happening. This is like what Morpheus and Trinity were all involved in... Not really, but... The shiny hunters collective... No, listen.

Known for several large-scale data theft operations, followed by pay or leak extortion demands. They exfiltrated more than a terabyte of data, including hundreds of thousands of national identifiers and millions of payment card records. Then there was one conducted by Chinese speaking operators likely residing in Changsha in China's Huanon province. And they were students. Two of the operators

were identified as undergrad students at the Chinese University in HUNAN studying computer and communication engineering. There's probably not in there, but I'm sure you saw the headlines that the Houthis were using Claude to build the AI guided missiles? Yes, I have that in here. Yemen-based guided guided weapons engineering cell.

And there's one... This one's scary. We identified and removed an account that used Claude to mass produce and rewrite political content. The operation used the model to rewrite and distribute

fabricated news stories across approximately 70 fabricated news websites. This network did not stick to one political ideology. Instead, they shifted political stances to support different sides of the political spectrum based on whoever was paying

at the time. Who else are you going to do it? They found some stuff where they were fucking with elections in Malaysia, and just to rattle off the last few. China-based surveillance and recruitment targeting Uyghurs in Syria. China-based intel operation targeting

Catholic, Tibetan Buddhist, Taiwanese Christian communities. China-based public opinion monitoring and dissident surveillance operation, Russia-based operation to engineer an autonomous military drone swarm. So that's fun. And finally, a China-based operation to collect Intel on directed energy weapons and their supply chains. All in the last eight months, just from Anthropics' own internal investigations, for their tools.

So you can only imagine what's being used with Google stuff, with the open model stuff, like Quinn. weird or open-al-ed-al-ed-all. Oh, yeah, and Open AI, of course. And this is just what they caught. It is pretty impressive what they're, I mean, again, this report is so lengthy and there's so much jargon that you've got to understand what you're looking at. It's really impressive what they're capable of doing and finding and figuring out internally. So that part makes me to counter how scary this stuff is. That does make me sleep a little better at night, that they're able to, at least be able to pinpoint some of the shit. And they shut them all down. too. I don't know, man.

Let them go. Let me use the thing the way I want. Let me create a little swarm. Let me start some, let me start some shit with the Department of Education. Oh, because you're loans? I think I would, I think I would do the IRS. Do the IRS. Do them all. And make it, make it so that they made a mistake and they've got to give me all my taxes back. I don't want to be too greedy. Just give me my taxes back. I don't like, I want my money back, please. Thank you very much.

Thank you very much. I would say make billioners pay their taxes. Make no mistakes. Ooh. Make, make, give me all the assets from every other Ben Con out there because there is only one true Ben Con and it is me. Ooh, that's good. That's pretty good.

I couldn't do it for mine. I'd make like no money. Do you, there are truly no other Emil de Roses out there? Have you checked? I don't know. All I know is that every time I need a... Actually, I'm not even going to say it because I don't want someone to circumvent me. Can you say it in a roundabout way? Just like... I don't have to...

Oh, you need like a... I know what you're going to say. Yeah. A username or something. Yeah, yeah, yeah. People, yeah. That used to be the fun thing. Like, you know, someone... Like my friend had MapQuest on Instagram. Oh, yeah. Xavier. Well, Mew's got...

It's funny, you brought Mews up. Muse got their stolen. Well, it's unclear whether or not meta offered... So if you don't know, I think meta tried to buy Instinct, which we were talking about last week, the stupid chatbot.

They tried to buy it? AI assistant. I think so. And instinct was like, no fucking way. And then... Honestly, because Meta does what it always does, and they go, okay, we'll just build it. Yeah. And then release it to their millions of users. And they have Mews

and they obviously wanted a way to promote it and have people a little home base for it. The band, Muse, had the handle, Muse. And so either Meta offered them money or they just booted them because now they're Muse Band. Muse Band. Is it Muse The Band or just Muse Band? I think it's Muse Band. I think it's Muse Band. God, that fucking sucks, dude.

Which they shouldn't be able to do. I remember, I forget who was it was like a friend of a friend who had Kim at Kim. And Kim, whoa. And Kim Kardashian offered her some money for it. And she was like, no. So she still has it? I think so.

But I don't know. Dude, take it. I would say, yeah, I'll give it to you for a million dollars. Start high. I'll give it to you for $10 million. Oh, yeah, you're right. Start higher. Start high.

$10 million. That's cool, man. At Kim. Someone's got Ben Khan, the email address at Gmail. Some fucker, some shit. Someone has at Ben. I mean, that would be so cool. if you were that smart to get it that far in advance?

How you doing? This was a good episode. I'm good. I'm like, uh, it's hard to not get overwhelmed by, you know, these people,

it's just, yeah, we live in a world where it seems like people are just completely uninterested in governing. I think, especially the American government right now, and for a long time, feels like,

I think people were like, oh, you know, like it sucks. We have Trump during this time. And I'm like, I guess, but like, okay, before that we had Biden. I don't think Biden was particularly suited for this. And then you're like, we had Obama before that. Obama was very tech friendly in a way that I don't think was good for us. I think he probably would have continued to give big tech everything they wanted, right? I mean, again, completely uninterested in governing.

I think it's a full, the complete tech consolidation all happened under Obama. You get these massive, massive tech companies, which I think is part of the problem we're dealing with now. And American tech went from being that kind of like scrappy thing of like, oh, you know, we're trying to like come up with new things and what to just. I think the frontier labs want the same moat that the big tech companies got in the 2010s where we no longer have to compete with companies.

If anyone starts to do anything remotely competitive or unique or cutting edge, buy them up. Buy them out or so. Buy them up. And I think yeah, I think that's part of it, right? it's like they've sunk all this money into fucking chatbots and they need chatbots to work out.

You know, um, like stuff I definitely don't comprehend, but I think like Google DeepMind and Google Labs and whatever has, has given up a lot of the work that could have been like genuinely groundbreaking. Like, I don't know, they had some like proton or protein fold. Yeah. Where it's just like, from what I understand, which is very little, uh, really little. regarding that kind of stuff, could have been like genuinely... Did they stop it? I believe they've gutted a lot of the resources going into it.

And so it's stuff like that where they're all just like, we need chatbots to work. I'm so glad that they didn't all follow Zuckerberg and do the Metaverse and like VR. Where, oh God, instead of we need chatbots to work because we put so much money to it, we need VR to work. Sorry, guys. it's only it's it's it's metaverse from here on out thank god having that shit foisted on this instead i just i i that might have been better who knows i i don't know if i believe in our ability when i say our i mean the government's ability to work with the private sector like this to create meaningful regulatory frameworks that everybody can agree on and everybody can

effectively implement and follow and regulate. I just, I think everybody's too stupid, lazy and distracted, frankly. I don't think we've got it in us anymore, especially something this advanced and technologically complex and stuff. You're really putting a lot of trust in these institutions to follow the rules and stuff. At least with banking now, it's like, okay, you know, there are, it's pretty easy to understand and pretty straightforward. Aside from the complex products that can be created that then gets convoluted, you know, packages wrapped in packages type shit. Yeah, I don't even know if it's stupidity or if it's just like, um, the role of a politician has just not been to regulate for a very long time. It's, it's been for decades now, they've been corporate managers where they are,

there to be front-facing to constituents and eat some shit from time to time tell people that you're doing the best we can but this is the way it's got to be and sell them on American excellence while the whole thing is stripped for parts I just don't

I don't think any of these people got into politics to there are some you know when we're talking about Lena Khan I'm like please fucking I'd love for her to be in charge Let us know what you think in the comments, everybody. What's going on? How are you doing?

How do you feel? If you're still here, you're a fucking freak. You are a freak. We haven't told you that in a while, but if you are still here, you are a little freak. Leave a comment if you're a freak. Say I'm a freak if you've made it. Say, I'm a little freaky freak.

Say, I'm a little freak and only real heads know what that means. Yes. Otherwise, hey, join us in the bonus, won't you? Ben andamielshow.com, sign up there. It helps us out and to continue putting good shit out into the world. We're going to talk about the Chinese song that this lady put out there that's really good that I have stuck in my head. Stinky fish in Mexico. I got to tell Ben about my new obsession with...

Emile's obsession? It's not a bag, but it's bag related, I'd say. It's driving me crazy. I need someone to set me straight. I think Ben, you might be able to... And maybe I'll tell you why I might look good right now. I don't know, maybe not. Oh, why you look healthy?

Yeah. Do I look? Do I look healthy? Is that what you were trying to say earlier? No, you said, do I look healthy? Well, you said, you look different. And I'm like, is that good or bad? Is it just that I have a healthy tan? Fuck!

Okay, we'll get to the bottom of this. We'll see you in the bonus. Thanks for tuning in everybody. Bye. But I... Coming up on this week's episode of Ben and Emile Show.com. I've got kind of a similar situation in my life as you. I've got a bit of a Santa Barbara.

What are you talking about? Is it a tennis racket that you want? No. Shoes? A suitcase. A suitcase. A bag. A bag. Is this what you were teasing? Or is this a different thing? Yeah, that's what I was teasing.

Oh. Emil's obsession. And it's not my wife, even though it is my wife. I already have my wife. She's not in Santa Barbara. She's at my house. Oh, boy. God, that was such a depressing era for me personally. Why?

Because I wanted to date Zoe Deschanel so bad. And I remember being legitimately bummed because she was married to... Ben Gibbard. And I was like, there's no way she'd be with another Ben after that. That's not how it works. You thought the barrier was only your name. Yeah. Well, I mean, no.

But I was like in my fantasy, I somehow... have access to... I will protect your heart, love. I fucking hate Death Cap for Cutie. Yeah, but he wants... So, he's saying, you come across a guy who signs a check with his left hand or whatever.

You go like, oh, you're left-handed. You're an abomination. I don't think that he's being that... I do. Prescriptive. I think that it's more just like, you know, these are things that you can say to your...

Bits to use. He let his bits to use. Yeah, but he's not... He's not nearly. as serious as the other one. These are just for... No, I agree. Yeah, yeah.

But I... I think this guy's better than her by a mile. Click bits to use. These are bits you can use in everyday conversations. Welcome back, class. I'm your professor Paul

and we're once again looking at bits submitted by you guys. First up on Instagram.