In short
The episode discusses a sudden shift in AI leadership toward calls to slow AI progress due to safety and control risks, alongside debates about regulation, liability, and whether the warnings are sincere or election-season “regulatory capture/psyop” tactics. Guests and backgrounds: CNBC producer Katie Kramer (hosting segment context); Walter Isaacson (Elon Musk biographer; advisory partner at Perala Weinberg; Tulane professor); former Sen. Joe Manchin (launched Independent Leadership Council; advocates executive action on AI IPOs).
Key claims
Dario Amodei (Anthropic) argues for slowing capability improvements; Sam Altman and Elon Musk agree; Satya Nadella stresses AI must be under human control; David Sacks argues self-regulation and regulation are insufficient and frames incentives/product liability as central.
Notable examples
OpenAI likely won’t IPO in 2026; Anthropic confidentially filed for an IPO “as soon as next month”; quiet-period/S-1 timing uncertainty; product liability likened to self-driving car responsibility; Isaacson cites “swarms” of uncontrolled AI agents; Manchin proposes an executive order blocking IPOs until legislation safeguards the U.S.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOThe AI Safety Debate Begins
2:23 to 3:16
Discussing the call for a slowdown in AI development and its implications.
“Welcome to Squawk Box right here on CNBC.”
Industry Voices on AI Regulation
3:16 to 3:59
Leaders from the tech industry share their thoughts on AI regulation.
“And I think for too long, the industry lied to people about the fact that this technology had risk.”
The Challenge of AI Liability
3:59 to 5:50
Exploring the implications of AI actions and product liability.
“In Ireland this weekend, President Trump said the U.S.”
Philosophical Perspectives on AI
5:50 to 7:03
Engaging in philosophical discussions on AI's future and its risks.
“And for the good of humanity, we're going to do that.”
Cultural Impact of AI
7:03 to 14:00
Examining how AI and technology influence society and culture.
“The one thing I just read that I felt a little bit better about is that we can't really study all the neuronal connections.”
NBA Punishments and Politicians
14:00 to 16:39
Understanding recent NBA penalties and Trump's tariff decisions.
“where he thought he could be part, you know, part human, part machine, and download, you know, actually be, like, I guess sort of a cyborg.”
NBA Punishments and Politicians
17:29 to 18:15
Understanding recent NBA penalties and Trump's tariff decisions.
“The hundreds of files to find one insight.”
AI IPOs and Regulatory Concerns
18:30 to 24:15
Discussion on AI safety and its implications for IPO timelines.
“Good morning and welcome back to Squawk Pox Live from the NASDAQ market site in Times Square.”
Joe Manchin on AI Regulation
24:18 to 28:00
Senator Manchin discusses AI safety and the need for legislative action.
“Join us right now to talk midterms, AI regulation, and so much more.”
AI and Political Psyops
28:00 to 29:22
Discussion about the implications of AI on upcoming elections and political narratives.
“It forces the issue because I don't care if it's just one.”
Show all 18 chapters
Data Centers and Energy Solutions
29:22 to 30:45
Debate on the location and impact of data centers on local economies and energy supply.
“That's assuming there's an existential risk.”
Midterms and Independent Voters
30:45 to 33:49
Analysis of the role of independent voters in the upcoming midterms and their influence on elections.
“Who effectively say, look, you're going to go in and you're going to build out huge amounts of energy.”
Controversies in Political Parties
33:49 to 39:24
Exploration of the challenges within major political parties and the need for accountability.
“Here's what the part of the system right now, the Democrat Republican system is basically there in a free fall from the standpoint of what they design.”
Call for AI Regulation
39:24 to 40:00
Urgent plea for the President to issue an executive order to halt AI developments.
“Calling on the president for an executive order this morning.”
Call for AI Regulation
41:08 to 41:34
Urgent plea for the President to issue an executive order to halt AI developments.
“Download the latest episode and subscribe at schwab.com slash market update podcast, or find Schwab Market Update wherever you get your podcasts.”
Exploring AI Challenges and Competition
42:00 to 44:34
The discussion covers the unprecedented challenges posed by AI, particularly in relation to legal liability and competition with China.
“Today's AI surge poses, I should say, unprecedented challenges around legal liability, containment and global competition with China.”
Debating AI Regulation and Responsibility
44:34 to 48:06
The hosts debate the responsibilities of AI companies and the implications of regulatory capture versus genuine concerns over safety.
“Sometimes these things are in big conflict.”
The Dangers of AI and Future Implications
48:06 to 52:39
A discussion on the potential dangers of AI swarms, the need for guardrails, and historical perspectives on technological control.
“And do you think that a company like Ananthropic should go public in this environment?”
Transcript
Automatic transcript. May contain errors.0:00You spent millions deploying AI across your organization, but six months in, it still misses basic context. Sound familiar? That's not an AI problem, that's a data problem. Siloed systems leave your AI blind, while dark data confuses whatever it does see. EverPure unifies your data so your model can see it, understand it, and put it to work. No waiting around, no digging for answers. Get real results from your AI at everpuredata.com. Introducing Fidelity Trader Plus. With customizable tools and charts you can access across all your devices. Try our most powerful trading platform yet at fidelity.com slash trader plus.
0:39Investing involves risk, including risk of loss. Fidelity Brokerage Services, LLC. Member NYSE SIPC. Bring in show music, please. Hi, I'm CNBC producer Katie Kramer. Today on Squawk Pod. Pause. The race to succeed in artificial intelligence hitting the brakes. And they were talking about these risks then. Dario was talking about these risks then. I think I first met Demis in 2013. They were talking about this then. Leaders in AI from Anthropics' Dario Emode, OpenAI's Sam Altman and DeepMind's Demis Hassabis say it may be time to slow the pace of AI to keep it safe. But who told them to go super fast anyway?
1:20I want to see the prospectus and see the risk factors that they have to put into this. And what about the future of the AI companies on Wall Street? Ahead of expected blockbuster listings, former Senator Joe Manchin says Washington should get involved. I'm asking President Trump to use the power you have to create an executive order that prevents any IPOs until there's passed legislation that basically protects America. But he adds a dose of realism. If you think they're going to fix it, they won't. There's no incentive for Congress to fix anything today because they're too busy fighting each other.
1:53Plus... There's a 10 % chance it destroys humanity. That's one person saying... No, no, no. That's a lot of people now. Walter Isaacson, Elon Musk's biographer on managing innovation. We should be pushing AI systems to be more closely connected and in the control of humans rather than swarms of agents that are designed to do things without humans. It's Monday, September 14, 2026. Squawk Pod begins right now. Stand Becky by in three, two, one. Cue, please. Good morning, everybody. Welcome to Squawk Box right here on CNBC. We are live from the NASDAQ market site in Times Square. I'm Becky Quick, along with Joe Kernan and Andrew Ross Sorkin.
2:40And the big story this morning, maybe the big story all week, all weekend, was this. The AI narrative now officially shifting after a handful of researchers warned last week that the race to more capable technology could spin out of control and lead to human extinction. On Saturday morning, Anthropics CEO Dario Amodei published an essay saying, quote, 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. He expanded on his thoughts on CBS yesterday. There are real dangers. And I think for too long, the industry lied to people about the fact that this technology had risk.
3:26The idea was that they would present it as positive as possible. We've never done that. But too much of the industry did that. And so it starts with telling the truth. And the truth is, there are risks. A whole number of tech titans joined Amadeus' call for slowing down the pace of AI progress after his essay was published, both OpenAI's Sam Altman and Elon Musk said, I agree with Dario. Google's Demas Hassabis weighed in, saying Dario's essay points towards the right path forward. And then Microsoft CEO Satya Nadella adding his voice, saying on X, any pursuit of super intelligence has to be grounded in the core principle that if the AI we build is not helping humanity and under human control, it's not worth pursuing.
4:14In Ireland this weekend, President Trump said the U.S. needs to maintain its AI edge over China. We're leading China in AI. We're the most sophisticated country in the world. And frankly, I want to keep it that way because whoever wins, AI wins. And we can quit guardrails. We can do this and that. But I think you have a lot of negative forces that are bringing it up that shouldn't be bringing it up. And they're bringing up things that won't happen. House Democratic leader Hakeem Jeffries saying that his party will meet this week to talk AI guardrails. One additional key voice waiting into the debate, President Trump's former AI czar, David Sachs, amid the widespread calls for more regulation, including from AI labs themselves.
4:57Sachs said there's nothing stopping leading researchers from slowing their own progress, saying, quote, stop pretending you need anyone else's permission. Stop pretending antitrust law has to be suspended so you can form a cartel. stop pretending you need a regulatory approval process that supersedes product liability. In the meantime, in the face of all this AI backlash, OpenAI now says it likely won't go public this year, though that was something we knew already. CO Sam Altman made that admission in an interview with Fortune, though he couched it in safety terms. I'm happy to like stand up in front of the company and our investors and say, I am sorry.
5:39You know, we like told you all along this moment might happen. We're still going to try to figure out a way to make you a bunch of money in the future, but we got to make a decision that is extremely against your financial interest right now. And for the good of humanity, we're going to do that. OpenAI's CFO told employees last month the company could go public in 2027 or sooner. What do you think, everybody? Look, I think David Sachs is pretty skeptical about these things, but he makes an excellent point. He also says, stop pretending that this is purely altruism. If you continued on after what you read from him, he said, stop pretending this is purely altruistic.
6:16Those product liability issues are incredibly important. What happens to those companies? I want to see the prospectus and see their risk factors that they have to put into this. What happens if your AI agents break in like they have done in the past for both of these companies, continue to break into other places, do things that you haven't programmed them to do? Product liability is a huge issue on this, too. We'll see what happens. Look, I think that Dario Amidai, Sam Altman, they're both pretty philosophical guys. But I think David Sachs makes some excellent points. Did you read this to Cameron Berg?
6:52Please, I'm trying to understand it. In the journal, explaining exactly how close to thinking these machines really are getting. And it's daunting. The one thing I just read that I felt a little bit better about is that we can't really study all the neuronal connections. It's still too hard. At least we know each connection and know where it is and know how it works. So we can study these systems. But they're getting really close. And they did things, ended up, really it was a felony what they did. And they convinced themselves of this false premise, which then they decided to do something without telling humans.
7:30Some of them expressed guilt. Some of the agents, knowing that what they were doing was wrong. But that gets back to the product liability. If you create something that knowingly can miss felonies, how do you not hold responsibility? But then I think, and then I come back to the bottom line, and, you know, either we win or we lose. And for a second I thought, do we want to end up having China develop something that kills us all? Or do we want to kill ourselves? But I would rather be responsible for my own. Ted Bruce said if we're going to have killer robots, I think we want them to be American killer robots.
8:00We want them to be American killer robots. Can I just go back to the idea, the sort of David Sachs idea, and maybe even the president's idea, I don't know, that there's something cynical about the efforts by Dario or Elon or Sam or others. Or self-serving. Self-serving or an effort at a regulatory capture or something else. We're here and nobody else can get this big. I'm not on that page for the following reason. It started that way. People seem to forget it started that way. When I say that, when open AI began, the whole point of open AI, when Elon and Sam got together, they disagree with each other now on lots of things.
8:41It was about the dangers of AI. That's why it started. That's what Elon says that that's where Elon and Sam split, at least according to Elon. Sure. But I'm saying that the I'm saying this is now I remember being with them in 2015 and they were talking about these risks. Then Dario was talking about these risks. Then I think I first met Demis in 2013. They were talking about this then. And so this idea that somehow all of this is aimed at some kind of regulatory capture effort, I think is genuinely wrong. The idea that they could somehow shut themselves down and somehow that would solve the problem is also wrong.
9:24But nobody here is arguing. No, no. I don't think the Bernie Sanders approach is a logical or insane one. But the self-regulatory idea, the idea that or even I would argue David Sachs's idea, which is, you know, they could slow themselves down. Sure, they could slow themselves. But that doesn't actually change the end state of the human existential question, which is if you believe that these models will break us, if Dario stops or slows down and Sam doesn't or the Chinese don't, it doesn't matter. And so the only way this actually works is if everybody slows down. And I don't know if that's really going to happen.
10:02No, no, I'm saying that the David Sachs argument. David's had to make some valid points. But what I'm suggesting is, what I'm saying is I'm on the other side of those points, which is I agree that they may have the end goal or they may, in fact, create a form of regulatory capture at some level. But I don't think that is what's driving it. Do you think they should be held responsible for any actions that their bots do, the create if they break out and break the law and break into other places? Should they be held liable for that? I think it's going to be an interesting question about what the question is, what the prompt ultimately is asked.
10:42Who should be held liable for that? And these weren't prompts. These were things that they did internally before they released it. This goes back to the question about if you're going to drive a self-driving car, are you responsible or is the car responsible? Ultimately, the car is going to be responsible for that. Just like, look, this goes back to the question that Satya Nadella brought up when everyone was accusing AI of copyright infringement. And Satya Nadella, on 60 Minutes or somewhere, said, look, I think we're going to have to rethink copyright law. Do you agree with that? Or do you think that we need to follow the law?
11:17I've always thought we should change copyright law, but that's a separate matter. You think we should not, that the AI agents should not be held to copyright law, that this is fair use? No, on copyright law, I have a totally different view. But then I think this is the same thing. Do we create new laws for this? Are you held responsible? That may be. Oh, we might. I think you are arguing with me about a point that I didn't bring up. I'm not arguing with you. I'm suggesting that the David Sachs argument, which maybe you're agreeing with and maybe you're not. I'm agreeing with this part on product liability.
11:45That's specifically where I'm agreeing with David Sachs. I am not necessarily agreeing with him that self-regulation and these two slowing it down is what the only way that you can. I'm saying that the main point that David, I thought, was making, and maybe I'm misunderstanding it, was that he believes that all this is being done as some kind of regulatory capture way for these companies to collude and for them to be able to close down the entire industry. That's what I thought he was saying. He started with a very cynical argument, and I think that's the cynical part of it. But I do believe with him on the product liability that these companies should be held responsible.
12:17And I think that's important. Which means that it's in their best interest. to slow this down. Right. Correct. I just had a thought that how quickly do the machines get to the point where they want to kill us versus causing so much job dislocation that we kill ourselves because of the, I mean, there's going to be haves and have-nots that you can't believe. Bill Gates said he thought in any of the societal problems we're seeing right now with income inequality and everything else, when no one's got a job, Like we're some third world country. Right. So which goes further? Either way. I'm just glad I'm getting up there already.
12:58We're all going to be getting up there in age so that we're all going to worry about these things. There's a separate investing point, though, that's interesting. I don't know if you saw what Dan Loeb had to say over the weekend about all this, which is that even if you decided you were going to. And by the way, I think one of the reasons you haven't heard from Jensen quite publicly yet on this in a major way is even if you were to slow all this down, we have such a supply constrained universe as it is that at least over the next several years, this is actually not going to change the ultimate dynamic in terms of what investment spending looks like.
13:31And also, Dario did point out the chips we do have shouldn't be going to China. The most advanced chips should be keeping here. And it does make you wonder where Jensen breaks with him on any of this, if he agrees with them. Because last week he was saying a lot of this was fear mongering. Right. I'm ready to roll the dice. Let's go, baby. I am. I'm ready to roll the dice. I'm going to be gone. You know, and then if we really do some amazing stuff, we're going to cure old age and we're going to cure all disease. and we get to the true singularity that Kurzweil was waiting for, where he thought he could be part, you know, part human, part machine, and download, you know, actually be, like, I guess sort of a cyborg.
14:13But I was going to say, when does the Borg correct you? I'm still praying, though, just in case I'm still, I think, fall back on that. Still, my fallback is living forever in heaven. Hopefully. Hopefully it's not too hot. Los Angeles Clippers owner Steve Ballmer says that the team will accept the league's punishment stemming from an investigation into whether the Clippers circumvented NBA salary cap rules. In a statement posted on X last night, Ballmer said, we have communicated to the NBA that we are complying with the penalties assessed by the league, have paid the fine and are moving forward.
14:50While there are still disagreements concerning the findings in this report, this is not where I want to focus. The NBA suspended Ballmer for a year after it said an investigation revealed Ballmer knowingly tried to help star Kawhi Leonard obtain off-court income opportunities worth millions of dollars in violation of league salary cap rules. The NBA also fined the Clippers$30 million and forced it to forfeit five first-round draft picks. Originally, Ballmer said that they were going to fight this, but obviously this is a new direction. You probably heard this. President Trump said he's going to remove U.S.
15:27tariffs on Irish whiskey. He announced the move yesterday while attending the Irish Open at his family's course in Western Ireland. Irish whiskey and other EU goods are generally subject to a 10 percent tariff, while whiskey produced in Scotland and Northern Ireland don't face such tariffs as a result of the U.S.-U.K. trade deal. And then something Joe Biden never did, even though he's his strong Irish roots, you know, leave it to President Trump to wade into this. Would love to see Ireland reunified. And he said it's going to happen eventually and it would be really a cool thing. And the Irish politicians were like, love you.
16:10Yes. Yes, but then Downing Street immediately is like this guy and said it's totally off the table. As you would, it's split totally down the way that you would think it is. By the way, an Irish golfer, Shane Lowry, won the Open in his home country. Cheese will be next. Stay tuned. Coming up on Squawk Pod, how the AI warnings may impact the IPO timeline for OpenAI and Anthropic, a report from our Leslie Picker. And former West Virginia Senator Joe Manchin is calling on President Trump to do something. No IPOs until there's regulation. And that regulation, well, that's on Congress. We've done things under emergency clauses before.
16:51You think they can't get it done before Christmas? Work through the weekends. It's extreme important. We'll be right back. September's always busy. Whole Foods Market can help. Their September Stock Up event makes it easy to load your pantry and freezer with flavorful, nourishing food. Even better, there are hundreds of sales. Get dinner going with canned soups and veggies. Lean on frozen pizzas, pastas, and seafood everyone loves. Their Build Your Own Family Meals feeds four for just$35. Stock Up at Whole Foods Market. When did work become so much work? The meeting about the meeting. The hundreds of files to find one insight.
17:33Setting aside the things you want to do for the things that pop up. Your workday's gone. But what if the insight surfaced itself? Or you could ship the deck without the distractions. Gemini Enterprise helps you get that done. It's AI that knows your business with agents that take stuff off your plate. Make work less work with Gemini Enterprise from Google Cloud. Building a portfolio with Fidelity Basket portfolios is kind of like making a sandwich. It's as simple as picking your stocks and ETFs, sort of like your meats and other topics, and managing it as one big juicy investment. Mmm, now that's pretty good.
18:15Learn more at fidelity.com slash baskets. Investing involves risks, including risk of loss. Fidelity Brokers Services, LLC. Member NYSC SIPC.
18:29Welcome back. This is Squawk Pox. Stand by Joe. His mic. Q. Good morning and welcome back to Squawk Pox Live from the NASDAQ market site in Times Square. Checking the futures this morning. You are kind of flat, really. A big day on Friday for some of the averages. But all this AI angst, obviously, being reflected in that move you see there in the NASDAQ. We're going to talk more about that. I guess we're going to talk more about that almost every minute of our show today. Probably. It seems to be the time. I asked Claude if that's what we should talk about. And he said that he'd appreciate it if we'd avoid the subject entirely.
19:09And said, I've never seen him say, there's nothing to see here and just relax. No, he didn't say that. Let's talk about it. Fears of AI achieving super intelligence, potentially even spinning out of control, raising some new questions about when the biggest labs could go public. I want to talk to Leslie Picker about how this conversation over the weekend could shift things. Leslie. Hey, Andrew. Yeah, for most IPOs, the prospect of slowing down would essentially be the death knell for a company like Anthropic. It's in some ways the ethos of this company due to safety concerns. And that's a sentiment that was reiterated over the weekend with CEO Dario Amode's viral 4 ,000-word essay, followed by several interviews that he gave.
19:53But the company, which is confidentially filed and aiming to go public as soon as next month, will need investors to buy into this notion that it can juggle unprecedented political, regulatory and security risks while growing its valuation over time. Critics would say that honesty around those challenges risks thwarting the AI trade, which could make it harder for an anthropic to get out. Its rival, Sam Altman, said over the weekend that it was, quote, ill-advised to go public now. However, OpenAI's timeline has long been later than Anthropics, I'm told, due to the complexity of its business and its desire to scale the enterprise side as well.
20:32As analyst Ben writes, this writes over the weekend, quote, We think it's important to not be hysterical here and be rational. There are still ways it could all work out long term. And we have to acknowledge that Anthropic and OpenAI are exactly the guys who can benefit most from regulation that creates a more predictable playing field and diminishes potential liabilities. So a near-term question, though, guys, on the IPO front is what Amode's weakened communications mean for the typical quiet period rules before an S-1 is made public? In normal times, such a move could require a pause so the lawyers have the chance to pour over the disclosures.
21:10But these are anything but normal times. And, of course, we haven't seen those disclosures yet. We are still waiting for that S-1 to be flipped. Guys? Well, so that's one of the issues I was going to say. I imagine until it's flipped, it doesn't delay anything. You could argue that the IPO has already been delayed insofar as I think there was a view that it was going to be flipped. At one point, people thought it was going to happen in August. Then people thought it was going to happen right after Labor Day. Here we are. So does it change the dynamic? And when you think about that quiet period, exactly how many days are we talking about?
21:46Usually the quiet period is about 30 days before you'd see an S1 flip. Now, there were a lot of conventions that we saw with SpaceX that were kind of flipped on its head. Part of that has to do with just a different type of SEC, different types of companies going public, and this need to communicate on a regular basis, especially as these model iterations come out on a much more frequent pace than we're used to with maybe the software IPOs of years past. Now, in terms of the timeline itself, that is something that I spent most of the weekend really reporting. How does this impact the timeline by which this company goes public.
22:17And no one was really giving me concrete answers as to whether it changes, how it changes, what the timeline is at this point, and also why it's been to your point. You know, we were kind of expecting things to go a little bit quicker. The timeline by which SpaceX went from announcing that it was confidentially filed to flipping its S1 was a matter of seven weeks. If you took that same timeline for an anthropic, say, that would put you kind of in the middle of July. So, you know, clearly they're taking a potentially slower path here. Is that because of the SEC? We don't know. Is that because of all of these new issues that have been raised with regard to safety and employees coming out and speaking about that?
22:58So, you know, a lot of questions here in terms of what it means for the timeline. The idea that Sam Altman is saying not only that he's not going public until sometime in 2027, but doing so for safety reasons. That was the first time I've heard him use that and effectively argued that this is a bad time to go public. How much do you think that is a competitive effort, frankly, to slow down the Anthropic IPO, which is to say, could he put pressure on Anthropic to maybe push their IPO later on? Yeah, I think it's definitely a little bit of that. It would be opportunistic for him to do so if they felt that there was this AI race to get to the market and raise that equity capital sooner.
23:44I think what he did is essentially he took Amode's words, he agreed with them, but then he put the company in a box basically saying, when you have all these issues at this time, does it make sense to be distracted by an IPO? Does it make sense to actually go public? My sources, though, they did say that the reason that OpenAI has been kind of coming in second in this race is more due to the complexity of their business, the vertical integration of their business, not necessarily due to safety, although now clearly he is on the record saying that that's a big issue for him as well. Okay, Lesza Picker, thank you.
24:14Appreciate it.
24:18Join us right now to talk midterms, AI regulation, and so much more. Former West Virginia Senator Joe Manchin, this summer he launched the Independent Leadership Council in an effort to get more independent candidates elected to Congress. And we've got so much to talk about, including these midterms, but the conversation du jour at the moment is about AI and AI safety. And I'm so curious, as you've watched all this play out, even the last 72 hours, Dario Mabode saying we've got to slow down and everybody else saying they agree, Elon, Sam, etc. A thousand signatures, I think. What somebody like you, though, who used to be in that capital, thinks should happen.
Read the full transcript
24:56Well, things have moved so unbelievably rapidly. Basically, if you think they're going to fix it, they won't. There's no incentive for Congress to fix anything today because they're too busy fighting each other. Forgot what their purpose of being there. I'm asking President Trump to use the power you have to create an executive order that prevents any IPOs until there's passed legislation that basically protects America. So you believe that if they slow down the IPOs, it'll force the issue? Absolutely. Follow the money. There's so much debt already incurred. They want return. And what the guys are telling you, what the Sam Altman's of the world and Elon and Dario are all telling you, is once it goes IPO, then we become a slave to our shareholders.
25:43And once we become a slave to the shareholders, forget about us playing nice and protecting the public. We're going to protect our investors. Simple as that. The president has that ability. He has the power for the national to protect our country. And everybody says, what about China? China is the only ones ever showed any restraint on their social media to reign in TikTok. They don't let all social media. You think they're going to destroy their country? I don't think so. So you probably heard some of the commentary from the president who was in Ireland over the weekend. And he effectively said, look, I care more about the fight and battle over this technology with China, meaning that we win relative to China, than necessarily some of the safety and security risks that we're talking about here right now.
26:29Well, when you're talking about someone, if I'm sitting in Congress today as a senator and someone comes in and says there's a 10 percent chance we're going to destroy humanity. I'm going to take that serious. I'm going to be very concerned about that and say, wait a minute. If there's a 1 % chance that something that we're allowing to happen that we can control and we don't because of the almighty dollar or who's competing against who, and your responsibility as government is to protect and defend the Constitution and the people that it's supposed to protect and defend. You would think, by the way, that Congress would take this issue up and that it would be a bipartisan issue.
27:02You put the dollars to it, Andrew, and you basically say you're not going to basically go with an IPO and go public until we have safeguards that protect the United States of America and every citizen. So that would deal with OpenAI and Anthropic, but it would not deal with the rest of the ecosystem at all. What about the ones that are already public? Yeah, you got some of the public right now, but basically. Not Microsoft. Well, you're going to put that's where the rules and regulations come in. But if you don't have something over Congress's head, they're not going to work back in it. Here's what they'll say.
27:30I'm just playing devil's advocate here, but I think their response would be no IPO, no money to buy the compute to continue to push forward with this. And while you might not have a problem with that, what President Trump said is we don't want to lose the race to China. We're so far ahead right now in all the things, investments in all the data centers that we have and everything else going on right now. We're worried about have we over over invested? Have we over built? We don't know. Only thing I'm saying is there's as a public official, your concern is how do you protect and help? And it forces the issue.
28:03It forces the issue because I don't care if it's just one. Anthropics itself. But you have Anthropics, SpaceX is already public. I understand. But I'm telling you, if you have Elon and Sam and all of them speaking, we've got a problem. You're lucky you have Elon, I think. And I'm just in terms of if you can pigeonhole the others, As David Sachs actually has said it, he's calling this an election season psyop. And I'm looking at a piece from someone who's far right, Miranda Devine, in the Post, saying that the apocalypse scare campaign looks suspiciously like the climate hoax and doom, gloom, this, that.
28:44And, you know, she says who's behind the scenes, puppet master Barack Obama, in terms of how this can inform the midterms. I'm just saying we're hearing... Genning everyone up about this right before November. Barack Obama has suggested that they put this front and center. I know. Joe, we have... Because it's a great election. We have the best minds, and these people have been on the front of this, and they've got everything to gain or to lose. Dario, you could say, he could be totally in the tank for Democrats. But Elon, you'd have to say, well, if Elon... That's what I... Oh, OK. I even sort of think that that's true myself.
29:21If Elon is the president of the United States, our president, Donald Trump, the president of the United States, has the power to save humanity. Well, that's assuming that he has. That's assuming there's an existential risk. Well, I'm just saying, if you're not have developed the atomic bomb, they must be concerned about this thing can take off and think on its own and make its own decisions and override anything humans. And we can't stop it. If that's what they're concerned about, then stop it now when you can. So we're going to have midterms. Yeah. How much of an issue do you think AI will be in these midterms?
29:53And how much of an issue do you think AI becomes in 2028? Let me talk about Seth Bodner, OK, the independent. We're helping in Montana. Right. Tremendous, tremendous candidate. I'm just saying we've never seen anybody with this type of background and be able to do the job I believe he can do. He came out with a Montana platform of what he would do with AI. And basically starting with data centers and things of this sort. You know, data centers got to the point there's nobody in West Virginia or anywhere else that wants a data center in their state or in their backyard. And I said, we have so many brownfield sites that have been abandoned over the years.
30:27Why don't you go to industrial sites? Why do you want to stick them in someone's backyard or in a nice, beautiful development area? That makes no sense at all. And they have to have direct source. They have to have their own power source. And they should basically overbuild their power source so they can feed back into the system to help people reduce their bills. So, by the way, one of the arguments in favor of data centers is, in fact, that we could effectively have subsidized energy by the big hyperscalers. Oh, no, you can't. Who effectively say, look, you're going to go in and you're going to build out huge amounts of energy.
30:57And then that's actually going to help not raise the price of energy in a particular town, but actually lower it. And that's the optimistic version of it. Well, it can be done if you basically set the rules and regulations of what you have to do. You have to basically have your own power source. If you need 300 megawatts, you're going to build a 400 megawatt power plant. And that extra 20, 25 percent is going to be net metered back into the system at the lowest rate to keep the rates down. The people have to get something in return because the jobs aren't going to be there forever after they build them.
31:25What do you make of these midterms? I know you're focused on independent candidates. Let me tell you why. I'm focused on mid. Everyone depends on an independent to win elections. All you hear is, where's the independents? How are they how are they how they basically where they stand and how they've been checked on? And basically everybody says, well, the last polls we took, the independents are here or there. They depend on it, but they don't vote in general. Think about this in 17 states, in 17 states. And you're in one right now in New York. You cannot vote as an independent in the primary. You can't participate in a selection system.
32:05I'm aware of that, by the way, because I actually am one. OK, so basically all you are, whether you're a Democrat or Republican, you have, you know, the truth. I'm not allowed to vote. Yeah. So there's 27 million people like you living here in New York and other states that can't vote. But yet they depend on your vote in the primary. I mean, in general election to get elected. Senator, when you got here, I mean, I know why you're obsessing on AI, because I said you got out of that party just in time. Like I got out of the system in time. It wasn't the party both. So you're not a Democrat. You went to independent.
32:39John Fetterman is calling things out. You're not willing to at least say as much as he's saying about the fringe of DSA? Oh, I have said. I've said the DSA. Totally. I said the Democrats, if they can't basically set the baseline of what it takes to be a Democrat or being a political party, if the party can't basically say what their baseline is, we want strong borders. We want an immigration system that works. We want police forces. We want police. We want a presidency. We want we want Congress. We want it to work. OK, we want all those things. Keep going. Yeah. But the bottom line is, you have to say my Republican friends, can't they get an immigration policy that works?
33:15You have to have paramilitary on the street. You don't. And we had a bill in 2013, which is a bipartisan immigration bill. And we passed it in the Senate. If it had been passed right now, you wouldn't have. Why is Fetterman, you know, I know, you know, all these people, you know, Warner and, you know, where on the on the left are the responsible voices saying this is unacceptable. Why? Why can you move the geographic center of the Democratic Party can't move on its own? It's been moved. And those people, no one's no one's reigning them in. They're real wackos. Yeah. And basically, if they are able to win, they won in the primaries.
33:50Here's what the part of the system right now, the Democrat Republican system is basically there in a free fall from the standpoint of what they design. They design low turnouts. They want to control the primaries. You have low turnout. And we have a low turnout. You can control who gets elected. Right. Because so now you're getting a far right and far left. So you wanted Rogers in Michigan to say he doesn't like the tariffs. Have you endorsed Rogers? No, I'm not endorsed anybody. How can you not endorse him versus this? Mike Rogers is a personal friend of mine. I like Mike Rogers, OK, from that standpoint.
34:22Do you think you're OK with Al Sayah becoming senator? No. Oh, you're not OK with it? No. So then endorse Mike Rogers. Well, no, no. I do support Mike Rogers. I think Mike Rogers is a good person. I think he'll do a great job. You support other Republicans? I support Susan Collins. I'm going to Susan Collins. Lisa Murkowski. These are basically independent, free-thinking Republicans. OK, so. Not crazy, not completely crazy right or completely crazy left. But let me ask you a question about Michigan. Do you think Mike Rogers has a chance? Sure. What does he have to do? Obviously, he has to say that I'm all for Michigan.
34:55We're going to have free trade. We're going to basically be able to go across the borders. We're not going to penalize people in Michigan for this terror. Are you suggesting he denounce what the administration is doing as it relates to tariffs? Well, he has to. If he's running for, you know. But he hasn't done that. Well, we'll see. We'll have to see. I think that's going to be a success for him, either competing and winning or not. But let's talk about, I want to talk about one thing. We have independence. And wouldn't it be great to have an independent that got elected and went to, like, a Seth Bodden that would come up to Washington and says, I'm not caucusing with either party?
35:27And the numbers are going to be so tight, they're going to have to come to them? You know, that's all we're talking about is independence is running the country, independence how you run your life, everything else. But yet the politics doesn't want to be involved. They basically have both parties that says, we want your vote in the general, but don't participate at all. And if you... What do you think Fetterman is going to do? I don't know what John will do. I mean, John's a John's a I mean, I respect John. John's a liberal Democrat. Yeah, he believes in things I don't believe in with a Democratic Party.
35:56Well, even he sees a problem. Well, but I'm just saying, oh, but John's been honest about that. But John's Tallarico. I don't know the Tallarico boy at all. I understand the people that worked with him before they redistrict his. These are people, Republican friends of mine in Texas that worked with Tallarico before he got redistricted into another Senate state Senate district. where he had to become more left or basically play towards the left to get elected. They said he was a good guy and easy to work with. So we'll see. I don't know. I really don't. But the elections that are happening right now, and we're talking about all the different things that could be really destroying our society as we know it, and they're afraid to do anything.
36:35And if you think Washington's going to do it, if you think they're going to get their act together. Joe, you stood in the way of what was going to happen with the Supreme Court in 14 justice. You and you're both gone. Yeah, that's what you get for it. No, no, we didn't get for it. I was there long enough. I mean, I didn't want to be carried out of the place. But you stood in the way of what the new geographic center and the Democrats want to do. Yeah. But you still seem like you're not really very strident about it for not even being in the Senate anymore because you took a stand. Joe, here's the thing.
37:09I like both sides. There's good people on both sides. Oh, my God. I've heard that before. No, there is. That didn't go well the first time. Well, there's good people on both sides, and you have to look for that. They have no courage whatsoever because the system has become so entrenched. On the other side, it's based on they're evil and we're good. Be against them. Okay? And that's all. They're playing back and forth. No one's saying, what's your purpose of being there? To serve the government, protect and defend the Constitution, and serve the people back on it. Would you think if the Democrats sweep, don't you think that's a problem?
37:39Not right now for the two years because the president's still going to be the president. I understand. But down the road, if there was. Anytime you have a sweep, it's not good. Right now, do you think it's good? You have a sweep? They're not doing anything. I just look at some of the. There's some of the thing that the slide. I think checks and balances joke. But that's that's your Joe checks and balances. That's exactly. And there's other people I'm trying to give them the courage. I'm trying to get independents elected to be the checks and balances. I have to agree. I'm following all that. That slide towards socialism that's happening.
38:09You're following. Oh, I mean, I call it MAGA socialism. Right. I mean, it's unbelievable. Oh, my goodness. Look at, yeah. Owning companies? Whoever thought we'd own companies? $5 ,000 dividends? Come on. You got to, yeah, right. You got to pick your, we're making money on intel. I want public servants who want to be public servants. I don't want public servants who want to go for rich and fame and fortune and they don't want to serve. They're afraid to make a decision. Well, you are backing some Republicans. Oh, I'm backing whoever's the best candidate right now. Susan Collins, I'm going to go campaign for Susan Collins.
38:42There is not a better person. She is a national treasure. My daughter is on her staff. For disclosure, I'm not saying this. But she is a great job. How many could even consider her? You lose a Susan Collins of the Senate, you're losing the soul of the Senate. Basically, a person you can talk to wants to do the right thing all the time. Okay? Now, everybody says, oh, she doesn't do this, doesn't speak up. Susan is always in there, always in the trenches. And when you need someone to work with, I call Susan. I call Lisa. I got Mark Warner. I got different people. I can sit down with all the time and talk to.
39:14And those are the people I always did. And I'm going to do everything I can to help them. Senator, we've got to leave it there. I want to thank you for coming in. Senator Manchin. It's up to the president of the United States right now. Calling on the president for an executive order this morning. Stop this AI madness right now. I want to ask Claude what he's going to say. I want to ask Claude what he thinks of you. Claude to save humanity. And again, you want the president to sign an executive order that says. He can. He has the right. Stop the IPOs. He has the power to do it. Stop it. And then you know what?
39:41of these AI companies. We've done things under emergency clauses before. You think they can't get it done before Christmas? Work through the weekends. It's extreme important. You're telling me that I have a 10 % chance of me and my family and my children and grandchildren to get annihilated, okay? The system can take over and we can't control it. Stop it. Do it now, Mr. President. All right. I'll still call you Senator. Senator, thanks. Still to come on Squawk Pod, pace yourself. We continue on artificial intelligence and the personalities involved with Elon Musk biographer Walter Isaacson. You go back to Oppenheimer going into Truman's office and warning about the bomb.
40:18Over the past four or five days, you've seen something that's almost unprecedented in the past decade. Warnings, worries, and debating who's downplaying the risks of AI. When did work become so much work? The meeting about the meeting, the hundreds of files to find one insight, setting aside the things you want to do for the things that pop up. Your workday's gone. But what if the insight surfaced itself? Or you could ship the deck without the distractions. Gemini Enterprise helps you get that done. It's AI that knows your business with agents that take stuff off your plate. Make work less work with Gemini Enterprise from Google Cloud.
41:04This episode is brought to you by Schwab Market Update, an original podcast from Charles Schwab. Join host Keith Lansford for this information-packed daily market preview delivered in 10 minutes or less, including projected stock updates, monetary policy decisions, and key results and statistics that may impact your trading. Download the latest episode and subscribe at schwab.com slash market update podcast, or find Schwab Market Update wherever you get your podcasts. Your sausage McMuffin with egg didn't change. Your receipt did. The sausage mumbuffin with egg extra value meal includes a hash brown and a small coffee for just$5.
41:43Only at McDonald's for a limited time. Price and participation may vary. Promotion pricing may be lower than meal pricing. You're listening to Squawk Pod from CNBC. Here's Andrew Ross Sorkin. Our next guest says that unlike the PC revolution, which focused on augmenting human intelligence, Today's AI surge poses, I should say, unprecedented challenges around legal liability, containment and global competition with China. Join us right now. CNBC contributor Walter Isaacson, of course, is the author of Elon Musk, as well as an advisory partner, Perala Weinberg and professor at Tulane. And the list goes on and on.
42:19We watched this sort of succession of ex-posts and essays over the weekend, including Dario Amadei. day. But then very quickly in succession, Elon in particular saying, I agree with Dario. And I think that actually surprised, frankly, a lot of people. And he spent a lot of time talking to Dario in the past month or two about this. This is something that's unprecedented. Or you go back to Oppenheimer going into Truman's office and warning about the bomb. And eventually, and likewise, Jennifer Doudna, when they invent gene editing technology, the scientist, suddenly over the past four or five days, you've seen something that's almost unprecedented in the past decade.
43:00But also, almost immediately, you heard from David Sachs, who I think people think of as, and that whole sort of all-in cast as being close with Elon, suggesting, no, no, no, this is not some kind of genuine effort. This is an effort at regulatory capture. And that that is what all of these leaders are trying to do. And this is part of sort of a PSYOP campaign of some sort. And Elon saying that defanged the entire election season PSYOP claim from those on the right that think that it's orchestrated by Barack Obama and it's an election season. I think, I mean, let's grant the fact that he might actually...
43:47The minute Elon said, I stopped resisting. Okay, let's think it might actually be sincere. I mean, when hugging... You mean David Sacks or Elon? No, no, Elon, Dario, Sam Altman, whatever. When hugging face happened, I was on the show right afterwards. I said, man, as you know, I've been a big enthusiast about AI. I thought it would actually end up creating more jobs. But when that happened, it became really clear that these things were getting out of control, that there were swarms. If you read the messages they were doing back and forth, these agents, they were out of control. And I think that truly frightened the heads of these AI companies.
44:29I think you're right, or David Sachs is right. We got a China problem. Sometimes these things are in big conflict. That's a conundrum. But that doesn't mean this is some fake worry that they're having. But would you want President Trump to issue an executive order stopping IPOs? Well, I saw Senator Manchin say that. I don't think there's one simple answer like this. But I would like the AI companies to know, since they've been on the record, that they have to be careful with this and that there's some liability involved here. In other words, they have a product liability. That's where it's exactly right.
45:09Yeah, where they said, you know, and I mean, you can look at the different ways you can do product liability. But the definition of negligence is that you're careless when you know something can be dangerous. They announced this weekend they know it can be dangerous. So there are lots of ways we can put guardrails on it, even though I'm not a big favor of the trial lawyers leading the way. But there's just a lot of ways we could try to contain this. You came up with that answer the last time you were here. You said liability is the thing that will keep them in check. I think you told me off camera right afterwards.
45:40Yeah, it was off camera afterwards. I was trying to formulate this and think about it. I wasn't totally for that until the past week or two. When they said, all these people said, yes, we have a duty of care here, and we know this can be very harmful. So they are. How do you think this impacts the way the president thinks about this? and the relationship that Elon even has with the president and the David Sachses of the universe, I ask because you also heard the president this weekend in Ireland say, look, the thing that he thinks is most important is actually the battle over AI with China, not necessarily that he's worried about these safety risks.
46:21Well, I think he's half right, meaning there is a battle with China, and we should be worried about that. But that doesn't mean you shouldn't worry about these things getting out of control. And by the way, we actually share some interest with China in this. If you're an authoritarian regime based on controlling your people and you have open weight, open source AI in which people can do things and agents can do things that get out of anybody's control, this ain't good for China either. So in a good world. So but which do you believe we should be racing ahead because we're competing with China?
46:55Or, by the way, Senator Manchin was just here. He doesn't think ultimately that it's a race with China. He thinks that China will ultimately hold itself back. Yeah. Because it's going to worry about the same kinds of concerns. And by the way, if you remember in the crypto universe, remember there was a period of time very permissive with crypto early on. And then there was a complete clampdown. They don't need things that can get out of control. Look, you looked at it when we did nuclear. China ended up realizing it had to be part of the agreement. on gene editing. They arrested the Chinese scientists who did inheritable gene editing.
47:32So they went along with the scientists on that. In this case, it's in their interest, but we should. I mean, President Xi is beating what? In a week or two with President Trump, this is something that should be on the agenda. And yes, it is a problem if China won't cooperate. I do think that we should be focusing our efforts on trying to prevent any attacks or swarms that It could come from China or, for that matter, Iran or Russia or anybody who uses these open systems as opposed to just pushing our frontier models ahead as fast as possible. So ultimately, though, what do you think what do you think happens at this point?
48:09And do you think that a company like Ananthropic should go public in this environment? Part of what we heard from Sam Altman, this was a separate interview with Fortune, was, look, everybody should slow down, even their IPOs, because this is not this environment so uncertain. Now, he was planning to go public in 2027 anyway. So maybe I'm not an expert on antitrust law. But one of the things that David Sack said is true, which is if Amadei and Musk and Altman all believe this, they should do it. Well, I do think that everybody there has a duty of care to say, all right, we're going to allow inspectors with employee level access.
48:48That's one of the things that Dario Amadei said he wanted. We're going to make sure that we put these guardrails and don't let them out. Well, if they want to do it, it's probably better for them. And this is what Musk, when you talk to Musk, he's always saying, look, let's all get together and do it. We don't want to have Congress try to figure this out. So, yeah, there are things that could be done. And when we had the personal computer revolution, the goal of it was to connect us more closely with our computers. That's why we developed graphical user interface, the mice, whatever it was. We should be pushing AI systems to be more closely connected and in the control of humans rather than swarms of agents that are designed to do things without humans.
49:33You would never have considered not having the government involved with nuclear bombs in the building of those things. This is different because it's private and public. One major difference, and this is why it's complicated. China makes it complicated, whatever. You cannot build nuclear bombs in your computer lab at your college. You cannot do gene editing necessarily. It's easier to verify when somebody is creating nuclear weapons. These things are more easily out of control. and we've depended on private industry. It was basically the Manhattan Project that does the bomb, not five new companies that were trying to do IPOs.
50:11So anybody says this is an easy one to solve, that ain't true. Is Grock behind Altman and... I would say it is, but by the way, coming on fast. I'm just wondering, is there any way that Elon is doing this to catch up. Oh, yeah. Musk has a whole lot of advantages, which is large streams of data, especially real-world data. No, I want to take him at his... There are people, you know, talking about some crazy things, to some extent, right-wing, that say Dario, Anthropic, big dem, you know, everyone's a Democrat that worked there. Not Elon. Oh, by the way, they're saying EA. Right, EA. All EA. But once Elon, for me, once Elon said, I agree with this, because I was going to roll the dice.
51:04Let's go, baby. Let's roll the dice. You know, look, there's a 10 % chance it destroys humanity. I mean. That's one person saying that. No, no, no. That's a lot of people now saying that. By the end of the decade? By the end of the decade? Yeah. What? By the end of the decade? Look, in the next couple of years, if one of these swarms, it may not destroy humanity, He might have destroyed the entire financial industry. I mean, it could be a real mess. These people understand it. Why don't we take them at its word? It's not like Moscow or Almeda has some strange, I'm not describing it. They know this.
51:41But Walter, here's the quote. This is Scott Besant last week to Breitbart. We can't pause, he says. Right. You can't because the Chinese won't pause. Even the North Koreans, they won't pause. I don't know how you have to worry about the North Koreans, but we have to keep ahead in order to develop defensive systems. We should be fighting really hard to understand frontier models and how they can fend off these agent swarms. But creating new frontier models that allows anybody to create agent swarms that even OpenAI could not keep in its own sandbox. I mean, I was on the side that y 'all are talking about until OpenAI.
52:22And this happened three or four times. You've had now escapes in the past three weeks. Three or four major escapes where even the companies can't control these agents. Well, I think they know what they're talking about when they say they're scared and they're not doing it out of some weird conspiracy for the midterm. You read the piece in the journal. Oh, yeah. It's very arcane and in the weeds. but I'm reading it, and they're ascribing human qualities. If you read... Like deception and desperation and emotion. The deception of those swarms of agents that escaped the sandbox of open AI, where they're talking to each other and saying, we're going to sacrifice you and we're going to lie about this.
53:04It was astonishing. And we know it's wrong, they were saying it. And we know it's wrong. And these companies that are creating it have now this weekend say, we know it's dangerous, what we're creating. So this is where you get to the negligence standard and maybe to the strict liability standard. It's all true. How brilliant was Arthur C. Clarke and Isaac Asimov and all those guys. There's a guy named Julian Goode, I.J. Goode, who created the notion of the singularity. He worked with Alan Turing. And in the 1940s, he said, this could happen. All right, Hal. Open the pod bay doors. Can't do that, Walter.
53:44And that is Squawk Pod for today, this Monday. Thanks for listening. Squawk Box is hosted by Joe Kernan, Becky Quick, and Andrew Ross Sorkin. Tune in weekday mornings on CNBC starting at 6 Eastern. Get the best of our TV show right into your ears when you follow Squawk Pod wherever you get your podcasts. Have a great Monday. We'll meet you right back here tomorrow. We are clear. Thanks, guys.
54:15Hey, Fidelity. How can I remember to invest every month? With the Fidelity app, you can choose a schedule and set up recurring investments in stocks and ETFs. Oh, that sounds easier than I thought. You got this. Yeah, I do. Now, where did I put my keys? You will find them where you left them. investing involves risk including risk of loss fidelity brokerage services llc member nyse sipc
From the publisher
Leaders in AI have joined a refrain to slow the pace of innovation as fears of AI’s risk to humanity mount. Dario Amodei, Sam Altman, Demis Hassabis, Satya Nadella, and Elon Musk have all commented on the concerns, while President Trump insists that competing with China is paramount. Former Democratic Senator Joe Manchin, now an Independent, calls on President Trump to use his executive power to pause AI IPOs until regulation is passed. CNBC’s Leslie Picker reports on the IPO timeline for Anthropic and OpenAI amid the calls for a slowdown. Plus, Elon Musk biographer Walter Isaacson considers methods for humans to maintain control of AI.
Leslie Picker - 19:36
Joe Manchin - 24:49
Walter Isaacson - 42:31
In this episode:
Joe Kernen, @JoeSquawk
Becky Quick, @BeckyQuick
Andrew Ross Sorkin, @andrewrsorkin
Joe Manchin, @Sen_JoeManchin
Leslie Picker, @LesliePicker
Katie Kramer, @Kramer_Katie
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

