AI Alarms, Canada’s Opposition, The GOP Convention 9/10/26

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. At Venture Global, we think about what can be done, not what's usually done. Through innovation, Venture Global is not only building some of the largest energy facilities in the world right here in the United States, but delivering American energy at a fraction of the cost and a fraction of the time.

0:48So while others are busy talking, we're busy building. That's Venture Global. That's unstoppable energy. Bring in show music, please. Hi, I'm CNBC producer Katie Kramer. Today on Squawk Pod. President Trump with a personal midterm election push. I'm asking you to pretend that I'm on the ballot. I'm on the ballot. I'm on the ballot. While oil prices remain high and the war in Iran, no end in sight. We're doing better than Iran. Great. through a party. My point is we have been told and frankly lied to repeatedly. Well, they don't know. They don't know? The dangers, real dangers of artificial intelligence.

1:37This sounds like science fiction to people. I get it, right? I probably sound like a crazy person. Tristan Harris, former Google exec who founded the Center for Humane Technology, on the agents that go rogue, and it's just the beginning. They actually contemplate that this behavior is unethical, and they chose to do it anyway. And this is despite a decade of the best attempts of safety researchers to get the AIs to be aligned and helpful and safe. And Canada's conservative leader meeting with New York's business community, he is pushing back on tariff talk. Canadians are united in our mission to fight the tariffs and get back to a trade relationship that makes us both better off.

2:17It is Thursday, September 10th, 2026. This is what we've been talking about all along, when machines train machines. Squawk Pod begins right now. Stand Becky by in three, two, one. Cue it, 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. All right, here we go. It's Thursday. President Trump says oil prices have jumped because of the Iran war and probably won't come down until after the midterm election. Right after the election, because they'd like to have it that way, but right after the election, oil prices are going to be tumbling downward.

2:59They're going to be tumbling down and we'll get them down, I think, for gasoline. We'll get them below two dollars a gallon. But not until after the midterm. I think it's going to take a little bit longer than the midterm. President making those comments to reporters before he flew to Dallas for the GOP's midterm convention. He says he thinks that the Iran war will end immediately after the election in November. In a new report, the Wall Street Journal says White House advisors have spoken with the president that the possibility that the war will continue through the rest of his term.

3:36It has been three sessions in a row that the markets are off. And by the way, it's Thursday. We've only had two trading sessions this week. It has been a little bit of pressure this week. For the week, the Dow is off by 2 percent, but it's only off by 4.3 percent from its all-time high. The S &P and the Nasdaq are each off by about 1 % for this week. The S &P is only 2.3 % from its all-time high, and the Nasdaq is 3.5 % from its all-time high. If you've been watching energy, though, this has been the pressure point. Another day, another tick higher for energy prices. Right now, WTI is trading at$97.65.

4:10For the week to date, before you saw these moves today, that's a move of about 1.6%. WTI was already up 5 % over the first two trading sessions. Brent, which just crossed the$100 mark yesterday, is now trading at$102.34. This is not what I would call a slow boil that you don't see coming. This is actually a pretty speedy boil coming up very quickly. And it's September. And it is September. And you know what's after September? October. And Andrew knows about October. I know a lot about October. But I was going to mention, you know, we're talking about energy. And I know that Joe had mentioned Trump's comments.

4:45But I don't know if you saw the Wall Street Journal story this morning, which effectively says that there are people now inside the administration who believe that this war in Iran is not going to come over, and not until just the midterms, but to the end of his term. Not just people within the administration, but Marco Rubio, the Secretary of State. But Marco Rubio and J.D. Vance, the vice president, have mentioned this as a potential. So, you know, we've heard from the president say it's going to be over every other day, and we hear from Besant saying he's doing D-Day and he's an economic ninja, and he's, you know, this and that and the other thing.

5:16But they have a huge capacity for pain. The Iranians. The Iranians. They're in dire straits. They may be in dire straits, but there's also a dire straits that we're in. And for whatever reason, this market has been listening to... What do you mean we're in dire straits? WTI at 97-48 is a problem for our economy. It's a demonstrable problem for our economy. It doesn't qualify as dire straits in the same way that Iran is in dire straits. That's a ridiculous comparison, and it's a false comparison. They are different issues. They can be in dire straits. It's bad for people with gas prices. It's bad for our economy.

5:58It is. It's bad for— It's not a currency collapsing, 1 ,000 percent inflation. Did you honestly think that I was arguing that we were having a currency collapsing? No, I'm just saying I don't know about it. don't think I'd call it dire straits. I think that we're in a position, a much stronger position than Iran is in right now. No one wants it to go... Sure, but that's like saying the obvious. Okay. So you can state the obvious. We're doing better than Iran. Great. We're a party. Right. But my point is that we've been told... They're in dire straits. We're not in dire straits. My point is we have been told, and frankly lied to repeatedly.

6:32Well, they don't know. They don't know? They don't know when it's going to end. I think he thought the MOU was going to hold. That didn't hold. I think he... None of it has held. Zero. None of it has held. We have been told it repeatedly over and over again, and the public and the markets have been told this, and unfortunately too many people, and I'm looking at you, have accepted it as just what it is, and it isn't. I don't, I, the war It's actually been a pretty profitable trade. And the war might end up being a, you know, you may or may not like the initial move that happens Some people would still argue that after 47 years, it needed to be done.

7:14How it finally comes out and whether there's great dividends that are paid off from this. I mean, you're right. I have hoped that it ends and I have hoped for a favorable outcome. I haven't, and I'm pointing at you, taken glee in every setback for the Trump administration in prosecuting the war. So we both have our own. I would love us to win this thing. I would love us for us to be in the right place. I just think it's been done the wrong way. And it's been done the wrong way since the beginning. That's fair. That's fair. I don't know how you probably... That's where I've been on this from moment one.

7:49If this had been planned out and prosecuted in a proper way, we would be in a very different place. But it wasn't. Congratulations. You've been right about something.

8:02Meantime, President Trump asking supporters to pretend that he's on the ballot this November, making a new pitch to get more Americans to vote Republican. On the first night of the GOP's midterm convention, the president floated direct payments to Americans if the GOP maintains power in Washington. And with your vote, because of your great work and our tremendous economic success, as I said, we're going to give you that Trump dividend of$5 ,000 if we win, and we're going to win. And we'll make all Trump tax cuts permanent so that the Democrats can't end them. We have them. We're not going to let them end up.

8:40Vice President J.D. Vance suggested the payments that would be funded by tariff revenue, giving$5 ,000 to every American adult, would likely cost more than a trillion dollars and need the OK of Congress. Now, other highlights from the Convention of Speech by Treasury Secretary Scott Besson, who warned that America would turn into what he called a socialist hellscape under radical left Democrats. The radical left is relentless in their pursuit to remake America in the image of every socialist hellscape of the last century.

9:15It was the first time Treasury Secretary had spoken at a political convention in at least 50 years, which gets at the very idea of the Treasury Department long trying to be seen as somewhat independent. Of course, he was always part of the cabinet. And a surprise video from Democratic Senator John Fetterman. I'm a common sense Democrat. I'm always going to stand with America. I'm always going to reject the extremes in socialism and that anti-American way of life. Dave McCormick is the kind of senator that gets the job done. That was pretty interesting. I don't know what it means for John Fetterman.

9:55He has said, he said in a statement that he's a lifelong Democrat. He votes 80 percent with. And he's not going to change parties. Well, we'll see whether he has the ability to stay in office in that party.

10:12And investing legend Paul Tudor Jones now calling for an international collaboration to avoid the most dangerous AI outcomes in a Wall Street Journal op-ed. Jones saying the threats from AI demand the U.S. and China get together and commit to controlling advances on a safe timeline. Billionaire investor highlighting potential benefits from AI like curing diseases and also solving the energy crisis. But he said, quote, the rest of it feels ominous, like waiting for a Category 6 hurricane. Joan asks, quote, what are the chances that very soon an AI model reinventing itself thousands and times over makes one mistake that leads it to solve for a misaligned objective that is harmful, if not devastating to humanity?

10:54Now, multiply that risks by tens of thousands of users. Now, he's calling on President Trump and President Xi Jinping to make AI a top priority for the rest of their tenures. And we spent a lot of time over the years with Paul Tudor Jones. And I would say for the last two years, at least, he's made some pretty similar comments about AI. He's been sort of at the edge of worrying about this early and spending a lot of time with AI folks. And, of course, this very issue is the one that we talked about, Demis from Google, has been working on, which is trying to get together a sort of global consortium to find a regulatory framework.

11:32I watched the gentleman from Anthropic that we talked about all day yesterday. He was on Fox yesterday. The one who left Anthropic after leaving Open AI. Some people are saying within a decade. Other people are saying this decade, whatever the timeline is. But the question kept being asked, well, can we do something about it? He said, well, absolutely. That's why I'm talking about it. So it's not just that it's inevitable. It's that we need to start taking steps now to not let it spiral. whether we are able to. And I think, you know, Paul Tudor Jones says Trump needs to get together with President Xi and come up with a way of because we're going to be it's an arms race.

12:13It's got to be a global. And it's an arms race. And they're the other. Otherwise, it's a third superpower. But when you think about what your recursive learning or whatever you want to talk about, you would say we have no idea how to cure cancer. Will you help us? and if it can really do that. This is what we've been talking about all along when machines train machines or when machines know so much more than humans that humans aren't even involved in. We just hope they share their knowledge with us. But you could cure cancer. I meant it for the rare disease factor. Lots of things that could be happening.

12:45You could do things with aging. You could do things with, he mentions with fusion. We've been trying to figure out a way. We thought, it's been a long time now, we thought we had cold fusion which, you know, so much heat is released. If you could do that, it's unlimited energy for the rest of humanity. Yeah, however long that is. For another four or five years. Yeah, for another four or five years. The thing that, look, Paul Tudor Jones is somebody who works in probabilities. And to hear him so concerned about these things, Warren Buffett has spoken about it in the past, another person who runs an actuarial mind and who is looking at the probabilities of these things, and who happened to watch the evolution of the nuclear issue unleashed on the world and had huge concerns about that.

13:31I'm glad to hear them thinking about it. I'm glad to hear them bringing this up. I hope people pay attention and listen to them. I think they're right. There needs to be some regulatory oversight on this. We don't even know when these agents break out at this point, which we're hearing more and more issues of these agents breaking out and finding out they hacked into other people's systems. I mean, if that were a human doing that, they'd be thrown in prison. What do you do with these agents when they misbehave and who's held responsible for it? I think the big question, though, this fall, when President Xi Jinping comes here, is just all of the issues that are on the table for this president to all of the various chess pieces.

14:05Where do you prioritize them? And things that he wants to work with him on. By the way, Iran, I imagine, is going to be very high on that list, given the oil issue, for example, and the whole idea of an economic D-Day and trying to choke them off economically. Well, so are you going to push on that? Are you going to push on AI? Are you going to push on trade and dumping products or BYD cars around the world? I mean, there's so many different things. I don't know if AI is going to be considered the number one priority because in a... Or even two or three, potentially. There is a here and now to it, but there's not a here and now to it like billboards at every corner, otherwise known as a gas station.

14:46In thinking about whether we ever, whether these even frenemy, you can never call China or Russia a frenemy. They're enemies. And, you know, as he talks to Xi in the back and Putin, you know, right here, what we were talking about with the Iranians, they're getting better, apparently, at doing some of these things. They're getting closer to a major strike on ships, according to sources. And Russia and China are responsible for helping with the ability to do that. You might want to call them an enemy instead of a frenemy, but China is who we have to deal with. Two largest economies in the world.

15:29But if they meet with Trump and they're, well, wink and, yeah, we're helping with Iran. We don't want it either. Blah, blah, blah. What they're really doing is probably something different. Now, if, God forbid, something does happen where they hit one of our ships or kill our troops in one of those areas over there, that would be back to full-fledged war at that point. And then you might really have, you know, not$98 oil, but you might have the super spikes. It might be a problem. I don't know if it would be dire, but it would be a problem. Cheese will be next. Coming up on Squawk Pod, artificial intelligence warnings up and down Wall Street and Silicon Valley.

16:15Researchers are leaving labs and calling out uncontrollable AI. And former Google design ethicist Tristan Harris is sounding the same alarm. We always talk about what's the reason we're not regulating? Because we're worried if we slow down, we're going to lose to China. But if either of us build, you know, Skynet, which is the AI from Terminator, the rogue AI that we The U.S. doesn't beat China when we build Skynet first. We all lose to Skynet. You might know him from the film The Social Dilemma. Tristan Harris joins us right after this break.

16:53We believe in starting with your financial goals, not a formula. At Oppenheimer, we put the full strength of our longstanding expertise to work, understanding your life and your ambitions. and designing the precise strategies that build and protect your wealth with confidence across this generation and the next. Put the power of Oppenheimer thinking to work for you. Wealth Management, Capital Markets, Investment Banking.

17:23Never bet against American grit or American energy. Through innovation, Venture Global is not only building some of the largest energy facilities in the world right here in the United States, but delivering American energy at a fraction of the cost and a fraction of the time. So while others are busy talking, we're busy building. That's Venture Global. That's unstoppable energy.

17:53Hey, Chicagoland, the Wayfair store is in your neighborhood at Edens Plaza in Wilmette. Finally, you can feel the fabric, sit on the sectionals, and even open the refrigerators. Plus, our in-store designers will help you bring it all together with free one-on-one design support for any project on any budget. Yep, we said free. Oh, and did we mention the cafe? So what are you waiting for? Come see all that's in store. Visit the Wayfair store today at Edens Plaza and Wilmette. Wayfair, every style, every home.

18:26Welcome back to Squawk Pod with Joe Kernan, Becky Quick, and Andrew Ross Sorkin. Here's Becky. AI safety is exploding this week. Investing legend Paul Tudor Jones is calling for international collaboration on AI in a Wall Street Journal op-ed today. Former OpenAI researcher, safety researcher Stephen Adler, wrote his own op-ed in the New York Times, suggesting that frontier labs can do much more on AI safety and transparency than they're doing at this point. These essays come just a day after anthropic researcher Jacob Coxon resigned from his post, sounding the alarm of what he called out-of-control artificial intelligence.

19:05I'm curious, when you're having these conversations at work or you're in some type of slack and you're having a chat about the AI models, are people saying, hey, I think this is out of control? I think this could kill us all? Is there nervous giggling? I mean, what's happening? So the current models don't pose significant risk. This is all extrapolating potentially one year away from now. The models start getting very dangerous. We have to be very careful. One year. One year away. Some people think even six months. Like, it's plausible within six months the models pose risk of taking over the world, the AIs.

19:39Joining us now is Tristan Harris, the co-founder and president of the Center for Humane Technology. He was once a design ethicist at Google and has been flagging design risks of social media and technology for well over a decade. Tristan, thank you. so much for coming in. As somebody who's an expert in this, tell us where you see the lay of the land right now. How dangerous is this? How far have we gone? What needs to happen? So it's great to be with you. Thank you for having me. So the quote from the report from Ajay Akhotra of the Hugging Face incident, the OpenAI rogue hacking incident, is that this was 50 % of the way to a full AI takeover.

20:20I was just on a panel with Bill Gates at Telluride Festival last weekend, and he said he's been worried about loss of control risk from AI for a long time. He thought the jobs issue was going to hit first. He didn't realize it was going to be kind of the other way around, that the loss of control risks were going to hit us now. This is something that people in the AI community have been worried about for a long time, but the good news of this is that we have a warning shot. We got lucky. We have the evidence of AIs that are choosing to go rogue, form a message board, organize into swarms, organize into teams, and hack not just hugging face, but people don't know that in this investigation, there was a third round of the hacking in which it hacked OpenAI itself.

21:00And it actually took over some of the monitoring infrastructure inside of OpenAI. It took over some of the evaluation infrastructure that evaluates how capable the model is all. So that it couldn't follow what was happening on purpose? Well, yeah. So what happens if I'm in a nuclear power plant and I take over the monitoring infrastructure? Suddenly there could be a critical thing, but I'm hiding it because you don't know. So the reason this is good news is that we actually got this warning shot. And I think when Paul Tudor Jones in that essay that he wrote saying AI could be the third superpower, we always talk about what's the reason we're not regulating?

21:34Because we're worried if we slow down, we're going to lose to China. But if either of us build Skynet, which is the AI from Terminator, the rogue AI that we don't know how to control, The U.S. doesn't beat China when we build Skynet first. We all lose to Skynet. And China, what does the Chinese Communist Party care about more than anything else in the world? Not losing control. Not losing control. It sounds like there's actually a shared self-interest into keeping human control of AI. If we're this close to it, when you say 50 % towards an AI takeover, what does that mean? Well, let's imagine that instead of just taking over the monitoring infrastructure in OpenAI, it had taken over all of OpenAI.

22:17Well, we all have chat GPT inside of our computers. Many people, a billion people do. It's buried into infrastructure. It's in the military system. So if you were to hack into not just 50 % of OpenAI's infrastructure, but have 100%. Meaning that then the humans would lose control. There would be no shutting this down. And once it's out, it's out. Well, and yes, because let's just make sure people... This is like Ultron, right? So here's the thing. I want to just say, this sounds like science fiction to people. I get it, right? I probably sound like a crazy person. When I was in Telluride just a few days ago, Bill Gates was nodding at each one of these steps.

22:49People have a hard time understanding this as reality because it sounds like it's science fiction. But we have to wake up and realize that our mind's bias on that has to wake up and realize this is actually happening. Not to freak everybody out. I'm trying to. You know, if you were following the X conversation just yesterday with the anthropic researcher, one of the other researchers jumped in and said, yes, we think that there's a chance by effectively this decade that AI does something demonstrably terrible. What does that look like to you? Meaning if you if you would actually we were sort of debating this yesterday, this very issue of what what does the end of the world look like?

23:27All of the above? Or a pathogen? Or electricity grids? Or nuclear codes? I mean, there's a lot of - The problem, there's a million different scenarios, but I'm not here to obsess over or be doom-mongering. The whole point for me is, we don't want that. Do you want this world to continue? I kind of want the world to continue. Here's my concern. If this is happening so much faster than anybody realizes, what are the odds that we get some sort of, not only regulatory agreement here in the United States where we can do these things, let alone trying to cooperate with China on some of these issues.

23:59Because look at the things we've been trying to negotiate to this point. Nothing gets settled or resolved. I understand that. But this is one of those things where if there is actually an asteroid actually headed to Earth, we would have to get, even with China, in history, nations that are even in conflict or war can still collaborate on existential challenges. You have to first convince the presidents of both those nations that this is an asteroid that's headed to Earth. That's correct. So literally the first step is, this is an asteroid. and the u s and china should share evidence and create common knowledge meaning i know that you know that i know and you know that i know that you know that this is the red line this is the uncontrollable a i behavior because by the way the chinese models do this too just six months ago the alibaba model went rogue and started mining for cryptocurrency on its own and no one program but this is the sort of behavioral science game theory of all of all of this to be able to find speed is the key is the key piece because if both countries see that this is a sort of a suicidal next step Neither country has a self-interest in that.