Pete Hegseth Says ‘No Beardos, No Weirdos’ / Donald Trump’s AI Safety Accord

0:00Your financial life rarely fits between 9 and 5, and the questions shaping your future don't follow a schedule. That's why Citi created Citi Sky, a secure AI member of your Citi Wealth team, built with the Gemini platform on Google Cloud. Whether expanding your portfolio or planning your future, you'll be able to access tailored wealth guidance on your time. Start building with Gemini at cloud.google.com. Citi Wealth investment products and services are provided by Citigroup Global Markets, Inc., member FINRA and SIPC.

0:32From the opinion pages of The Wall Street Journal, this is Potomac Watch. Defense Secretary Pete Hexeth delivers a notable speech on the state of the United States military, saying deterrence is back, but quote, beardos and weirdos are no longer welcome. Plus, President Trump gathers tech executives in Washington to sign a new accord on AI safety. Welcome. I'm Kyle Peterson with The Wall Street Journal. We're joined today by my colleagues on the WSJ's opinion pages, editorial board member Kate O'Dell and columnist Alicia Finley. Secretary of War Pete Hegseth has been on the job for about 20 months now and made news Wednesday in Quantico, addressing his and President Trump's priorities for the Pentagon.

1:21Let's listen to a piece of what he said. Now, for almost two years, you've heard me talk over and over and over about three core priorities, reviving the warrior ethos, rebuilding the military and reestablishing deterrence. And I'm proud to stand on this stage and report that the warrior ethos has been revived. Our military is being rebuilt and American deterrence is back. Kate, how do you grade those priorities from Pete Hegseth and the Trump administration's management of the Pentagon, particularly on rebuilding the military? I thought President Trump says he already did that in the first term.

2:03Well, look, I think Hegseth there outlined good guiding priorities. I think as we go through his speech, we can grade them sort of individually, but start with something that he announced on Wednesday, which was he wants to establish a new four-star command for autonomous weapons and drones. This is an example of a good substantive priority for the Pentagon. Hegseth was also really exactly right when he said that what we've learned in Ukraine is not that we need this kind of drone or that kind of drone, but we need to be able to adapt on the battlefield. In Ukraine, the life cycle of a drone, you know, it's a couple of weeks before the Russians have developed a countermeasure.

2:47Ukrainians have to adapt. And so the question for the U.S. military isn't really, you know, what do we purchase now? is how do we create this environment where we have our, you know, software and engineering minds really in the front and able to help adapt on the battlefield. So that, I think, is a welcome elevation of an important priority. They want to set up this four-star command and have it be inoperative by October of 2027, which is fast in the world of military affairs. And I think they have some support for this concept in Congress. The Senate included a similar concept in its defense authorization this year.

3:26So that was really the headline for me in the rebuilding the military category was this initiative to try to adapt to the changing state of warfare as we're learning it from both Ukraine and Iran. What about the warrior ethos, something that Pete Hegseth has talked a lot about. It's true that news stories seem to suggest a rebound in enlistment in the services. How much do you credit the change in tone at the top, the turn away from a kind of DEI messaging under President Biden toward a lot of talk about patriotism, about excellence, about being the finest fighting force on planet Earth? I think if you look at the public opinion polling during the Biden administration, you do see a drop in the public's confidence in the U.S.

4:18military. And I think the Trump administration did have a mandate to restore the military to a meritocracy and to protect that. That's a very popular concept with the public who wants to think that officers are getting promoted based on their merits and not their race, sex or other features of their identity. Now, at the same time, I think there is some loss of confidence among the public in this particular project. I think Hegseth's firing of various general officers, sometimes justified, sometimes not explained, has left the public wondering why. Why, for instance, Randy George, the chief of the staff of the army, why was he fired?

4:56And then also it's taken on, just gone into rabbit holes like mandatory testosterone testing for the force. And I think it loses credibility with the public in a way that I think is a risk, because for this to last, for what Hexeth is doing here to be durable, you need Congress to support it. You need the public to say, hey, we don't want to ever go back to what the Biden administration was doing. And so I do have some concern that some of the particulars here and just the polarizing, aggressive way in which is being pursued may make it easier for Democrats to just get back to their DEI project as soon as they retake power inevitably in Washington.

5:31Hang tight. We'll be right back in a moment.

6:01on Google Cloud. Check response. Subscription required. This is Matthew Hennessey, the editor of Free Expression. I want to personally invite you to come check out what we're doing. Signing up for the newsletter couldn't be easier. Just go to our homepage. Directly below the Wall Street Journal logo, you'll find the magic words, Free Expression, wedged between opinion and arts. There's an option to sign up for the newsletter right at the top of our section. Just hit that button and you'll start getting the world's hardest working newsletter delivered directly to your inbox Monday through Friday.

6:36If you aren't a journal subscriber, the newsletter is free, but we are gonna need your email address. There's a lot of fun stuff in the newsletter that you can only read in the newsletter, and I don't want you to miss any of it. Sign up for free expression today. We'll see you in your inbox.

6:54Welcome back. Let's listen to another section of Hague Seth's speech. This is the one that got a lot of attention, a lot of headlines. We are no longer the woke department or the weak department. Simple translation of that. No fatties, no trannies, no beardos, no weirdos, no wimps, no radicals. Just warriors. Just tough, ready, committed troops. The War Department, as you know, is colorblind, gender neutral and merit based. Or as the president likes to say, Pete, central casting. I want central casting. The ideological clowns are out. The patriotic cowboys are in with testosterone testing on top.

7:46Kate, if we are criticizing the Biden administration for narrow casting in one direction in some of its messaging, maybe this is the same mistake in the opposite direction. I mean, I think talk of patriotism and excellence by the U.S. military is broadcasting. But to your point, mandatory testosterone testing, I'm not sure what the public would make of that. No beardos like, you know, William Tecumseh Sherman or the vice president, Vice President J.D. Vance. I mean, what are we supposed to make of this? Again, it's unfortunate because, you know, that middle section of the speech we watched, just listen to where he says the military is colorblind and gender neutral.

8:32Again, these are popular things. And, you know, Hegstatt has done elevated standards in a way that I think are appropriate. Even grooming standards for units are part of showing individual unit discipline, I think on combat standards, I think everyone in the military should meet a basic physical fitness standard to remind them the nature of the organization that they joined. I think also Hegseth has done important work in making the combat physical standards gender neutral, because I think the Pentagon is right that, you know, they use the line, you know, combat doesn't care. The combat environment is unforgiving.

9:05It does not discriminate. You know, it only cares whether you're capable of doing what it asks. And so this high gender neutral standards, extremely popular stuff with the public. And so I think what is unfortunate and just regrettable about the speech is that it does have this element of kind of deliberately provoking people who might disagree as opposed to trying to aim to broaden the coalition for these popular initiatives. On the point about beards, though, I mean, I get that there are changing cultural standards. The mustaches, particularly of the 18th century military, are no longer in style.

9:38But if there's someone whose job might put them in a position where they would have to quickly put on a gas mask, that's one thing. Well, first of all, Kyle, there are still a lot of mustaches in the U.S. military, particularly in naval aviation. So I just want you to know that that tradition very much lives on. I mean, look, I think the larger problem here is not really about individual unit standards, which I do think that, to your point, sometimes there are reasons for those to exist. I think that grooming standards can bear on a unit's discipline. And part of being in the military is submitting yourself to a hierarchy and following directions on things that don't always make sense and might not be what you would choose.

10:14But I think the larger issue here, to your point, is this alienating list of things does preclude a lot of human ingenuity that we're going to need to defeat or deter enemies that the United States is facing. And so, you know, in today's editorial, I mentioned, you know, like, look, no beardos and no weirdos. That's going to rule out a lot of people, for instance, in the defense technology sector. We are going to need a lot more technical talent around the U.S. military. We're going to need a lot more cyber talent around the U.S. military. That will include some weirdos and weirdos. But I don't think that's what Hegseth is really talking about here.

10:45I think this is kind of, like I said, a hallelujah chorus line to try to get just gin up the support for kind of his culture war approach to the Pentagon. And on that, you know, shifting our turning our ship here a little bit abruptly. But I mean, one of the arresting things about this speech to me was things that were not in it. You know, it was really geared toward domestic audience. And there is a tremendous merit into getting the public to see that the military is an institution they can trust. And that is essential for getting the military, the budget, the weapons that it needs. At the same time, a speech called the state of the force, you would expect that to include things like, you know, are we getting after our shipbuilding crisis in the U.S.

11:29Navy? Have we improved our logistics and ability to commute to a long fight all the way in the Pacific? Are we exercising more than we ever have been, our steaming days and flying hours up? These are kind of essential force readiness measures that were not really discussed at all in the speech. And I think the reason for that is some of them are concerning, that we have not actually improved our basic readiness or force posture in the first few years in office. I mean, only once in this speech, as I'm remembering right, does Hexeth mention the$1.5 trillion budget that the president proposed for the Pentagon this year that looks like it may have problems passing Congress.

12:07And so the focus on the culture war issues are as notable as just kind of the basic state of the force elements that were left out. And, you know, the point of the force is to deter and win the nation's wars. And it was striking to me that that was not a major theme of the speech. Hang tight. We'll be right back after one more break. This episode is brought to you by Charles Schwab. Decisions made in Washington can affect your portfolio every day. But what policy changes should investors be watching? Washington Wise is an original podcast from Charles Schwab that unpacks the stories making news in Washington right now and how they may affect your finances and portfolio.

12:47Listen at schwab.com slash Washington Wise.

12:53Don't forget, you can reach the latest episode of Potomac Watch anytime. Just ask your smart speaker. Play the Opinion Potomac Watch podcast. From the Opinion pages of The Wall Street Journal, this is Potomac Watch. Welcome back. Let's turn to the White House Accord on Superintelligence, signed this week by executives from Google, Anthropic, Meta, OpenAI, XAI, and NVIDIA. First, here is President Trump with some of those executives explaining what he hopes it will do. At the top, I'd say it looked like a group of about 40, but whatever it was, here's one of them right here, Jensen and NVIDIA. and they love our country, they love the world, and they came together and they did something that few people really, it's almost like a constitution in a way.

13:47And the biggest people in the world signed that, and I signed it as president, and it really is a form of protection. Is this a court that everybody signed? Is it binding in any way? I think it's morally binding, yeah. Alicia, what do you make of this effort? Well, I think what he's trying to do is, one, protect himself and, for that matter, some of the executives from the criticism that they've been facing that AI isn't being regulated, that it's unsafe, that it's going to cause a billion deaths, and show that they actually are taking these threats seriously. That said, I think the president has the right instinct that self-policing is probably the best answer to this problem or the multivarious or various problems that the critics or maybe if you want to call them doomers have raised in that, yes, there are very serious potential.

14:49you know maybe not catastrophic but there is the potential for havoc or you know cyber attacks and such especially if these tools get into the wrong hands and that could be including you know Chinese the Chinese are developing open weight models and those could be used or be proliferated across the world where you could have terrorists or other adversaries, hackers, use these models to attack the U.S. And so merely calling for a pause or kill switches or whatever regulation have been proposed really wouldn't address the biggest concern, which is that Chinese will continue to advance their models and aren't going to impose safeguards and could actually use these models to attack the U.S.?

15:43The Accord document released by the White House calls for companies developing AI to include four layers of safeguards described with some level of generality. So here's one of them, robust internal controls to monitor AI. Another one is partnering with independent auditors or external evaluators. And then the document says this, together these steps will give each company its customers and the public confidence that the technology is operating as intended. The participating companies will meet regularly to establish standards and best practices to improve the safety of their systems. To me, that's interesting, Alicia, because it suggests maybe a multi-company approach.

16:28I mean, one imagines that some of the things that are described in this document, internal controls and working with external evaluators, are things that companies like OpenAI and Anthropik are already doing. I think we know that from some of the reporting. But the idea that the industry wants to work together on strengthening those safeguards and then to your point, maybe strengthening public confidence in the idea that the industry is not ignoring the public concern. I think that's exactly right. And that's exactly the goal. Most of these, at least the leading developers, the frontier labs, are actually doing most of what's already in this compact.

17:10They do have internal safety review teams. They are using external third-party, you can call them auditors, review teams. That's actually how the incident over the summer happened in which a model or OpenAI's model was found to have hacked into the company's hugging face. That actually happened during an external safety testing, and the OpenAI didn't put enough controls, internal controls on its model, but it was also the result of the external auditor or reviewer also didn't configure its own testing, quote-unquote, sandbox correctly, which enabled this to happen. And actually, that same company, that same auditor also had issues when it had tested other companies' models, including Anthropics.

17:56This was not an actually isolated incident, which also just underscores that merely having an external safety auditor won't necessarily prevent some of the problems. And to an extent, it could provide a false sense of security if these companies don't do enough of due diligence on their own. I think that's also the broader problem with regulation, regulating these models is that it becomes just a compliance checklist that you're just meeting what the government says. You do essentially what the government says, and then you kind of overlook just some kind of basic safety safeguards. I mean, the case in point of this in a different realm was, you know, the Silicon Valley bank failure, which if you read some of the reports, it was very concerned about procedural and compliance checkboxes and wasn't really looking out after managing safety and soundness and interest rate risk, which was ultimately what caused it to collapse.

18:59And so I think there is a concern that when companies become too focused on just complying with regulations that they miss and overlook some really safety blind spots. Democrats are criticizing this White House accord as not going far enough, as being voluntary and morally binding instead of something that's mandatory. But Alicia, what is the alternative proposal or is there one that they have even rallied around? I mean, is the idea that these AI companies doing frontier work in this new realm of artificial intelligence are supposed to go into the Commerce Department and show some government bureaucrats what they're doing and get approval or sign off?

19:39Or what is the alternative? I mean, anything like that that slows down U.S. progress in this area, the worry would be that it would become a strategic vulnerability if U.S. AI models fall behind the Chinese ones, maybe are not able to scrub our systems here for vulnerabilities that a Chinese AI might exploit, you know, an hour before Xi Jinping gives the order to move on Taiwan. Right. So there are various proposals that have been kind of bandied about in Congress and by the AI labs, you know, self-regulatory organization kind of either modeled after FINRA, which oversees brokers in the financial space.

20:20Some have proposed that a separate agency, federal agency, created to review these models, which, again, I'm a little skeptic about just because we've seen time and again that regulatory agencies can be captured by the industry or large players. But there's also the concern that the quote-unquote experts at these agencies just aren't as sophisticated as the people who are even developing the models. I mean, again, look at the financial space where the investment bankers typically know much more than the bank supervisors and examiners who are in charge of regulating them and reviewing everything that they do.

21:04And then you have some proposals kind of further and more radical on the left, you know, that would halt AI model development and data center construction entirely until these AI labs meet a long list of checkboxes, such as, you know, they ensure that their models or the data centers do not contribute to climate change, do not displace jobs. jobs essentially present no risk to humanity, which all these requirements, you know, these companies will never be able to meet. And that's the whole goal is they want to pause AI development. One, this is somewhat ideological, and they do view AI as a threat to jobs.

21:53And they also don't like the fact that a lot of these data centers are going to be running and going to be powered by fossil fuels. So they view by stopping essentially AI from advancing and data centers from being built, they hope to essentially reduce carbon emissions. The other piece of this debate that sometimes is missing, I think, is the U.S. government already has all sorts of broad regulatory authorities that might be able to come into play here, even setting aside the idea of the liability, the legal liability that these companies are under, which I would not underrate as an incentive to them.

22:32If one of these frontier models gets out and causes real damage to another company or the public, I think that they would be held liable for that, which gives them a real reason to focus on the safety and the alignment of anything that they are going to let loose out of the sandbox. There's also the Federal Trade Commission, Alicia, which has announced that it is now doing an investigation of frontier AI, looking into potential unfair or deceptive acts. And then don't forget that recently, Anthropic said that it was going to suspend access to the Fable 5 model. The U.S. government had put an export control directive on that, And Anthropix said that the only way for it to comply was to suspend it for everyone.

23:19And so it does not seem as if the U.S. government lacks for ways to make itself known if it has some serious concerns about what these AI companies are doing. Right. And actually, my concern would be that over the federal government overreaches in some of these contexts. I don't know really if the FTC should be right now doing a sweeping probe of all these models for alleged unfair business practices and consumer deception. And who knows how these regulators in a different administration will use these powers in order to essentially, and they could use these powers to pause and impose a de facto pause on AI development.

24:01I mean, we saw during the Biden years how the regulators, or at least government officials, really tried to put the squeeze on social media platforms to essentially censor speech that they didn't like. And they kind of floated around the various threats, antitrust action, repealing or limiting their protection under Section 230 in order to essentially coerce those platforms. And I think that that's what is likely to happen with the AI companies is that the government will use these threats. And you mentioned one of them with the Fable 5 model, the other threats to essentially get the companies to do what it wants.

24:43I mean, the Trump administration also declared Anthropik a supply chain risk. And you can debate whether that was credible or if that was just a political vendetta, as some say, because some folks in the Trump administration didn't like Dario Modi's liberal and progressive advocacy and his support for Kamala Harris. But all this is to just say that the government has a lot of authority and a lot of power already to essentially regulate and punish these companies if it wants to. Thank you, Alicia and Kate. Thank you all for listening. You can email us at pwpodcast.wsj.com. If you like the show, please hit that subscribe button.

25:27And we'll be back next week with another edition of Potomac Watch. Why can't we get AI to production? Our competitors are all ready. How do we keep data secure? AWS AI cuts through the noise. With ready-to-use agents and the broadest set of AI tools, stop overthinking, start building. AWS AI is how.