AI-Made Viruses, A July Jobs Report, & Meta’s Child Harms Case 8/7/26

7 Aug 2026 · 35 min · 10 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

The episode links AI to labor and economic disruption, AI-enabled biological threats, and current business/policy news. It covers a weak July jobs report (non-farm payrolls -23,000; participation rate 61.4; unemployment 4.1; U6 7.9), Iran’s draft plan to block U.S./Israeli ships through the Strait of Hormuz, Airbnb’s AI-driven results, Meta’s New Mexico child-harms ruling, and OpenAI’s upcoming donut-shaped smart speaker.

Guests

Rep. Greg Kazar (Texas; chairs the House Progressive Caucus; co-author of the AI Tax and Work Protection Act). MIT evolutionary engineer Kevin Esvelt (bioscience defense analyst).

Key claims

Kazar argues for an AI tax that discourages job-automation and scales up funding for new jobs if unemployment spikes, while not taxing “productive” AI like medical research. Esvelt says AI can design bacteriophages that work (potential antibiotic alternative) but warns future work could extend to viruses infecting other organisms, enabling harmful variants.

Notable examples

Block layoffs citing AI; Airbnb CEO Brian Chesky credits AI; Meta ordered to fund a teen mental-health program and limit minors’ access; Esvelt references AI trained on ~15,000 bacteriophage sequences and “downstream” misuse risk.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

Discussing Job Losses and AI's Role

2:25 to 3:51

An analysis of the unexpected job losses in July and the implications of AI on employment.

“Good morning and welcome to Squawk Box here on CNBC, live from the Nasdaq market site in Times Square.”

AI and Potential Biological Threats

3:51 to 9:15

Exploring the dangers of AI in creating biological weapons and its ethical implications.

“Well, you have to feel for these people.”

Airbnb's AI Integration

9:15 to 11:05

How Airbnb is leveraging AI for business growth and the future of travel.

“It hasn't, I mean, to my knowledge, and it could be happening, I haven't seen that yet.”

Meta's Legal Challenges

11:05 to 11:43

Meta's recent court ruling and the implications for children's mental health.

“The munchkins don't really come out of the bagel.”

Upcoming Smart Speaker from OpenAI

11:43 to 14:00

Insights into OpenAI's new smart speaker and its potential market impact.

“No, I don't want any device listening to me.”

Congressman Kuzar on AI Tax and Worker Protection

14:00 to 23:17

Learn about the proposed AI Tax and its implications for future job markets.

“Joining us now with the details, Congressman Greg Kuzar.”

Congressman Kuzar on AI Tax and Worker Protection

23:22 to 24:04

Learn about the proposed AI Tax and its implications for future job markets.

“With Progressive Boat Insurance, you can choose coverage for most mistakes you or other boaters could make, helping you float carefree all season long.”

July Jobs Report and Economic Discussion

24:44 to 28:00

Explore the implications of the July jobs report and economic forecasts.

“We're just seconds away from the employment report for July.”

Understanding the Underemployment Rate

28:00 to 31:58

Learn about the latest underemployment rates and their implications.

“The last part of this report is called the underemployment rate or U6.”

AI and Virus Genome Innovation

31:58 to 38:04

Explore how AI is creating new virus genomes and its potential impacts.

“I mean, the headline sounds nefarious, But this is actually a good development in terms of the fight against drug resistant bacteria, correct?”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00Trading at Schwab is now powered by Ameritrade, bringing you an expanding library of education with even more ways to sharpen your trading skills. Access new online courses, insightful webcasts, articles, engaging videos, and more. All curated just for traders. Plus, guided learning paths with content designed to fit your unique interests. And no sifting to find exactly what you need, so you can spend your time learning to trade brilliantly. Learn more at schwab.com slash trading. Insurance isn't one size fits all. That's why drivers have trusted Progressive's Name Your Price tool for years. Just tell Progressive what you want to pay, and they'll show you coverage options that fit your budget.

0:38Visit Progressive.com to find a car insurance rate that works for you. Progressive Casualty Insurance Company and Affiliates. Price and coverage match limited by state law. Bring in show music, please. Hi, I'm CNBC producer Katie Kramer. Today on Squawk Pod, can we blame AI for everything? The U.S. economy unexpectedly lost 23 ,000 jobs in July, with a labor participation rate the lowest in five years. Buckle up on this one. Minus, minus 23 ,000 on non-farm payrolls. I can hear the bond yields falling right now. Taxing AI companies that cut jobs to create a new era works progress administration.

1:24Congressman Greg Cazar of Texas joins us.

1:26Greg Casar:I'm not talking about just stopping all of AI in its tracks. I'm saying we should lead in AI, but that doesn't mean we have to be in a country with large amounts of unemployment. And AI created viruses. Haven't we been through enough? A new study on just how willing artificial intelligence is to create biological weapons. MIT scientist Kevin Esvelt. There is no reason to think that future researchers cannot and will not attempt the same thing for viruses that infect other organisms. Probably not humans at first. That would be too obvious. Plus, the rest of today's news, the clogged Strait of Hormuz, Airbnb's rise, and OpenAI's new smart speaker will be shaped like a donut?

2:11Spherical, a color, a twist, a long john, a bear claw. Novelty shapes like hearts, stars, seasonal outlines. It's Friday, August 7th, 2026. Squawk Pod begins right now. Stand by Joe in three, two, one. Here's Mike. Good morning and welcome to Squawk Box here on CNBC, live from the Nasdaq market site in Times Square. I can't help it. I can try to not smile, but it's impossible. I'm Joe Kernan, along with Melissa Lee. There it comes again, Becky and Andrew are off today. Why does that happen? What? When certain people are gone, you're smiling? Well, now you're actually on with, no. I was doing it in a positive way because you're here.

3:00Thank you. You know that. I'm happy to be here. I think it, is it years of this that we've shared together? Is that what it is? It feels like it. And I was talking about, you have people on all the time on your multiple shows. Not all of them at the beginning of the year were at 7 ,700 or 7 ,800 or 8 ,000. No. Somewhere in the sixes. A lot of them, my favorites, because they're sell-side people, they really aren't able to make very good calls, and they don't even try. What was your call at the beginning of the year? A lot has changed during the year that nobody could have guessed in January or February.

3:39I don't, you know, luckily we just get to second-guess other people. Exactly. Which is good. However, I don't know if I would just wear the market clothes and say that's my target until the S &P goes up 100 and then I raise my target. Well, you have to feel for these people. My point is you have to feel for these people. They make money. They tell their poor retail sales force that they pretend they actually know something. They don't. It's a no-win job. That's for sure. You know who you does know are buy side people because they they look right at their what they have and they go, oh, God, why did I buy that?

4:14They know they bought a year and a half ago. They know they bought something like some of your traders. I think they feel because they're able to change. They actually it's not according. I use this term a lot. I have determined and luxury beliefs. Everybody has a luxury belief that doesn't cost me. You can be negative on Bitcoin from from all the way up. And you can still say, and it doesn't cost you a penny to be negative. Then it finally comes down and you start pretending that you called it coming down. Are there some predictions? Which predictions do you want from me? I don't know. Well, I think it has more to do with AI.

4:53I think it has to do with deregulation. I think it has to do with the big, beautiful bill and expensing. I think it has to do with, this is heretical, but onshore. And some of the, there have been some reindustrialization. And obviously AI on top of all those things. It's huge. And I think that, and I said it about a week and a half ago, that when we had had the pullback in AI and we started to, oh, okay, this is the beginning of the unraveling. a lot of people thought with that hedge fund. After it comes back a little and we're at new highs again, that's the hardest time to buy. The easiest thing to do at that point, and that's what all the sell-side people do, would be, you know, oh, I think we need...

5:38And to buy things that have been breaking out to new highs is harder. That's true. I mean, although we had sat on the S &P 500 at 7 ,500 for about 11 weeks. Yeah. I mean, it was a long time just churning around like at the index level with a lot of the sectors and stuff turning around underneath. But it never... A lot of people are at 7 ,100. It never really threatened to go back. It looked like it could a couple of times, but it never did. Do you want to continue? I will. Okay. It's up to you. Okay. Thanks. We're monitoring developments out of the Middle East alongside the price of crude oil.

6:15Iranian state news publishing a draft plan for the Strait of Hormuz that would ban U.S. and Israeli ships from transiting the waterway. The draft also says other nations that harmed Iran would not be allowed to sail through the strait until they paid compensation. Iran's state news said the plan is under review by a parliamentary committee. President Trump said yesterday he thinks the Iran war will end, in his words, pretty soon. Iran's chief negotiator accusing the president of staging what he called theater diplomacy. Yeah. Got a lot of factions over there, and some of them are hardliners. Do any of them really care about the Iranian people?

6:50No. Airbnb shares jumping on better than expected, unless they consider the Revolutionary Guard the Iranian people. And I think they do. Expecting second quarter results. The company's current quarter sales forecast also topped estimates. Airbnb raised its full year projections for both revenue growth and adjusted EBITDA margins. And in an interview with CNBC, CEO Brian Chesky credited AI for the company's acceleration.

7:20Greg Casar:I think a year ago, it was really a question. Is AI good for Airbnb or is AI bad for Airbnb? The verdict was out last year. I think now it's safe to say AI is the best thing to have happened to Airbnb. I think we're becoming an AI native company. And I think that is probably the number one explanation for our results. I'm going to figure out exactly how you'd use it. Do you understand it? I guess I could. Airbnb? How you would use AI to streamline everything at Airbnb. I'm going to think about it. Regarding AI and staffing at Airbnb, Chesky said the company's philosophy is not to employ fewer people, but to use the technology to make people more productive.

8:00And I say this because we have an interview coming up with a former chief information officer of Lululemon who wrote a book about, on the one hand, AI, there is a lot of promise, but there's a lot of companies that are already saying, oh, yeah, it's amazing. when it's real, they're just saying that, saying it's real. And there are other kinds, she points out, there are other places that, yeah, you can use it, but you need a lot of human supervision and, you know, employing it the right way and making sure that it's functioning properly. You've got to roll your sleeves up and really work hard for the payoff that hopefully is going to.

8:39Well, the last mile, I think, for AI. I mean, originally the argument was that AI would displace Airbnb. It would displace the online travel platforms like an Expedia and Booking. For instance, both, by the way, had strong earnings. So all of them had strong earnings. But the argument was that you don't need them anymore. It disintermediates those sorts of platforms. But AI agents can't execute. You can say, find me the itinerary. But how do you actually buy the ticket? So there would be a bunch of people that are ready to rent out their houses. and it'd be somewhere other than Airbnb, just you'd find it through AI.

9:17I don't know. It hasn't, I mean, to my knowledge, and it could be happening, I haven't seen that yet. And as a consumer, do you want to just deal directly with the homeowner? Not always. Sometimes you want that insurance. You want the vetting process. And all these things in five years, I don't know, it could be totally different and much better. And it could disemediate everyone. Facebook parent Meta ordered to pay more than half a billion dollars into a mental health fund for teens in New Mexico after a state court found the company harmed children's well-being. A judge finding the social media giant had created a public nuisance.

9:55He ordered Meta to make changes to protect younger users, including monthly limits on use of Facebook and Instagram and stricter controls on adult contact with minors online. Meta said it will appeal the ruling. It's preparing for a case in California next week over claims by dozens of states. The company designed its products to addict children. Got some details on OpenAI's upcoming consumer smart speaker like product, which is designed to challenge Amazon and Alphabet. According to Bloomberg, the device will look something like a donut and be about the size of a hockey puck. And I assume, Melissa, it would be the Taurus-type ring donut that they're talking about.

10:40As opposed to like a... Spherical, a color, a twist, a long john. A bear claw. Novelty shapes like hearts, stars, seasonal outlines. See, I'm using AI that talk about every type of donut. In this case, I think if you say donut like it's self-explanatory. Like a bagel. Well, that doesn't really have the inside thing. What are those little things called? The munchkins don't really come out of the bagel. It's sort of just an indentation. This would be, I hope it doesn't, these don't taste like a hockey puck, allowing it to easily be carried from room to room. You could carry it like that. Through the hole.

11:23The report says the device will function like an AI-enabled computer aimed at helping users fulfill daily tasks or at least a slated for next year. And it will be expensive, likely above$300 compares to Amazon's entry-level smart speaker device, which is about$40. Google's, which runs about$100. So do you have any of that? Do you have those? No, I don't want any device listening to me. It's like, who? I'm worried enough about this, especially when I'm talking about something and then remedies for foot fungus comes up on here. Why would that happen? It's like, wow. That's why it's called a smart foot.

12:05Cheese will be next. Coming up on Squawk Pod, how Congress wants to protect American jobs from AI. And yeah, there are politics involved. Chair of the House Progressive Caucus, Texas Rep Greg Kazar joins us.

12:18Greg Casar:There are productive and positive uses of AI. And our bill, in fact, doesn't tax those things like medical research. We want AI development that's going to help us cure cancer. Trading at Schwab is now powered by Ameritrade, bringing you an expanding library of education with even more ways to sharpen your trading skills. Access new online courses, insightful webcasts, articles, engaging videos, and more. All curated just for traders. Plus, guided learning paths with content designed to fit your unique interests. And no sifting to find exactly what you need, so you can spend your time learning to trade brilliantly.

12:59Learn more at schwapp.com slash trading. Insurance isn't one size fits all. That's why drivers have trusted Progressive's Name Your Price tool for years. Just tell Progressive what you want to pay, and they'll show you coverage options that fit your budget. Visit progressive.com to find a car insurance rate that works for you. Progressive Casualty Insurance Company and Affiliates. Price and coverage match limited by state law. You know what's not a great use of time? Getting data ready for AI. That's why IT leaders turn to CDW. We help unify data and govern it, building the foundation for your AI and agents to deliver results.

13:35Because when your data is ready, AI stops being an experiment and becomes an advantage. We configure, optimize, and deliver the tech that runs business. CDW. Make amazing happen.

13:53You're listening to Squawk Pod. Here's Joe Kernan. New legislation in the House is pushing for a tax on AI to help save workers from what could be mass unemployment. That's the fear anyway. Joining us now with the details, Congressman Greg Kuzar. He chairs the Progressive Caucus and is the co-author of the AI Tax and Work Protection Act. And Congressman, it's good to have you on. We're all trying to grapple with, you know, with with how to deal with these things. And as usual, both sides of the aisles probably view it very differently on what the most effective way is. In this case, you think it's a tax on AI to to help retrain workers that could lose their jobs.

14:40Is that the rationale? Yeah. Good morning.

14:42Greg Casar:It's good to be here. And we need a tax on AI to, one, disincentivize the current model where these AI companies and big corporations essentially get a tax break for automating your job. And second, we should take that tax revenue to be able to create millions of new, good American jobs. And basically, we've structured this tax so that it's very low right now when we aren't facing a mass unemployment crisis. But then we'd scale up. If AI does start putting millions of people out of work, we could scale that tax up to make sure that the unemployment rate does not skyrocket. That's what the CEOs have been warning us about.

15:27Greg Casar:That's how they actually, in many ways, are able to have these high valuations as they are saying they want to put people out of work. And I think that Americans want to work. They want to contribute to their communities. And this bill says we're not going to let a few billionaires become trillionaires while putting us out of work and us not do anything about it. But that, I think, is just a fundamentally different way of viewing what AI could mean to society, I guess, in the future. It made me think when we had, I don't know, agriculture probably employed 100 times the number of people it employs now.

16:10And we know that we have more food than we've ever had. So or the auto industry, you know, prior to that, we had, you know, we had buggies and horses. If if at the time something like this had been done to just maintain the status quo and to make it impossible to move forward with with the new stuff. Do you think that would have been a positive back then to really sort of handicap the the you're at? ahead of the progressive caucus. This sounds like you should be ahead of the regressive caucus, Congressman.

16:50Greg Casar:Well, Joe, as you and I talked about last time, there are productive and positive uses of AI. And our bill, in fact, doesn't tax those things like medical research. We want AI development that's going to help us cure cancer. We want AI that's going to help us make our society better and stronger. But we also know that these AI CEOs are out there talking about, quote, the end of work. We know Elon Musk has said people won't be needed for, quote, most things. That's not the kind of society we want, where a few of these Silicon Valley billionaires say, well, we will have, you know, Great Depression levels of unemployment, but we'll make that more tolerable with a little bit of universal basic income.

17:41Greg Casar:That's not what most people want, whether you're progressive like me or conservative or independent or whatever. We want to make sure we have positive uses of AI, but should not have mass unemployment. So that's why we've released this bill, the AI Tax and Work Protection Act, that keeps this tax rate very low right now, but goes upward if there is mass unemployment. And of course, nobody can tell you four years from now exactly what the unemployment rate might be. Four years ago, we didn't even have chat GPT. So we need a plan to make sure that we are prepared because Americans are asking us, what is Congress's plan?

18:21Greg Casar:If what the AI CEOs are saying is going to happen, we saw, for example, Block lay off 40 percent of its workforce, citing AI. If that is going to happen, then what's our plan? And this is a plan we can put in place to make sure we don't have to have mass unemployment in exchange for technological progress. We're back to the fundamental discussion about whether AI is going to long term be a a boon to society or not, because if you tax it, you'll get less of it. And if we tax it and China doesn't. Obviously, we're going to be trailing China eventually. It's a tough one because in certain ways, I agree with you that this is different than than some of the transformations that we saw in the past.

19:09I mean, you would have to say not everyone should be in agriculture right now. I mean, we were an agrarian country, you know, centuries ago. And we didn't want to stay like that. I don't think, Congressman, is there nothing analogous to what we're talking about now with this? This is this is revolutionary. And I don't know whether you can if you want to cherry pick the good things that you were talking about in terms of medicine and efficiency. I don't know how you can have those develop as quickly as we would like and then stop everything else that might be disruptive to certain sectors of employment.

19:50I don't see how you can really do it that way.

19:53Greg Casar:Well, that's exactly the question, Joe, that we've been wrestling with here. How can we make sure we develop the good uses of AI, be a leader in the world and technology, but not pay for it with mass unemployment? And when I look around the country, I see all sorts of great work that is worth doing where we don't have folks in those jobs. We have a shortage of teachers. We've got a shortage of firefighters. We've got a shortage of people to care for our elderly. We know there's a lot of people that aren't able to go to work because they don't have affordable child care or they're not able to go to work because they've got to care for an aging parent.

20:32Greg Casar:And so in the age of AI, if we suddenly start losing a significant number of workers and work in a structural way, we can actually have a modest tax on AI, use that funding to hire millions of workers. In fact, that's why my bill creates what we call the Work Protection Agency, just like FDR created the Works Progress Administration and built things like the Hoover Dam and the LaGuardia Airport, which actually improved our economy and our country and all of our lives. And so, Joe, I'm not talking about just stopping all of AI in its tracks. I'm saying we should lead in AI, but that doesn't mean we have to be in a country with large amounts of unemployment.

21:15Greg Casar:Since I was last on your show here in my own district where Oracle is headquartered, they laid off 21 ,000 people citing AI. And I just don't think and I think that conservatives don't think either that the cost of technological progress should be what we heard from the Anthropic CEO, which is Great Depression levels of unemployment. We can make sure we have a fully employed America and lead the world in AI. I was going to ask you about some of the things happening in your in the Democratic Party is a caucus of the team. I guess we're going to lose a lot of jobs in terms of law enforcement when there's no police left, when we get rid of all that and people employed in prisons.

21:59But we don't let's talk about that. Let's talk about that next time. Some of the proposals we're seeing. Look, what I think is really important. Talk about job losses when we shut down all the police departments and in prisons and everything else. There's going to be. What are all those cops going to do?

22:17Greg Casar:We got Joe. Come on. You know, look, I came to be on. You know what? You're a Texas Democrat.

22:25I don't think I don't think that's to say. I think you're probably looking at some of the stuff we're hearing from some of these candidates and go, whoa, I can't sign on to that. But let's talk. Let's talk next time and come back. And you're, you know, we'll hash it around. We're out of time, though. I know I introduced that. That sounds good.

22:43Greg Casar:We can talk AI safety next time we're on, because that's a real threat. Thank you, Joe. All right. All right. Thank you, Congressman. Next, the surprisingly weak July jobs report and lab-made viruses. MIT scientist Kevin Esvelt on the latest breakthrough to make germs. This could help replace antibiotics. That's why they did it. It's great work in that regard. The worry is downstream. What will they think of next? Is it life, Kevin? We could never agree whether viruses are life or not. Squawk Pod will be right back. Support for this podcast comes from Progressive, America's number one boat insurer.

Read the full transcript

23:21We've all made mistakes on the water, but there's one mistake you shouldn't make, being uninsured. With Progressive Boat Insurance, you can choose coverage for most mistakes you or other boaters could make, helping you float carefree all season long. Quote today, Progressive Casualty Insurance Company and Affiliates. Number one rating based on 2024 filings with State Departments of Insurance.

24:03That's schwab.com slash market update podcast or find Schwab market update wherever you get your podcasts. Ever finish your to-do list just in time for three more things to get added? That's what managing technology can feel like. There's always another ticket, another update, another thing demanding attention. That's why CDW's here. We help run, optimize, and evolve complex IT environments so your team can focus on moving the business forward. We've been there and solved that. When operations stop getting in the way, it's actually pretty amazing. We configure, optimize, and deliver the tech that runs business.

24:38CDW. Make amazing happen.

24:44Welcome back. This is SquawkPod from CNBC. You're watching SquawkBox on CNBC. I'm Melissa Lee along with Joe Kernan. We're just seconds away from the employment report for July. Rick Santelli is standing by at the CME in Chicago. And I don't know, Rick, Steve is still mad that Kevin Warsh hasn't called him and told him exactly what he's going to do with this. None of that is true. It's absolutely true. None of that is true. Rick, is it possible that with the two year where it is and Fed funds where they are, that just removing the easing bias and keeping neutral or maybe a slight tightening bias, can't you just hope that something positive happens in Iran and that maybe we don't have to get hikes?

25:32But you've wanted higher rates for a while, I think. But I'm still hoping that we don't need to choke off the economy to deal with supply issues that are causing some inflation. I agree. I agree. I agree that I've been for higher rates because of debt deficits, but not from the Federal Reserve. I believe that the market rates potentially ought to be higher to penalize Washington for all the spending. But on the short end, why should we make our payments to service debt higher? Why should we try to slow the economy and use a blunt instrument? Keep that rate where it is for longer. That's the way to go.

26:14Now the wild card is the war. And here we go, Joe. July employment figures populating. Buckle up on this one. Minus, minus 23 ,000 on nonfarm payrolls. I can hear the bond yields falling right now. Minus 23 ,000. That would be the second worst number of the year outside of the only other negative number. Minus 156 in February. Now, if you look at the two-month revision, another whopper, minus$103 ,000. Manufacturing payroll is actually a bright spot. Manufacturing up$5 ,000. That would be, what, the best since it was up$15 ,000 in March. And that was the high watermark going all the way back to$23 ,000.

27:01On the unemployment rate, 4.1%. And I'm sure that's because of the labor force shrinkage. 4.1 percent. That would actually be the lightest of the year. Once again, that would equal June of last year to find a lower number. You're in January of 25. Now let's look at the participation rate. And this is 61.4. We're expecting 61.6. It's one tenth lighter in the rearview mirror at 61.5. It equals where we were in Feb 21. You're in May of 2020. COVID era to find a smaller number. But as Treasury Secretary Besson has pointed out in numerous venues over the last week or so, that this is a demographic issue.

27:48It's not a weak economy. The rate's falling because of old and young. But that 25 to 54 age range is still doing well in terms of participation. And drumroll, please. The last part of this report is called the underemployment rate or U6. The other rates called U3. This also dropped 110 from expectations. We're expecting eight. Seven point nine arrived. Seven point nine equals where we were. We've had several of those readings to the lowest of the year to find a lower number than that. You're in June of 25. It's beautiful, Rick. Steve Leitzman, are you going to take back any of those three rate hikes that you've instituted as an imaginary Fed chair that you orchestrated earlier in the show?

28:36Are you ready to take back any of those? My concern is there's people listening on the radio and either don't hear the sarcasm in your voice or see the smile on your face. No one thinks you can raise rates, although you'd like to. I want to say quickly, Joe, that I am a little less concerned about the negative number here than maybe meets the headline I hear. And the reason is because we have this minus 40K in leisure and hospitality. And it strikes me that a piece of this, a good piece of this, could be the unwinding of the World Cup hiring. And there was some expectation that it would happen.

29:16And not many people have really I don't know how much that was built in. OK, the other thing I'm scratching my head about and, you know, there's no hair to scratch. Anyway, but is the minus 50 on local government education. That's a seasonal thing. You have to go back, by the way, to the pandemic to find a bigger drop on a monthly basis than than we had in this month. So I'm thinking there's a little bit of seasonality in this number. I don't know what's happening. I have to dig in further. What's happening with the participation rate? As Rick said, we thought it was going to be a little bit higher.

29:55It ends up being, people thought it was going to bounce back after. It did not. Somebody or some group of people are dropping out of the workforce for reasons we have to figure out. And I don't know the place I would look first as soon as I stopped talking, Steve, is the prime age workers. If that's happening, that's a problem. It could be at the senior end or it could be at the younger end, in which case it's less of an issue. Is it, have you come more into my, and Rick, I don't know, Steve says so many things you probably disagree with. You can take your pick, but do you feel more on my side that maybe we can just wait and maybe.

30:31I have felt we could wait. You asked me if I saw the fence. Here's something real easy. Everybody's talking about the two year, okay? The two year high watermark for this cycle, and it was obviously all war related, was a four point three five percent high yield close. We're basically on the verge of 20 basis points lower. The moral of that story is, is that the big run up in nominal rates that happened on February 28th and beyond due to the war, which affected the rise in real rates, can come down rather aggressively should we see some more issues in the Mideast get settled out. And therein lies the linchpin between the neokashkaris, the Bowmans, all these Fed officials that want to raise rates that before the war didn't want to raise rates.

31:20They need to focus on supply chain issues, energy issues, global energy issues, and how they may be resolved and impact the market should all these things in the Middle East start to go in a more fine-tuned, better optimistic fashion. That is the whole key here, in my opinion. A study from the Ark Institute found that after scientists' prompts, AI created new recipes for virus genomes. When scientists inserted those viruses into bacteria, more than a dozen turned out to be viable with the ability to infect other bacteria. Joining us now, associate professor at MIT, Kevin Esfeld. He's an evolutionary engineer and a bioscience defense analyst.

32:01Great to have you, Kevin. It's a pleasure.

32:04Greg Casar:Thanks. I mean, the headline sounds nefarious, But this is actually a good development in terms of the fight against drug resistant bacteria, correct? That's exactly right. So what they did is they trained AI on thousands of sequences from viruses that infect one kind of bacteria. And they used AI to predict new genome sequences for viruses that could infect these kinds of bacteria. And remarkably, some of them worked. the bacteriophages the the the smallest is that the the smallest number of base pairs of any in any and i'll say living and i'm putting quotation marks on living uh organism that's the easiest thing to make isn't it if you're trying to to create life and is it life kevin we could never agree whether viruses are life or not what matters is that they work and yeah these aren't even full these are just viruses that infect bacteria right how many genes 10 12 even fewer so typically five or six for some of the small ones that's true for some of our viruses too right so in principle you could do this on a human virus but the scientists here took a lot of precautions they didn't do it on human viruses they never trained their ai on any human viral sequences.

33:22Greg Casar:And as you said, this could help replace antibiotics. That's why they did it. It's great work in that regard. The worry is downstream. Yeah. So, Kevin, this is more than what we think might have happened in Wuhan. I'd be interested in your take on that because of the CGG repeating codons. And this is more than just gain of function where you have an existing virus where you just change it a little bit so it can maybe infect human cells. This is from scratch? It's not quite from scratch because they trained it on 15 ,000 viruses that normally infect this kind of bacteria. So it's more like imagining if you had trained it on all of the COVID variants, then the AI would spit out new COVID variants and some of those would spread.

34:13Greg Casar:Now that's a harder problem, but that would be the next phase. And thankfully, right now, due to new government policy, you cannot get funding to run that experiment to actually make the predicted human viruses in the lab. That's no longer allowed. And I think that's a good thing. You obviously saw that Dr. David Baltimore passed away a couple of months ago, who was, I mean, in terms of virology, reverse transcriptase and everything, he actually was one of my professors at one point. But do you know when he said that that smoking gun that we saw in COVID, did he get pressure to walk that back?

34:53Because even towards the very end, he said, OK, it's not a smoking gun, but it strongly indicates that this would be weird to occur in nature. And I'm sorry to get off topic, but this is something we've been talking about the last couple, obviously, with all the Fauci testimony. What do you make of that?

35:10Greg Casar:I think it was a bit of a travesty, frankly, how much politics got in the way of trying to decide where it came from, when really, even today, we don't have enough information to know. It could have been either route. And the rational thing to do would say, we don't know. China won't give us enough information. Either it was their fault because it was a wet market, or it was their fault because they did work in the lab that was unsafe. And in both cases, the thing to do is fewer wet markets and better safety practices in labs. So you would do both in terms of trying to make sure it It doesn't happen again.

35:43It is a little weird that the first cases were like a mile from the lab. It's a little weird. It's weird in a lot of ways.

35:50Greg Casar:It doesn't make sense either way. Something is strange about it no matter which hypothesis you prefer. Was it Ockham or Ockham? I think it's Ockham. Ockham's razor would say that that was inserted, those two codons. But anyway, sorry, dude. No, no. I think this is a good... But I want to get back to the study that was published in Science, Kevin. And so when you say the problem is downstream, I mean, how concerned are you that if, you know, in the wrong hands, that something more nefarious could result from this sort of experiment? So here's the thing. These scientists did everything right. Only works on phages that infect bacteria, not humans.

36:31Greg Casar:Didn't train their broader AI on any sequences from human infecting viruses. But there is no reason to think that future researchers cannot and will not attempt the same thing for viruses that infect other organisms. Probably not humans at first. That would be too obvious. But maybe other animals, pest animals. And the problem is, if you can predict new variants of a virus that will overcome existing immunity, and it would work in a mammal, then it probably would work in humans. Was this a piece? What was the level of containment that these were done in before? I mean, it would seem that that that and what was it in China?

37:15Like a P1. Right. I mean, that was right. It was a P2 at most, wasn't it? And pretty shoddy at that. Yeah.

37:23Greg Casar:So the problem with China is they did a lot of the key experiments in what's called biosafety level two. But here these experiments were only viruses that can infect bacteria. They have zero chance of getting into a human cell. So these experiments from the science paper were absolutely perfectly safe. There is no way that they could possibly have caused harm to any human. But it would not be safe at all to try to do the same thing for any virus that would even infect a mouse. Because if you can make it work in a mouse, then you can almost certainly make it work in a human. And then you've opened the door to a bad actor doing it in humans.

38:01Professor, we have to leave it there. Great to get your thoughts. Thank you. We're toast. And that is Squawk Pod for today. Thanks for listening. Squawk Box is hosted by Joe Kernan, Becky Quick, and Andrew Ross Sorkin. Big thanks to Melissa Lee for sitting in today. You can tune in on television weekday mornings on CNBC starting at 6 Eastern. Or get the smartest takes and analysis from our TV show right into your ears when you follow Squawk Pod wherever you get your podcasts. Have a great weekend. We'll meet you right back here on Monday. We are clear. Thanks, guys.

39:00Guided learning paths with content designed to fit your unique interests. And no sifting to find exactly what you need, so you can spend your time learning to trade brilliantly. Learn more at schwab.com slash trading.

From the publisher

In a surprise to Wall Street, the Labor Department reported a loss of 23,000 nonfarm jobs in July. CNBC’s Rick Santelli and Steve Liesman break down the report and what it means for the broader health of the economy. Representative Greg Casar (D-TX) is focused on AI’s impact on the labor force in his latest legislation to tax hyperscalers, while MIT Associate Professor Kevin Esvelt is closely watching AI’s achievements in science. He considers the implications of a new study from the Arc Institute that showed AI-created viruses are viable in the real world. Plus, in a child harms case in New Mexico, Meta was ordered to pay $567 million into an abatement fund, and OpenAI plans to release a home device to consumers. 

 

Rep. Greg Casar - 14:23

Rick Santelli & Steve Liesman - 25:24

Kevin Esvelt - 32:13

 

In this episode:

Greg Casar, @RepCasar

Joe Kernen, @JoeSquawk

Melissa Lee, @MelissaLeeCNBC

Steve Liesman, @steveliesman

Rick Santelli, @RickSantelli

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.

More from Squawk Pod

All 546 episodes
AI-Made Viruses, A July Jobs Report, & Meta’s Child Harms Case 8/7/26Squawk Pod · 35 min
Listen in VO