Transcription
Open AI and Anthropic. Something quietly legendary happened yesterday in business. Everybody's, we're going to get into the political story, but everybody was paying attention to this one headline 22 hours ago. Open AI co-founder Andre Karpathy joins Anthropic. Wait, what? Yes. Didn't we just get the news that Elon Musk was soon, uh, you know, Open AI and they dropped the case because it was too late to be filed? Yes. Wasn't that like a timing thing where, hey, maybe Sam Altman got himself a little bit of a victory? No.
Let me read this to you. Andre Karpathy, known as one of the best AI researchers in the world and a founding member of OpenAI, announced Tuesday that he's joining AI lab Anthropic. That's like going from the Yankees to the Red Sox. That, that's like going from the Dodgers to the Yankees, which is actually a good move if you do that. That's like going from, you know, the the Lakers to the Clippers. You don't join the Clippers. You don't join a different team, but they're leaving OpenAI and going to Anthropic. Why this matters? The hire is a major coup for Anthropic in the high-stakes competition for elite AI talent, and another sign the company is emerging as a magnet for some of the industry's most respected technical minds. Driving the news, Karpathy will start next week at Anthropic's pre-training program, which is responsible for the massive training runs that give Claude its core knowledge capabilities. And according to Anthropic, Karpathy will help launch a new team focused on using Claude itself to accelerate, uh, pre-training, uh, uh, research and increasing important frontier as AI companies race to automate parts of AI development. I think the next few years at the frontier and LLM will be especially formative. I am very excited to join the team here and get back to R&D. Tom, why is this such a big deal?
>> It's a big deal because he and Sam Altman have gone back and forth. I mean, he tried to leave before, left for a little while, tried to do his own thing and this and but he was always assumed to be out there kind of in Sam Altman's orbit. You know, you started here and everything and now he just says, you know what, I'm going to Anthropic. You have to remember Anthropic was started, you know, Adario left and went in the name of safety. He said, I think safety is more important. I think there should be safe AI. I think we need to put guardrails around it. I think we need to do these things. And he was claiming that Sam and OpenAI were just doing going too fast and not doing it. Well, uh, Karpathy felt this largely the same way. And now he comes up, he's going there and this is the guy who's taught like the first, there's an interesting set, Pat. He gets permission to teach this course on neural networks at Stanford and he got his PhD at Stanford. The first time he teaches it, they rounded up like 150 curious students. You ready for this? Two years later, there's a waiting list and every time the course was offered, there's 750 students who want to get into it. So, this guy has been seen at Stanford and in Silicon Valley as this neural network genius and now he is on the eve of IPO season, he is now at Anthropic with, uh, Dario.
>> This is a big deal. Look, what do you think about this story?
>> It's really interesting you bring up his point about safety. I didn't know that Karpathy was aligned that way. Um, because Karpathy put out on X three, four months ago a study that he had conducted that looked at the risk to existing US jobs by sector by, I think there's 192 job classifications or something in the Bureau of Labor Statistics. And he went by jobs category across the entire US economy and rated each individual job category with the risk of job loss from AI over the next several years.
>> And what he found was was shocking, which was that roughly 26 million jobs in America are at between very high or extremely high risk of being disintermediated by AI. He put this out. It's, it's a fascinating chart because it just shows each area as a square and then green is very low risk. So you, there, there it is right there. That's the, uh, um, that's, that's part of it there, right? So you, you can see the areas. So you know.
>> Red is bad, green is good.
>> Red is bad, green is good. And you can see it there. And, and so this genius, this guy that's worried about safety, which to your point, Pat, I think is really interesting going from to the to the more safety or the, uh, however you want to term it, focused, uh, AI model. Uh, this is something he put out three months ago, which is, I mean, look at it. Accountants gone, lawyers gone, market researchers gone, software developers gone, general office clerks gone. I mean, it is 26 million. I mean, but, but if you keep going, some of these jobs, you just said stuff that maybe we would say, "Okay, that makes sense though, Luke. Those are going to be gone." Bookkeeping. Okay. Gone. Receptionists. Okay.
>> Yep.
>> And then, but you go on the right side. Software developers.
>> Gone.
>> Gone. Gone.
>> Computer. What does that say on the bottom? Computer. I can't see the.
>> Some sort of. Yeah, it's.
>> CS. Let's just say computer science. What? Lawyers. Lawyers is in the red.
>> Lawyers is, lawyers gone.
>> Wow. Project managers, human resources, accountants, purchasing, uh, which ones are green? Let's, let's focus on the greening. Home health, hand laborers, uh, food and service, beverage cooks, waiters, electricians, you know, right there in the middle, construction laborers, carpenter, general maintenance, janitors, child care, you know, it is, it's a fascinating chart because you have one of the, one of the, you know, the best geniuses in AI three months ago saying what China was to blue collar workers 25 years ago, AI is going to be to white collar workers in the next five years. So if I'm, I just came from the eighth grade graduation right this morning and my niece is graduating as well, high school. She's valedictorian. She's going to a very good school. So these kids are doing everything that they're told to do. When they see this, like, uh, Steve Wozniak is giving a commencement speech. I don't know if you guys saw this or not. So we saw what, uh, Eric Schmidt got with Google. We saw what the other lady got when she spoke at University of Central Florida. Booed hardcore. Here's what Wozniak said. Go ahead, Rob.
>> All have AI. You all have AI. Actual intelligence.
>> Now, watch what he says, which is very funny what he says.
>> Life in the technical world, I've been following people that were trying to figure out how to make a brain, software, hardware, synapse, chips. And I was at a company where the engineers figured out how to make a brain. Takes nine months.
>> It's funny, right? So, he gave a promising message, right? He gave a message that is at least giving kids some confidence. How do you feel? If I'm a 19-year-old, 22-year-old right now coming out of high school going to college. What? I got a lot of worries. The kids are talking to each other. What do you tell them?
>> I don't know. I mean, I've got three boys, 25, 23, and 20. And we have this conversation with them a lot because the world is changing. So there are kids graduating high school today that are going to go and they've played by all the right rules and the world's going to be different in, forget about four years, it's going to be different in two years at the pace at which AI is moving. So I don't know what I, I think what you have to do is, and I know what we've done with our boys, which is be honest with them, communicate, and then just talk about critical thinking. Yeah, like, like Woz says, learn how to think, think critically, have character, you know, as I.
>> What skills are they taking? What skills are your three boys going? What are you majoring in?
>> So, the oldest one's already out of college. He's working, uh, in, uh, audit analytics at a major, uh, US bank. Um, and he's actually, uh, working on AI governance. His mother and I, he's going to be taking a, an online course through MIT that anyone can take that you have to, you know, there's a tuition associated with it, but it's around AI because he sees, he's basically, he sees the dragon coming. And, you know, we've taught them, and when you see the dragon, you don't run from the dragon, you run at the dragon. Um, middle one is actually going into, uh, funeral services, uh, very AI-proof.
>> Sales or?
>> Uh, he's, his, his, uh, longtime girlfriend, uh, his family owns, her family owns a funeral home. And so he's been working there. Uh, and so he actually likes the business and it's, and it's a great business. So he's moving there. And the youngest one's doing cybersecurity, which we specifically talked about. Let's find a something where there is you can leverage it where it's not, you know, you're, you're not going to be in a red square.
>> Luke, what's that MIT course and is that available for everybody or?
>> It's available for everybody. I would, you know, I'd have to ask my wife. I know she's watching.
>> Is it an online type of?
>> It's an online type of class. Anybody can take it.
>> Would you mind texting your wife and see if she can respond back? So maybe by the end of it, we can share with the audience. Richard, your thoughts on this?
>> Um, I'm, I'm actually quite, uh, positive. Um, we, we do know that, uh, in recent, uh, well, actually decades, there's been this tendency to, uh, for companies, you know, uh, to merge and companies get bigger and, um, staffing, um, in many big companies has become somewhat bloated. I mean, there was, uh, um, David Graeber wrote this book, "Bullshit Jobs," and apologize, I apologize for the language, it's the book title. Um, and, um, you know, the argument is that actually a lot of these jobs are sort of redundant and that has become part of the sort of European-American, you know, industrialized country business, um, modus operandi. And I think these jobs are clearly, um, now endangered. But at the same time, AI gives a lot of people new opportunities. And the most important area which I think is very important for economic growth is not going to suffer. What is that? And I think that's right up your alley and, and all the things you're doing, Pat. Entrepreneurs, you know, owners of small firms, small firms, uh, micro businesses, entrepreneurs have now more opportunities. And essentially, what, what, uh, you know, I'm, um, I would, you know, as, as a professor also at uni, what I'm telling the young generation is, you, you got to, you know, if you so far you haven't thought about being an entrepreneur, you should think about it now. And it's an exciting thing. And I try to, you know, give some examples and, you know, essentially think of ways to set up a new business, offer new goods and services. And there's just so many areas, um, because of the changes and there are going to be big changes, um, it's really a fantastic opportunity for small firms. Now, one important bottleneck exists, and that is small firms, and that's very well known. I mean, they are the biggest employer, and that's why I'm not really worried that much about the overall economy. Um, and I think if, if things go well and done well and the policies, the right policies adopted, then we can actually see literally higher economic growth as a result of this and also more job creation because it's not the big firms that are the big employer. 60% of jobs are with these small and medium-sized enterprises.
>> And in many ways, they're less effective because they're usually lean and they actually, they'd like to hire more people, but they can't. Because so what is the restriction here? What is the, uh, the limiting factor that has been limiting growth of small firms because they're the ones that deliver productivity and, and, um, you know, job creation and therefore really the most important players in the economy. It's just because they're small, we tend to ignore them. Is not famous big names there. There's a totally unknown tiny local, uh, players, but they could be exporters globally, as we see in, in Germany, for example, many of the family-owned small firms are global exporters. Well, the limiting factor is finance. Most small firms are in the position where they already have a lot of good ideas and great ideas to expand, but, you know, the money is the, uh, the limiting factor. And I mean, there's, there's, there's a lot of research being done on this. Small firms are mostly held back by lack of funding from their main source of external funding. Of course, there's internal funding, and that's the single most important source for everyone, is their retained earnings and also family and friends. But for external funding, small firms, and we're talking about these micro businesses, very small firms, they're not really able to access, you know, capital markets.
>> The way the bigger guys.
>> Even venture capital. I mean, it's all much bigger. So they get money, external money from banks, and there is one principle, and we did a study with my, uh, collaborator, uh, on the US over 20 years, um, in total, because there used to be, you know, more banks, uh, in total. It's, it's more than 10,000 banks that we followed for 20 years. Of course, during that time, many also merged, and we classified them into different size categories. And what we found is a very, very strong message. The bigger a bank is, the less it's interested in.
>> Small businesses.
>> In lending to small firms. And of course, that's, you know, makes perfect sense. In many ways, you can't blame the banks. Um, and so the, the banks that lend most to small firms are the small banks. And it's really the smallest category is the biggest lender to small firms. The second smallest category of banks is the second biggest lender. And so it goes on until, you know, you come to the biggest banks, and they're the smallest. You know, they have the lowest propensity to lend to small firms. And now, what we also found over time, over these 20 years, is that because quite naturally banks also grow, and as they grow and move into a new size category, they also lend less to small firms. That alone, that alone, nothing else, uh, you know, considering, is a reason why we should always create new banks, and the government should make it easy. The regulator should make, should make it easy to always have new startup banks, small banks, because, you know, banks merge and get bigger and grow, and then the lowest category, smallest category of small firms already not interesting, uh, for them. And we've had this problem in the past 15 years in the US and in Europe very strongly that there's been a wave of mergers. The number of banks has declined by thousands, by thousands.
>> And it's interesting what you said though. Here's the part on the correlation because when you said 60% of the jobs are small business owners, you're the one that shared with me and taught me the, you know, we had 14,000. We talked about it the last couple years. And so you got me thinking with this, but do you know what else has been declining, which is a concern? Is the same way bigger banks are picking up the smaller banks or the banks that are screwing up and going out of business. So Chase keeps getting bigger.
>> Yeah.
>> Do you know, you said 60% of jobs are created by small business owners, right? We've always thought about employment.
>> This is the.
>> So 60% of total employment is with small and medium-sized enterprises.
>> Do you know what that number is today? That used to be the case in the 70s and the 80s.
>> And, yeah, it's going down also.
>> It's at 45% today.
>> Yeah. Yeah. Which is unhealthy. It's unhealthy. It's not good. There is an opportunity for the younger entrepreneurs to do something with that. I want to remind you guys that the Vault Conference we do once a year, Rob, if you want to, uh, for those of you guys that are small business owners and you want to find how to compete in a market like this, don't go about it yourself. Be in a community with other people that they can share with you what they're doing. We grew this thing from the first event being 440 people to, uh, last year nearly 8,000 people. This year we're expecting 12,000 people at MGM Grand Arena. Rob, I don't know if they've given you the video or not. If you do have it, if you want to play the clip, go for it. And it's a great opportunity for all of us to get together in one place where we talk about business, family, there's no walking on eggshells. We talk about every single thing at one event. Go ahead, Rob.
>> Successful business people speak a language others don't. It's like when I lived in Germany at age 10 and a month later I'm speaking German because I immerse myself in a community of Germans. I speak that language. The same applies for business. So studies have shown that 65% of people that immerse themselves in a new language completely feel like a different person on the other side. So what does it mean? You go to a business conference. You're at the Vault Conference. 12,000 other people around you for four days. We're speaking a language that we're fluent in in business. How we walk into a room, how we negotiate a deal, how we hire somebody, how we raise capital, how we come up with a new strategy. The more you're around it, you all of a sudden start speaking that language fluently. So when you go back home, you do deals in a different way. You handle your business in a different way. If you want to get fluent in speaking a business language, once a year we host a conference called the Vault Conference. This year, for the first time, we're doing it at the MGM Grand Arena in Las Vegas where we go through this 296-page manual with 12,000 people over four days in one room. If you haven't yet registered, click on the link above or below. Bring your spouse, bring your partner, bring your peers. Let's spend four days to get out. Las Vegas, August 31st through September 3rd. Looking forward to seeing you there. Go to vault2026.com. Get your ticket. Again, vault2026.com. Get your ticket. Cannot wait to see you and your family at the Vault Conference together. If you enjoyed this video, you want to watch more videos like this, click here. And if you want to watch the entire podcast, click here.