Karl Mehta
@karlmehta • 144,217 subscribers
3x Exited Founder/ CEO of tech cos, Chairman Emeritus- QUIN(Quad), former VC@Menlo Ventures, Author of 2 books, fmr White House fellow. All tweets personal.
Shorts
Videos
Dario Amodei says Anthropic held its first version of Claude back about six months, and gave up the ChatGPT moment doing it: "When we had our first product, our first version of Claude, which is our model, we actually delayed the release of that model roughly six months because this was such a new technology that we weren't sure of the safety properties." "We weren't sure we wanted to be the ones to kind of kick off a race." "This was just before ChatGPT." "So we arguably had the opportunity to seize the ChatGPT moment and we chose to release a little later, which I think had real commercial consequences but set the culture of the company." ChatGPT launched on November 30, 2022. Anthropic launched Claude in March 2023. - Dario Amodei (Dario Amodei), co-founder and CEO of Anthropic (Anthropic), at the Council on Foreign Relations (Council on Foreign Relations) CEO Speaker Series.
Karl Mehta116,514 görüntüleme • 4 gün önce
Sam Altman says the heaviest AI user on the planet six and a half years ago used what the average person on Earth uses today: "Six and a half years ago, the world token leader was an OpenAI employee using about 100,000 tokens a month." "And this seemed ludicrous at the time. The worldwide average per capita is like zero." "Now, the worldwide average of tokens per month, which I think tokens are the dumbest metric, but it's what we have. It's like 100,000." "And the token leader at OpenAI uses something in the hundreds of billions." "If that happens again, which I think it probably will, then in another six and a half years, the average person uses, let's say, 500 billion tokens a month." If he is right, the average person's AI usage grows five million times over the next six and a half years, and OpenAI is already building the capacity for it. That is the number enterprise boards should be planning against this year. Demand at that scale changes what a reasonable AI budget looks like. - Sam Altman (Sam Altman), CEO of OpenAI, with Garry Tan (Garry Tan) at Y Combinator (Y Combinator) Startup School 2026.
Karl Mehta654,557 görüntüleme • 1 ay önce
Satya Nadella says LinkedIn merged four job titles, product manager, designer, front-end engineer and back-end engineer, into one: "I'll give you at LinkedIn, we used to have product managers, we had designers, we had front-end engineers, and then we had back-end engineers and so on." "So what we did is we sort of took those first four roles and combined them. In fact, increased scope and said, they're all full-stack builders." "So at the same time, as you can imagine, if we're to build an AI product today, there's a complete new workflow, right? It starts with evals, right?" "So basically, there's this eval to science, to infrastructure." "And so evals are done by these full-stack builders and what have you and product managers in the new form, the infrastructure is built by the systems engineers at the back-end because they support the science that supports the product." "So in some sense, there's a new loop and you have to structurally change." LinkedIn has already put this into hiring. Its Associate Product Manager program is finished, and the replacement, the Associate Product Builder track, teaches code, design and product management at the same time. Evals sit at the front of that workflow rather than the end, so writing them is now part of building the product instead of a check before shipping. - Satya Nadella (Satya Nadella), Chairman and CEO of Microsoft (Microsoft), with the All-In Podcast (The All-In Podcast) at USA House, Davos 2026.
Karl Mehta530,339 görüntüleme • 26 gün önce
Former Google CEO Eric Schmidt says a frontier AI model costs up to $500 million to build and is worth about six months: "Are the scaling laws stopping?" "And the answer right now is no, which is unbelievable." "A model costs $200 to $500 million, you release it, and it has a life expectancy of about six months." "And then there's another one." "And the competition between the leading players is so brutal, right, that they literally count in the number of weeks of value of these things, which costs hundreds of millions of dollars." Schmidt is talking about the build cost. There is a second cost underneath it, which is re-proving each replacement, because whatever testing cleared one version does not carry over to the next. A company that standardizes on a frontier model this year is signing up to do that re-approval work twice a year. - Eric Schmidt (Eric Schmidt), former CEO of Google (Google) and now Chair and CEO of Relativity Space, with Alan Blinder at Princeton's Griswold Center (Princeton University).
Karl Mehta492,935 görüntüleme • 25 gün önce
Matthew Walker is a former Harvard professor who showed 6-hour nights can push your genes toward heart disease and cancer in 1 week. On Jay Shetty's podcast, he revealed 10 "normal" habits that wreck your sleep, mood, and brain: 1) Checking your phone right after waking.
Karl Mehta2,117,941 görüntüleme • 3 ay önce
Geoffrey Hinton says a big language model runs on about 1% of your brain's connections and still ends up knowing more than you: "So in your brain, you have a hundred trillion connections, roughly speaking. Okay. That's a lot. And you only live for about two billion seconds. That's not much." "If you compare how many seconds you live for, with how many connections you've got, you have a whole lot more connections than experiences." "Now with these neural nets, it's sort of the other way round. They only have of the order of a trillion connections. So like 1% of your connections, even in a big language model, many of them fewer, but they get thousands of times more experience than you." "So the big language models are solving the problem with not many connections, only a trillion. How do I make use of a huge amount of experience?" "And back propagation is really, really good at packing huge amounts of knowledge into not many connections." "But that's not the problem we're solving. We've got huge numbers of connections, not much experience. We need to sort of extract the most we can from each experience." Two to three billion seconds is the whole budget. Everything you know, you learned inside it. So evolution built you to squeeze a lot out of very little. Hinton's point is that a language model has the opposite problem and the opposite fix, and backprop turned out to be extremely good at that fix. Worth noticing what this predicts about failure. A system running on 1% of your wiring and thousands of times your experience is not going to fail the way you do. You fail from having seen too few examples. It fails from compressing too many into too little, and the compression is where the errors get made. That is a strange thing to be deploying into hospitals and courts with no way to inspect it. We test these systems by asking them questions, which tells you what came out. Nobody can yet look at a trillion connections and say what got packed in. - Geoffrey Hinton, Nobel laureate and Turing Award winner, on StarTalk (StarTalk) with Neil deGrasse Tyson.
Karl Mehta660,338 görüntüleme • 1 ay önce
Elon Musk told SpaceX employees that AI revenue will exceed everything else SpaceX earns, starting in September: "Not probably, definitely, our AI revenue will exceed all other SpaceX revenue, probably in September, like next month," "and will significantly exceed all other SpaceX revenue in the fourth quarter." "So AI has become an extremely important part of SpaceX's future." "I think it's actually vital that we succeed in AI." SpaceX reported $7.81 billion in second-quarter revenue, up 92% from a year earlier, with AI revenue at $2.56 billion. The remark came in a company update SpaceX published on August 11. - Elon Musk (Elon Musk), SpaceX company update, SpaceX
Karl Mehta123,100 görüntüleme • 10 gün önce
Warren Buffett says he could collect 20 to 40 billion dollars a year from government bonds without taking any risk, and that is the number he holds AI spending against: "AI companies, you're putting out huge amounts of money. And I can put huge amounts of money into government bonds and get, you know, 20 or 30 or 40 billion dollars a year in terms of payments from them." "So a good business is one that earns a lot more than, and has prospects of continuing to earn a lot more than the returns on essentially riskless investments, which you could define as Treasuries." "But if you take something like American Express, you know, most of the banks earn 13, 14% on capital." "But it's so different that it earns 30% plus on capital, and it does not incur more risk in doing so than the banks that earn 13 or 14%." "And the trick in life is to find, I mean in investing, is to find businesses that are going to earn high returns on capital for an extended period of time." "A long period of time gets to be very important because those doubles later on are very big numbers." That is a hurdle rate, and you can hold the next AI capex announcement against it the same day you read it. His second condition is the harder one. Earning high returns for an extended period means somebody has to show the returns will hold, and almost every number that would show it for AI is published by the company doing the spending. When the entity being measured also runs the measurement, what you are reading is marketing. Buffett has been able to check American Express against decades of audited filings. There is no equivalent for an AI model yet, and building it is the whole job. - Warren Buffett, chairman of Berkshire Hathaway, on CNBC (CNBC).
Karl Mehta317,150 görüntüleme • 1 ay önce
Demis Hassabis says he sold DeepMind in 2014 because staying independent needed $7 billion, and he could barely raise $10 million: "He saw Google as an AI company. The reason I sold was I knew it would require 7 billion dollars of investment in order to take on, try and become one of those companies." "And I was barely able to scrape together 10 million dollar rounds at that point." "And the problem was is that we'd done enough, the Atari thing, we had alerted enough of the Silicon Valley titans. They were coming after my people." "We had no business, we didn't even have a product, obviously not, because I was trying to make the core technology." "I was paying myself zero and people were being paid like 100k a year and then they were getting offers of 10 million a year." "And then if that didn't work, they could afford to come back and say, here's 50 million a year." "So at some point, that was going to go wrong." P.S: I'm hosting the AI Assurance & Governance Summit 2026 on October 1 at the Stanford Faculty Club, one day and one track, to work on the part of this that money did not solve. Frontier labs, regulated industries, insurers and investors in the room. Link in the reply. - Demis Hassabis (Demis Hassabis), co-founder and CEO of Google DeepMind, at the Worshipful Company of Information Technologists Annual Lecture.
Karl Mehta302,818 görüntüleme • 1 ay önce
Sundar Pichai, the CEO of Alphabet, was asked point blank if AI is a bubble. He didn't dodge. He said yes -- partly -- and that not even Google gets out clean: "You see real demand, and we are constrained in our ability to serve that demand." "When we go through these investment cycles, there are moments we overshoot -- collectively, as an industry." "We can look back at the internet: there was clearly a lot of excess investment." "It's both rational, and there are elements of irrationality through a moment like this." "No company is going to be immune, including us. If you overinvest, we'll have to work through that phase." The tell isn't that he admitted the irrationality. It's that the man sitting on the most compute on earth is already pricing in the overshoot -- while telling you the demand underneath it is real. A bubble and a revolution were never mutually exclusive. The internet was both.
Karl Mehta409,103 görüntüleme • 2 ay önce
Satya Nadella reveals why every company may need its own AI model: the model becomes the new company database. "To me, a model is like the database market." "A firm should be able to take the tacit knowledge it has and embed it inside weights in a model that they control." "When somebody asks me how many models should there be, I'll say as many models as firms in the world." The contrarian part: the value may not sit in one universal frontier model. It sits in each company turning its private operating knowledge into a controlled model.
Karl Mehta468,212 görüntüleme • 2 ay önce
Dario Amodei, the CEO of Anthropic, went on CNN and put a number on the thing most AI CEOs will not say out loud: Half of all entry-level white-collar jobs gone. Unemployment at 10 to 20%. Inside 1 to 5 years. "The most salient feature of the technology, and what is driving all of this, is how fast the technology is getting better." "A couple of years ago, you could say that AI models were maybe as good as a smart high school student. I would say that now they're as good as a smart college student, and sort of reaching past that." "I really worry, particularly at the entry level, that the AI models are very much at the center of what an entry level human worker would do." "What is striking to me about this AI boom is that it's bigger and it's broader and it's moving faster than anything has before." "People will adapt, but they may not adapt fast enough. And so there may be an adjustment period." The tell: this is not a critic outside the industry. It is the man building the models saying the quiet part on camera -- because he thinks the people whose jobs are on the line adapt slower than the models improve. High school to college in two years was the warm-up. The entry level is where it lands first.
Karl Mehta341,287 görüntüleme • 2 ay önce
Sam Altman says OpenAI plans to give an AI research intern 500,000 A100-equivalent GPUs by September: "I think we're at a point where, and we've set this goal that by September of this year, we will use 500,000 A100 equivalent GPUs, like a lot of computing power, as an AI research intern." "And by March of 2028, that we will have a full end-to-end, very talented researcher, like figuring out complete new architectures." He said this at Stanford in May, which puts the September target three weeks away. OpenAI's published plan carries the same March 2028 direction, that a significant fraction of its research may be done by AI systems working alongside its own researchers. Neither milestone has been announced as met, and September is the first one with a number attached to it. - Sam Altman (Sam Altman), CEO of OpenAI, at Stanford (Stanford University), CS153 Frontier Systems.
Karl Mehta129,043 görüntüleme • 24 gün önce
Ben Horowitz, co-founder of a16z, says an AI version of him could get on a Zoom and tell his finance team to wire $500 million: "Oh my god, somebody's gonna go on a Zoom." "It's gonna be AI me, and they're gonna tell my finance team to wire like, you know, $500 million." "Don't believe anything from me unless it's got my cryptographic key on it, all that kind of thing." "I knew these problems were coming. They're coming so fast now." "First is just, are you a human or are you a bot?" "I think everybody is going to really, really want to know that, be it social media, a dating app, a Zoom call... Am I talking to an actual human?" Every company taking instructions on a call needs that control, and none of them can test whether the key or the human check works before the wire goes out. P.S. Come hear Ben Horowitz open the AI Assurance and Governance Summit on October 1 at the Stanford Faculty Club. - Ben Horowitz (benahorowitz.eth), Co-Founder & General Partner, a16z (a16z), with Alex Rampell (Alex Rampell) at a16z Fintech Connect, on the a16z channel, April 2026.
Karl Mehta61,387 görüntüleme • 13 gün önce
Mark Zuckerberg says he is not an AI researcher -- so here is the one number he obsesses over to win the superintelligence race: "I'm not an AI researcher. So it's not like I'm in there telling them what research ideas to do." "The things that I can do as CEO of the company are: I can make sure that we have the very best people and talent density. I can make sure that we have by far the highest compute per researcher, and that we do whatever it takes to go build out all that capacity." "We're building this Prometheus cluster, which is going to be the first gigawatt-plus single cluster, a contiguous cluster for training, coming online next year." "We're building a cluster in Louisiana that is going to be 5 gigawatt, and we're building several more clusters like that that are just going to be multiple gigawatts." "We're talking about many hundreds of billions of dollars of capital. So you both need to have a good business model that can support it, and you need to believe in it." The tell: the richest lab in the world is not optimizing headcount. It is optimizing compute-per-researcher, and betting hundreds of billions that whoever owns the biggest contiguous cluster owns the frontier. Talent gets you into the room. Watts decide who leaves with the model.
Karl Mehta255,587 görüntüleme • 2 ay önce
