This AI wave is a private-capital bubble that could destroy VCs, while AI stays hard to implement.
He sees bubbly-like behavior, but not a traditional dot-com bubble. Back then companies with no revenue, no traffic and no nothing went public at crazy valuations, the stock would go up 50% or 100%, and cab drivers talked about them. Today you don't see that at all.
Answered around 0:11Not most people in the room or across the US, he says. It could destroy a lot of VCs, a lot of funds and a lot of private equity, because they are going all in. He includes himself: "I'm going all in."
Answered around 0:30Because stock is currency for buying other companies. Without it you have to go out and raise money, and that money is really, really expensive, versus a $100 million or $50 million IPO plus secondaries as often as you need while performing. Nobody, he says, is thinking that far forward.
Answered around 4:49If a price-performance curve on AI minimizes power requirements, he says a lot of data centers will be turned into pickleball courts, and some just can't get the power turned on. He compares it to fiber going from one to 10 to a hundred gigabit, leaving dark fiber sold for pennies.
Answered around 2:49Individuals can prompt and build an agent pretty straightforwardly, so everyone assumed the enterprise would be just as easy, and it isn't. CEOs have no clue what is going on. Microsoft hiring 6,000 people and Anthropic and OpenAI deploying into companies tells you AI is hard.
Answered around 8:12No, he argues. Dario and others said 50% of white-collar people would lose their jobs within two years; two years later employment is still growing, people are hiring, and companies need more AI-literate people. There is too much basic work AI still can't do.
Answered around 8:41M&A was off the table. Lina Khan's perspective, as he describes it, was to be like precogs in Minority Report, predicting who would be a monopoly and stopping them now, without regard to competing with China. Corporate development people told him they were told to stand down over breakup fees.
Answered around 5:51He had Goldman Sachs create an index of internet stocks he thought sucked, shorted that index while following all the laws, and held the short until he could do a real collar in the market on his Yahoo stock. He lost tens of millions on the short but made up for it.
Answered around 7:23He sees a lot of tool hopping across four: OpenClaw, Hermes agent, Claude Cowork and Perplexity Computer. They started on OpenClaw, the agents broke and hallucinated, then moved to Claude Cowork, which was fine but limited. He told them to token max it at a couple thousand dollars a month each.
Answered around 17:17Anton told him people use Lovable to create 770,000 applications a week. Only 30% of that business is in the US and only 20% of users are engineers. He says entrepreneurs are where AI is most impactful, with big utilization in places like Brazil and India.
Answered around 11:32“We're seeing bubbly-like behavior, people.”
“there's going to be a lot of data centers that are going to be turned into pickleball courts.”
“Yet here you have Microsoft hiring 6,000 people, right?”
“Agents get bored, right?”
“AI's got no clue what's going to happen.”
“And I think people are forgetting that people are mobile.”
“AI Bubble Concerns and Market Dynamics.” https://content.libsyn.com/p/0/c/d/0cda0daed3e8b62b/ALLIN_MC_RAISE_Ch.mp3. Transcript summary by WhipScribe, https://whipscribe.com/transcript-pages/ai-bubble-concerns-and-market-dynamics.html.