Mind-blowing AI Video: OpenAI launches Sora + Is Biden too old? Tucker/Putin interview & more

The episode opens with a house tequila project whose best sample, per its taster, beat a benchmark bottle that retails for about $3,000. It then turns to the AI trade.

The lead AI-chip company, whose name the transcript garbles, is cited at about $1.8 trillion, the fourth most valuable company, behind Saudi Aramco. Its stock is up 6.5x in five years and rose on a report of custom chips for Amazon, Meta, Google and OpenAI. The host adds that the Financial Times put OpenAI's run rate at $2 billion, about $165 million a month, and that Google rebranded its AI suite as Gemini at $20 a month.

Friedberg says opinions differ on whether productivity gains match the compute, but spending continues because the infrastructure must come first.

you've gotta build out the infrastructure

He likens it to building a new internet. AI's core workload is matrix multiplication, which needs special-purpose chips, and the ceiling is hard to bound.

Arm's valuation is said to have doubled within weeks, then fallen to about $140 billion from roughly $156 billion. SoftBank bought the Vision Fund's 25% last August at a $64 billion valuation, owns 90%, and holds a stake of about $125 billion after paying $32 billion for the whole business seven years earlier. Critics had called its founder a one-hit investor (Alibaba), yet the AI bet now looks prescient.

The defense is the power law: one winner can erase portfolio losses. Another host says no credible investor called him a fraud, and survival matters most.

You only look like a moron to losers

A caveat: only 10 to 15 million of roughly 145 to 150 million Arm shares float, and about 12.5 million were short before earnings. Part of the jump may be a short squeeze, as with GameStop.

Sacks calls him a gambler who needs one big return, and doubts $500 million checks into effectively seed-stage companies were wise. Another host notes that big winners such as DoorDash were not seed bets.

The hosts contrast scaling businesses such as DoorDash with early companies that took huge rounds and struggled to deploy them.

If you don't have a business that can handle the gas you're putting in the tank, the engine's gonna blow up.

One host dates a "take the market" doctrine to about 2009 and recalls being pitched to raise more than he wanted. He says most such businesses could not turn the money into market share. Another cites Katerra, which raised about $900 million and which he believes ended near zero.