Video · 2025-10-31 · 1h 14m · 6 moments

All things AI w @altcap @sama & @satyanadella. A Halloween Special. 🎃🔥BG2 w/ Brad Gerstner

✦ AI generated

timeline · colored by role

Claim◆

OpenAI's restructuring created one of the largest nonprofits in the world, which is a bigger deal than Microsoft's equity stake.

Satya Nadella says the creation of the OpenAI Foundation as part of the restructuring is the real headline, even more than Microsoft's ownership stake, comparing it to the Gates Foundation.

transcript

Satya Nadella: what's pretty unique about OpenAI is the fact that as part of OpenAI's process of restructuring, one of the largest nonprofit gets created. I mean, let's not forget that... we are very proud of the fact that we were associated with the two of the largest nonprofits, the Gates Foundation and now the OpenAI Foundation.

Mechanism◆

Compute demand isn't fixed — it behaves like energy demand, where usage explodes as price per unit falls, so there's no such thing as a single 'enough compute' point.

Sam Altman argues that asking whether there's 'enough compute' is like asking about energy demand without specifying price — falling costs unlock huge new demand curves.

transcript

Sam Altman: If the price of compute per like unit of intelligence or whatever, however you want to think about it, fell by a factor of a 100 tomorrow, you would see usage go up by much more than 100 and there'd be a lot of things that people would love to do with that compute that just make no economic sense at the current cost, but there would be new kind of demand.

Fact

Microsoft's real supply constraint today is power and the ability to build out data centers fast, not a shortage of chips — chips are actually sitting in inventory unable to be plugged in.

Satya says Microsoft's bottleneck isn't a shortage of GPUs but the inability to get power and facilities built fast enough to plug chips in.

transcript

Satya Nadella: the biggest issue we are now having is not a compute glut, but it's a power and it's sort of the ability to get the builds done fast enough close to power. So, if you can't do that, you may actually have a bunch of chips sitting in inventory that I can't plug in. In fact, that is my problem today, right? It's not a supply issue of chips.

rebuts · 1supports · 2

Prediction◆

The cost per unit of AI intelligence is falling so fast (about 40x per year) that it's a scary exponent for infrastructure buildout, and like every past tech infrastructure cycle, some people will get badly burned by their compute contracts.

Sam Altman warns that the ~40x/year drop in cost per unit of intelligence is a dangerously steep curve for infrastructure investment, and predicts some people will get burned financially as has happened in every prior tech buildout cycle.

transcript

Sam Altman: if we can continue this unbelievable reduction in cost per unit of intelligence, let's say it's been averaging like 40x for a given level per year. You know, that's like a very scary exponent from an infrastructure buildout standpoint... Some people are going to get really burned like has happened in every other tech infrastructure cycle at some points along the way.

gives example · 1

Claim

A 50-state patchwork of AI regulation, like Colorado's AI act, is unworkable and a big mistake, and OpenAI genuinely doesn't know how it could comply with it.

Sam Altman says he doesn't know how OpenAI would even comply with Colorado's AI law and calls a 50-state regulatory patchwork a big mistake for the industry.

transcript

Sam Altman: I don't know how we're supposed to comply with that California, sorry, Colorado law. I would love them to tell us... I literally don't know what we're supposed to do. I'm very worried about a 50-state patchwork. I think it's a big mistake. I think it's there's a reason we don't usually do that for these sorts of things.

explains mechanism · 1

Mechanism◆

SaaS applications are being architecturally disrupted because the agent tier is replacing the old business-logic tier that used to be tightly coupled with data and UI, and low-ARPU/high-usage products like Microsoft 365 are best positioned for this shift.

Satya explains that AI is decoupling the traditional data/logic/UI architecture of SaaS apps, with the 'agent tier' taking over business logic, and argues Microsoft 365's low-ARPU, high-usage model positions it well for this transition.

transcript

Satya Nadella: the architecture of SAS applications is changing because this agent tier is replacing the old business logic tier... the way we built SAS applications in the past was you had the data, the logic tier, and the UI all tightly coupled... AI quite frankly doesn't respect that coupling because it requires you to be able to decouple.

explains mechanism · 2extends · 6supports · 1

Highlight slides

OpenAI Restructuring Creates One of the Largest Nonprofits✦ from: OpenAI's restructuring created one of the largest nonprofits in the world, which is a bigger deal than Microsoft's equity stake.Microsoft Now Linked to Two Giant Nonprofits✦ from: OpenAI's restructuring created one of the largest nonprofits in the world, which is a bigger deal than Microsoft's equity stake.Compute Demand Is Elastic, Not Fixed✦ from: Compute demand isn't fixed — it behaves like energy demand, where usage explodes as price per unit falls, so there's no such thing as a single 'enough compute' point.Price Drop → Usage Explosion✦ from: Compute demand isn't fixed — it behaves like energy demand, where usage explodes as price per unit falls, so there's no such thing as a single 'enough compute' point.AI Intelligence Cost Dropping 40x/Year✦ from: The cost per unit of AI intelligence is falling so fast (about 40x per year) that it's a scary exponent for infrastructure buildout, and like every past tech infrastructure cycle, some people will get badly burned by their compute contracts.Infrastructure Buildout Risk✦ from: The cost per unit of AI intelligence is falling so fast (about 40x per year) that it's a scary exponent for infrastructure buildout, and like every past tech infrastructure cycle, some people will get badly burned by their compute contracts.Historical Pattern Repeats✦ from: The cost per unit of AI intelligence is falling so fast (about 40x per year) that it's a scary exponent for infrastructure buildout, and like every past tech infrastructure cycle, some people will get badly burned by their compute contracts.AI decouples the traditional SaaS stack✦ from: SaaS applications are being architecturally disrupted because the agent tier is replacing the old business-logic tier that used to be tightly coupled with data and UI, and low-ARPU/high-usage products like Microsoft 365 are best positioned for this shift.Microsoft 365 is positioned for the shift✦ from: SaaS applications are being architecturally disrupted because the agent tier is replacing the old business-logic tier that used to be tightly coupled with data and UI, and low-ARPU/high-usage products like Microsoft 365 are best positioned for this shift.

Related episodes

Audio · All-In Podcast

SpaceX-Cursor Deal, SaaS Debt Bomb, New Apple CEO, SPLC Indictment, Colon Cancer Spike

→

Video · The Cognitive Revolution

Fable's Back, AI Engineer Recap, & SambaNova

→

Video · Machine Learning Street Talk

Intelligence is collective, not artificial — Prof. Michael I. Jordan (UC Berkeley / Inria)

→