CONCEPT
Intelligence as Compute
Sam Altman's foundational thesis that intelligence scales approximately as the logarithm of the computational resources used to produce it—converting the most contested philosophical question of the era into a capital allocation problem and the organizing premise of a half-trillion-dollar industrial commitment.
The most consequential sentence Sam Altman has written is not about safety or governance or the future of work. It is a quiet technical claim buried in his February 2025 essay “Three Observations”: The intelligence of an AI model roughly equals the log of the resources used to train and run it. If true, this converts the most contested philosophical question of the modern era—what is intelligence, and is it the kind of thing that scales?—into a predictable engineering curve. It is the premise on which the scaling laws bets made by every major AI laboratory rest, the premise that justifies the concentration of capital in frontier model development, and the premise that, if falsified, would render the financial architecture of the entire AI boom mispriced by an order of magnitude. The pairing of this claim with a second observation—that the cost of a fixed capability level falls by approximately ten times per year—produces
Keep reading with YOU ON AI
Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.