“Models keep getting more impressive at the rate that the short timelines people predict, but more useful at the rate that the long timelines people predict.”
@Dwarkesh Podcast
AI revenue stays orders of magnitude below AGI-pricing, so the capex buildout decelerates and its chief beneficiary derates.
“But people are trying to launder the prestige that free-training scaling has, which is almost as predictable as a physical law of the universe, to justify…”
@Dwarkesh Podcast
RLVR has no reliable scaling trend, so reasoning-model moats are overhyped and OpenAI's valuation gets repriced down.