AI coding gave us a false sense of speed. It showed how fast AI scales when deployment friction is zero.
But for most enterprise workflows, just defining 'what is a correct output' (evals) requires bespoke back-and-forth. The data isn't universal, it's messy and siloed. That's why adoption is slow. Every enterprise needs heavy lifting and integration, not just a turnkey model.
Infinite AI demand is colliding with finite corporate budgets. Right now, AI is not baking a new economic pie; it is just aggressively redistributing the existing one. This minimum token spend is essentially a mandatory tax just to stay in the game. Eagerly waiting to see the true builders who will bake the new pie.
Suspect we will start to hear about a “Pareto optimal” balance of computationally efficient humans, cheaper open-source tokens and frontier tokens.
Our internal AI spend @Atreidesmgmt will be roughly 100x higher in August 2026 vs. March 2026. Still roughly doubling every month. Note that is before Grok Bot moves to consumption pricing which will likely create a step function when it happens (at least for me).
AI means that scale has become more important to investing imo.
I think there will be a minimum token spend required to be competitive in most knowledge based industries.
The “compute inequality” referenced by Sholto in our discussion.
@GavinSBaker@Atreidesmgmt The minimum token spend is just the new rent. We are in a massive phase of wealth redistribution toward infrastructure, not new wealth creation. The Pareto optimal balance is just companies trying to survive the squeeze.
FDE 热潮在我看来恰恰佐证了将AI在coding的adoption线性外推到其他产业的这条曲线不会那么 smooth。
AI 的下一阶段,瓶颈至少不再只是 capability scaling,deployment scaling 正在变成同样重要的约束。
这就是我现在对 AI 中期最大的 uncertainty:Can deployment scale as fast as compute capacity is being built?
如果可以,现在 FDE 只是一个过渡层,未来会逐渐被产品化吞掉,今天巨大的 CapEx 很可能有需求承接。
如果不可以,FDE 会变成一个庞大的 services industry。AI 仍然改变世界,但企业 adoption 比资本市场预想的慢很多。