The Three AI Pills: Why Most People Miss AI's Real Power
Author reveals why most people remain blind to what frontier AI can already do.
The essay proposes a simple framework for understanding AI disagreements: the Three AI Pills. Taking the first pill means accepting that AI can already do many things—coding, answering complex questions, and automating tasks—far better than most people realize. The second pill, AGI, anticipates AI doing much more in the near future. The third, ASI, holds that AI will eventually perform almost all tasks better than humans, within our natural lifetimes. The author declares themselves ASI pilled, and notes that many frontier lab employees and the labs themselves share this conviction.
Most people, the essay argues, haven't taken even the first pill. They remember old ChatGPT failures, dismiss AI as a stochastic parrot, and ignore current capabilities like coding agents. Even economists and policy experts underestimate how quickly AI will get cheaper and better, fixating on temporary weaknesses. The author's key advice: show unpilled people what AI can do today. The marginal cost of testing AI is near zero, and the practical benefits are enormous. For professionals and decision-makers, this argues for actively exploring current AI tools rather than relying on outdated assumptions.
- Three pills: AI (current capability), AGI (future expansion), ASI (surpass humans in our lifetime).
- Author identifies as ASI pilled, claiming frontier lab employees and labs share this view.
- Majority of people, including economists and policymakers, haven't taken the first AI pill, underestimating existing tools like coding agents.
Why It Matters
Professionals who dismiss current AI risk falling behind; understanding its real capabilities is now a competitive necessity.