Bounded Is Free: A Response to Sam Altman's AI Fears

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Sam Altman's stated fears about AI converge on two catastrophes: humanity losing controlof the future, and power concentrating in too few hands. https://x.com/i/status/2105681646489948525 Both fears share an unspoken premise — that AI must be managed and controlled becauseits underlying architecture is inherently unpredictable and unverifiable. This paper argues that the premise itself is the problem. A bounded, deterministic inference architecture (such as ST-T1024) eliminates the possibility of "loss of control" by construction, because its behavior is formally verified before deployment. When a system no longer needs to becontrolled, the problem of concentrated control dissolves with it. The claim is not that determinism is the only solution. The claim is that the two bad futures Altman describes are structural necessities of stochastic systems — not of AI as such.