Infinite Computing Theory and AI Safety

geoffreyirving · x · 2026-07-08

The author argues that theories related to infinite computing limits—such as reflective oracles, logical induction, and AIXI—remain highly relevant.

They point out that rationalization isn't the only safety concern; the fastest path to a tractable theory of bounded rationality might begin by weakening certain infinite computing theories.

Related event: Geoffrey Irving: AI Safety Must Solve Post-Hoc Rationalization(8 posts)→

Original post →

More from AGI Musings

AGI Musings channel →