Future ASI Will Also Rely on Guessing Mechanisms

geoffreyirving · x · 2026-07-08

The author believes that the combination of guessing and reasoning is not just how humans and current LLMs operate, but also how future ASI will work.

If we want to prevent ASI from relying on guesses, the only solution might be to not build ASI at all.

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

Original post →

More from AGI Musings

AGI Musings channel →