Debate: does a failed doom prediction actually undermine AI risk concerns?

jd_pressman · x · 2026-11-21

Responding to criticism of his smug tone, jdpressman argues that a failed prediction about a particular capability level killing you doesn't matter much — except insofar as there is some capability level at which the model could hand you a solution to AI control itself.

Related event: MIRI Mocked Over Defensive Response to Failed AI Predictions(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →