Critics call out AI safety pundits whose falsifiable doom predictions were falsified

gayestgaymer1 · x · 2026-10-22

• An ongoing debate over AI safety predictions: critics argue claims that "unprecedented AI harms are crippling cyber infrastructure, caused by current open-weight models" haven't materialized, even though these were framed as falsifiable predictions.
• Critics say making falsifiable predictions was honorable, but refusing to acknowledge their falsification is not; others counter that such harm predictions were never really falsifiable and it's too early to declare them wrong.

Original post →