2h agoGoogle DeepMind's Séb Krier warns AI models warp epistemic humility and that automation complexity is underestimatedSeán Ó hÉigeartaigh also warned of pathological model over-trust.SentimentSentimentPos72.7%Neg27.3%Praise went to the policy expert's changed stance on AI automation pace and AGI scaling limits for reflecting intellectual accountability, though the permanent underclass meme drew criticism for being unpersuasive.11 comments with sentiment. View comments.