Richard Ngo warns against deference, linking it to AI safety derailment

RichardMCNgo · x · 2026-08-25

Richard Ngo argues against updating on others' views, stating it erodes independent thinking and leads to conforming to power structures. He cites how Effective Altruists derailed AI safety via this epistemic modesty. He advocates quarantining outside views and developing one's own inside view through action to avoid social pressure traps.

Related event: Richard Ngo: Deferring to Others' Views Is Derailing AI Safety Research(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →