Venkatesh Rao: 'EA promises to solve AI Safety. Now we have two problems'

akbirthko · x · 2026-09-19

Venkatesh Rao publishes 'EA Safety' on Contraptions, tracing his 20-year fraught relationship with the rationalist/EA community — from the 2006-07 Overcoming Bias era to its growth into an institutional landscape that shapes AI's future, with himself labeled 'postrationalist' by Scott Alexander.

His core thesis, per the title: EA's promise to solve AI Safety has created a second problem — the EA formation itself. The essay examines how a blog scene became a power structure with its own gravitational field over consensus and history.

Related event: Venkatesh Rao's Long Critique of EA and AI Safety(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →