Ex-MIRI researcher shares 2016 talk showing she once backed task-directed AGI
jessi_cata · x · 2026-11-22
AI safety researcher Jessica Taylor pushed back on claims about her 2016 views by linking a talk she gave while employed at MIRI, in which she said "Building task-directed AGI is a good idea, and we can make progress today studying how to do so." She argues it wasn't mainly her idea to say such things, in an exchange with @wolftivy and @ohabryka about MIRI's historical stances on task AGI.
More from AGI Musings
- "Interacting with the archive of all human knowledge is real sci-fi" — AI circle resonates — repligate · 2026-11-22
- AI Progress on Riemann Hypothesis Spurs Math Interest; Lean Creator on Formal Verification — bigdata · 2026-11-22
- Welcome to the other side of the singularity: the intelligence explosion meme — rand_longevity · 2026-11-22
- Gary Marcus Fires Back: I Admit Wrong on What I Got Wrong, Not Else — GaryMarcus · 2026-11-22
- Dev Veteran: In the AI Era, the Cleverer the Hack, the Faster It Dies — cocktailpeanut · 2026-11-22
- Gary Marcus mocks the biggest data center ever, tied to OpenAI, SB Energy and Nvidia — GaryMarcus · 2026-11-22