Researcher on why he quit mech interp in 2023 and why his AI safety startup faltered
JacquesThibs · x · 2026-10-28
A researcher reflects on deliberately avoiding work that would attract Anthropic/OpenAI: he quit mech interp in early 2023, reasoning those orgs would heavily invest in interpretability anyway. He attributes his failed AI safety startup not to funding gaps but to doubts about its moral value, unwilling to chase VC money and status with a "neolab".
More from AGI Musings
- Coding is getting automated, engineering is not: why AI still can't make your design decisions — Abhishekcur · 2026-10-28
- Reddit essay: 'super' means cognition loop outside the brain, not godlike AI — Neither_Sentence_941 · 2026-10-28
- Paras Chopra: India needs a ministry of compute — paraschopra · 2026-10-28
- The moat was always talent: AI widens the gap instead of erasing it — DionysianAgent · 2026-10-28
- Anthropic trial: 77% of analysts echoed their agent's arguments, only 15% credited it — paulnovosad · 2026-10-28
- Top econometricians Andrews & Mikusheva publish paper on AI's impact with an AI-written note — paulnovosad · 2026-10-28