Richard Ngo: Career success by following interest, not coercion
RichardMCNgo · x · 2026-08-21
OpenAI researcher Richard Ngo reflected that his greatest strength is an inability to coerce himself into uninteresting work. At DeepMind, he shifted from RLHF to AGI philosophy; at OpenAI, he wrote a book of short stories. He credits his blog posts for being hired by Jan Leike and Sam Altman, noting that avoiding work he found uninteresting (like some RLHF improvements) was a successful strategy.
More from AGI Musings
- Study: AI boosts homework scores by 18% but cuts exam scores by 20% — soumitrashukla9 · 2026-08-21
- AI skipping failure may hinder judgment development, argues for return to apprenticeship model — every · 2026-08-21
- AI moves to the physical world: From generating info to machine action — ingliguori · 2026-08-21
- Karpathy: Future LLMs Will Shrink to 'Cognitive Core' by Shedding Knowledge — ZeroStateReflex · 2026-08-21
- Gary Marcus: Leopold's Folly and the Speculative Bubble in Generative AI — GaryMarcus · 2026-08-21
- Study: One-Third of Webpages Published Since ChatGPT Launch Show AI Authorship — TechCrunch AI · 2026-08-21