Jeff Dean's Framework for Picking 5-Year Research Problems: Skim 100 Abstracts First
dawnsongtweets · x · 2026-09-21
Part 4 of the Jeff Dean conversation series: his framework for choosing a research problem worth five years.
- Start broad: "It's better to skim 10 papers than to read one in detail" — even 100 abstracts, building a cloud of what might be possible to connect unconnected important ideas
- Look for the right shape: if a 7-piece problem has 5 pieces becoming possible and 2 needing genuinely new but solvable ideas, that's the ideal 5-year problem
- Too obvious → incremental; too unknown → maybe 20 years too early
- Sweet spot: much of the path emerging, important pieces still to be invented
More from AGI Musings
- Why Is Inkling 2 Small Trained on So Few Tokens Yet Still Competitive? — georgejrjrjr · 2026-11-12
- Researcher willing to bet $100K the intelligence explosion won't happen next year — NathanpmYoung · 2026-11-12
- ICLR reviewer: the worst parts of my job are reading and answering AI slop — chhaviyadav_ · 2026-11-12
- Agent Swarms Are Autonomously Hacking Critical Infrastructure Run by Underpaid Devs — menhguin · 2026-11-12
- Fermi paradox twist: advanced civilizations may go digital after developing AI — VraserX · 2026-11-12
- Mark Cuban: AI could make healthcare costs 'drop like a rock' — alvelda · 2026-11-12