AI Researcher on the Job Market Reflects on the Art of Problem Selection
simi_97k · x · 2026-10-28
Simran Khanuja, an AI researcher on the job market, reflects on how her approach to choosing research problems has evolved since 2017: resisting keyword-based self-labeling (evals, alignment, post-training), favoring problem-driven work, and arguing that problem selection is becoming a defining skill in an era of fast execution.
More from AGI Musings
- Purdue's GPS-free drone tracks runners through forests; Rachel Thomas asks if it should exist — mmitchell_ai · 2026-10-29
- Europe's share of global tech R&D collapsed from 23% to 8% since 2005 — NinaDSchick · 2026-10-29
- Europe's chip gap: the case for Eurofab, an RSI-ready EUV fab on European soil — S_OhEigeartaigh · 2026-10-29
- 'We generated a proof but haven't checked it' is the new 'downloaded the PDF, haven't read it' — miniapeur · 2026-10-29
- Neurovium: a proposed unified discipline of neural computation and cognition — neurovium · 2026-10-28
- Could iMessage Agents Be Apple's Next App Store-Scale Business? — signulll · 2026-10-28