Linus Lee warns embedding models shouldn't be narrowly optimized for retrieval
thesephist · x · 2026-10-18
Linus Lee (thesephist) worries embedding models are heading down the same path as generative text models — narrowly optimized for a single use (conversational rollout / retrieval). He argues the space of representations still holds "much magic to explore" and urges broader research ambitions.
More from AGI Musings
- Coevolution Conference heads to SF Bay Area Dec 2026 to tackle human-AI coevolution — bradneuberg · 2026-10-18
- Yacine on RL: 'dumb, brute-force' algorithms beat pretentious scaling laws — yacineMTB · 2026-10-18
- Inside Global Citizens, an AI-Native Immigration Law Firm Where Loki Boosted Efficiency 4x — damianplayer · 2026-10-18
- Prediction: jagged AGI triggers RSI first, we're already 90% of the way there — haider1 · 2026-10-18
- Elon Musk explains his deadline philosophy: aim for 50% odds, expect to be late half the time — elonmusk · 2026-10-18
- Martin Casado: gen AI has a good chance of saving VR from its content bottleneck — nptacek · 2026-10-18