Stanford and NVIDIA's CLM picks actions by embedding, skipping generation: 44 ms vs 579 ms
Prompt Engineering · youtube · 2026-09-27
CLM from Stanford and NVIDIA Research takes a different approach from generative System 1 models like Jev: it never generates an answer. Instead it embeds the situation and every possible action into one space and picks the closest one. Since action embeddings are computed once and cached, scoring 1,000 options costs about the same as one — 44 ms vs 579 ms for Jev. The catch: options are encoded independently, so they're never compared side by side. Full demos ran on a DGX Spark, with blog and open-source code available.
More from Models
- Leaked: OpenAI's GPT-6.2 (codename betelgeuse) reportedly weighs in at ~15T tokens, drops Thursday — iruletheworldmo · 2026-11-11
- Running DeepSeek 4.1 locally on an $800 rig with a used RTX 3090 at 2-3 tok/s — LankyGuitar6528 · 2026-11-11
- Google open-sources Harness², an agent that rewrites its own setup per task — Arindam_1729 · 2026-11-11
- Dev dumps Codex desktop app over quality issues, falls back to codex-cli — tekbog · 2026-11-11
- Why do old models get dumber before launches? Lab party, says joking X post — glcst · 2026-11-11
- Grok Bot now detects hidden prompt injections, warns users and refuses to follow them — XFreeze · 2026-11-11