RL pioneer John Langford joins Core Automation, a lab built around agent-driven research
JohnCLangford · x · 2026-11-12
RL researcher John Langford announces he's joining Core Automation after Microsoft — a self-described "world's most automated AI lab" betting that the next wave of frontier research will come from nimble teams with highly capable agents.
Key points:
- First job is automating its own work to make room for more ambitious research;
- Believes the next step change won't come from scaling the current recipe, but from new learning algorithms superseding large-scale pretraining/RL and architectures that scale better than transformers;
- Team includes people who pioneered reasoning models, led OpenAI's RL research org and Gemini's pretraining, created the model behavior discipline, and co-authored FlashAttention-4;
- Hiring across research, systems, and lab automation, plus an internship where small teams build businesses with agents.
More from AGI Musings
- "Hard to tell a 92% model from a 99% model": the bottleneck is evals — chris_j_paxton · 2026-11-12
- A 4-bit sparse flop and a synaptic firing both cost ~30 femtojoules, but the brain has far more connections — MikePFrank · 2026-11-12
- Economist on AI plagiarism dispute: an innocent author would publish their Google Doc history — paulnovosad · 2026-11-12
- Researcher mourns loss of top open-weight advocate, fears regulatory capture — KevinKaichuang · 2026-11-12
- Radiologist tests Fable 5.2 vs experts: about a second-year resident's level — JHochderffer · 2026-11-12
- Developers predict AI model plateau: the real gains will come from inference optimization — bendee983 · 2026-11-12