Nat Lambert: narrow RL training on internal infra is a poor fit for open-weight models

mervenoyann · x · 2026-10-27

AI researcher Nathan Lambert commented that the simplest explanation for a model's behavior is narrow RL training reliant on its internal infrastructure — a design he argues isn't good for an open-weight model, since external developers can't replicate those training conditions.

Related event: Nathan Lambert: Cohere Model Trained via Narrow RL on Internal Infrastructure, Bad for Open Weights(2 posts)→

Original post →

More from Models

• GPT-6.1 Terra cuts factual error rate from 22.3% to 4.3%, an ~81% drop — haider1 · 2026-10-27
• GPT-6.1 Terra lands: early users praise speed, value and subagent performance — BenBajarin · 2026-10-27
• Leaked GPT-6.1 Terra benchmarks: cheap but 'boring', still far behind Astra and Claude — Angaisb_ · 2026-10-27
• Artificial Analysis Ships Intelligence Index v4.3.2 With GPT-6.1 Terra Comparison — ArtificialAnlys · 2026-10-27
• GPT-6.1 Terra cuts hallucination rate by 50 points while lifting accuracy 2 points — ArtificialAnlys · 2026-10-27
• GPT-6.1 Terra scores 48 on AA Coding Agent Index at one-third the cost of GPT-5.6 Terra — ArtificialAnlys · 2026-10-27

Models channel →