John Schulman Releases Inkling 2 Small With Detailed Training and Alignment Notes
ziqiao_ma · x · 2026-11-06
Former OpenAI/Anthropic co-founder John Schulman announced the release of Inkling 2 Small, with a blog post detailing the training process. He highlighted the 'Safety and alignment' and 'Training Inkling 2 Small' sections as the most interesting part of the release, noting his team encountered unpleasant surprises when early checkpoints behaved unexpectedly. The post drew excited reactions from peers.
More from Models
- Ricursive open-sources Ripple-65B Preview: 12T-token model scores 80.52 on TerminalBench 2.1 — annadgoldie · 2026-11-06
- Users claim Claude performs better with an empty system prompt than Anthropic's official one — repligate · 2026-11-06
- Palisade's Jeff Ladish: models better at everything but still not funny or persuasive — JeffLadish · 2026-11-06
- Palisade's Jeff Ladish calls to revive METR time-horizon estimates for new models — JeffLadish · 2026-11-06
- Leaker DM claims Qwen4 is dropping, but even the timeline is a guessing game — bclavie · 2026-11-06
- Artificial Analysis launches Endpoint Accuracy Index v1.1: nearly 1 in 4 serverless endpoints score below reference — ArtificialAnlys · 2026-11-06