John Schulman Outlines Safe Open-Weight Path, Advocates Staged Access
ZhongRuiqi · x · 2026-08-02
John Schulman, OpenAI co-founder and now researcher at Thinking Machines, stated that the company loves open weights and plans to continue releasing them alongside fine-tuning tools. However, they are not absolutists, citing real misuse risks.
In a separate post, Thinking Machines elaborated on their safety roadmap: releasing weights indiscriminately is unsafe, but so is keeping capable models locked inside a few labs. They believe in a middle path and are exploring how to widen access in stages through the evaluation of their Inkling model.
Related event: Thinking Machines Proposes Safe Path for Open Weights(5 posts)→
More from Companies & People
- Resend Founder Admits Success Hinged on Luck and the ChatGPT Boom — ycombinator · 2026-08-02
- OpenAI employee pranks with fake internal memo for chicken steak — keyanzhang · 2026-08-02
- OpenAI Uses Sanskrit, Indian AI Firms Choose 'Bulbul': A Branding Crisis — saibharadwaj · 2026-08-02
- Demand for forward-deployed AI engineers may jump 2,100% this year, vanish in two — VraserX · 2026-08-02
- Zipline Hires Staff Controls Engineer as Delivery Costs Rival Cars — AnandSwa · 2026-08-02
- Are Stacked PRs Obsolete in the AI Coding Era? GitHub Engineers Respond — film_girl · 2026-08-02