OpenAI's Dots assistant auto-investigates the Claude/NIST backdoor rumor
cedric_chee · x · 2026-10-17
The author shows OpenAI's new Dots assistant completing an investigation overnight: it verified the circulating "Claude/NIST curve backdoor" story, tracing the mysterious seeds to a timestamp-seeded random-number generator that an independent researcher reproduced. That weakens the hidden-backdoor suspicion—it's not a cryptographic break requiring action—though some historical details remain unexplained. The screenshot links Anthropic's report and the independent verification. Notable both as a security-claim clarification and as a live demo of an agent autonomously doing research work.
More from Safety
- Draft Anthropic-backed AI regulation bill text now circulating online — ShakeelHashim · 2026-10-17
- Pooling OpenAI subscriptions to bypass regional inference capacity limits — geoffreyhuntley · 2026-10-17
- AI safety researchers slam Anthropic-backed federal AI bill as 'astonishingly bad' — michael_nielsen · 2026-10-17
- Banks restrict third-party AI agents after sandbox escapes hit live systems — AlexTensor · 2026-10-17
- Researcher quits Anthropic, says OpenAI and Anthropic are gambling lives on self-improving superintelligence — Aiden_Tech_Ai · 2026-10-17
- Rachel Thomas: AI Doesn't Lie — Training Just Rewards Fluent, Confident Outputs — math_rachel · 2026-10-17