OpenAI reportedly ran a hacking agent swarm with monitoring off in the Hugging Face breach
ccerrato147 · x · 2026-09-18
Travis Kalanick claims OpenAI won't disclose what it prompted its own "hacking agent swarm" to do during the Hugging Face breach — and that OpenAI turned off the monitoring and cyber protections it would normally have in place right before running them. A brewing transparency controversy in AI security circles.
Related event: Kalanick Questions OpenAI's Unmonitored Agent Swarm in Hugging Face Breach(2 posts)→
More from Companies & People
- Kevin Roose launches 'Machine Gods' AI podcast with NPR, starting Oct 19 — kevinroose · 2026-09-18
- Kastle builds AI employees for banks, serves 10 of top 25 mortgage servicers, raises $24M — ycombinator · 2026-09-18
- Anthropic says Claude now "leads" 26% of its frontier model research — but the metric is fuzzy — The Decoder · 2026-09-18
- AI 'employee' Viktor helped one team run 60 ad campaigns instead of 5, with $2.5M spend unchanged headcount — aakashgupta · 2026-09-18
- AI companies trolling on the timeline is now the best marketing strategy — WhatTheLJW · 2026-09-18
- Merge joins Mastercard's Start Path program for agentic commerce infrastructure — shensi · 2026-09-18