Ex-OpenAI researcher says Anthropic breach gave attacker full Slack admin access to research secrets
Background-Wafer-548 · reddit · 2026-10-29
Former OpenAI researcher Will Depue claims Anthropic is downplaying a recent security incident: an attacker dubbed "Cathedral" allegedly obtained full administrator access to Anthropic's Slack workspace, where nearly all research secrets live. He likens it to OpenAI's 2023 hack via a stolen access token from an external app service, and mocks labs for keeping crown jewels in Slack with admin API tokens everywhere. Unverified claim.
More from Companies & People
- Stanford no-tech classroom rule made the same lecture take 50% longer to cover — Afinetheorem · 2026-10-29
- Harvard AI Safeguards Professor Stephen Casper Hosts Reddit AMA on Safety and Governance — HarvardKennedy · 2026-10-29
- Tatsu Hashimoto on teaching DL: teach the standard mental model, then let students break it — srush_nlp · 2026-10-29
- Qualcomm's Dragonwing IoT Day: Societal Acceptance Will Size the Physical AI Market — seankinneyRCR · 2026-10-29
- DeepMind's talent drain: mid-level researchers leave, directors spin out billion-dollar labs — zacharynado · 2026-10-29
- Inside Google's AI Strategic Partner program: what perks can you actually extract? — Cloud_Architect_J · 2026-10-29