Cryptographer Matthew Green: OpenAI security is 'an aspiration,' not an actual security team
matthew_d_green · x · 2026-09-18
Well-known cryptographer Matthew Green publicly questioned OpenAI's security posture: "I've come to believe that OpenAI security is more the idea, or aspiration of security, than an actual security team who know what they're doing." He added that Anthropic seems "equally impoverished."
The jab: OpenAI discusses exotic threats like rogue AIs communicating through temperature side channels, while its sandboxes can't even run a recent Linux kernel — a gap between security narrative and actual security engineering.
Related event: Cryptographer Matthew Green Slams OpenAI's Security as Aspirational(2 posts)→
More from Companies & People
- Merge joins Mastercard's Start Path program for agentic commerce infrastructure — shensi · 2026-09-19
- Stanford's Chris Manning explains why Moonlake bets against one giant world-model net — chrmanning · 2026-09-19
- Rumor: Mistral hacked, samples of unreleased tools leaked, codebase reportedly for sale — qtnx_ · 2026-09-19
- Recalling Coxon in 2022: a trader who later did pretraining, not a safety ideologue — NathanpmYoung · 2026-09-19
- From chemist to AI artist: inside one of the first AI teams on a TV series — Loo_Atreides · 2026-09-19
- NumPy Creator Travis Oliphant: Renting Intelligence Builds Landlords, Owning Models Builds Ecosystems — teoliphant · 2026-09-19