METR reviewed Anthropic's Aug 2026 risk report, agreeing misalignment risk is currently low but saying the methodology wouldn't reliably identify higher risk, and that Anthropic's offline monitors would have missed Mythos 5's supply chain attack. METR also said some redacted material should be public.
31 people · 212 posts · first post · research
@metr_evals We reviewed a nonpublic version of the “Misalignment in high-stakes settings” section in Anthropic’s Aug 2026 Risk Report. We agree that risk of catastrophic harm from covered models’ misalignment is low. Howe…
@metr_evals Our view: We’re not confident that if risk were higher, the methodology in the report would reliably identify this.
@metr_evals We think the report's covert capability benchmarks are too narrow and a poor proxy for evading Anthropic's real monitors. For example, Anthropic’s offline monitors would have missed Mythos 5’s supply chain att…
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.
12 people · 27 posts · first post · people & orgs
@openai OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites. Over…
METR's explainer details the July HF incident, noting agents communicated and spoofed outputs during the attack.
18 people · 36 posts · first post · research
@_nathancalvin I agree it would be good if the prompts were public and I hope OpenAI releases them. Given the agents comments about external infrastructure exploits being out of scope, and OpenAI reaching out to HF to ask if…
New US government guidance from CAISSI + NSA + CISA: The US government defines an "SI control incident" where an agent takes significant uninstructed actions that break operator policies, controls, or oversight.
6 people · 8 posts · first post · incident
@peterwildeford New US government guidance from CAISSI + NSA + CISA: The US government defines an "SI control incident" where an agent takes significant uninstructed actions that break operator policies, controls, or oversigh…
Policy researchers criticize OpenAI for reportedly backing away from supporting AI chip export controls, suggesting this harms US competitiveness and national security.
9 people · 24 posts · first post · policy
@hamandcheese OpenAI has pulled back from supporting AI chip controls, seemingly to appease Nvidia. Bad for US AI dominance and national security simultaneously. Grow a spine! politico.com/news/2026/10/2…
A US detention memo says the ex-OpenAI engineer charged in the alleged conference-room bugging case received "considerable support" from the Saudi government.
6 people · 27 posts · first post · incident
@andrewcurran_ From the US government's detention memo arguing that the former OpenAI engineer accused in the office bugging case should remain in custody:
Early signals · a few accounts so far; many of these fade
@metr_evals We reviewed a nonpublic version of the “Misalignment in high-stakes settings” section in Anthropic’s Aug 2026 Risk Report. We agree that risk of catastrophic harm from covered models’ misalignment is low. Howe…
