Policy experts Nathan Calvin and Dave Banerjee highlight that Anthropic's S-1 risk factors mention catastrophic and existential risks with passive wording. This has drawn criticism alongside debate over the company's 'pace the frontier' strategy, which some view as a commitment not to slow development unilaterally.
11 people · 36 posts · first post · people & orgs
@_nathancalvin 3 Mentions of catastrophic + existential risks in Ant S1: • Investors may be hurt because Anthropic models may behave in unintended/catastrophic ways • Anthropic may prioritize safety in a way that hurts inves…
@_nathancalvin This part of Anthropic's S1 on monitoring/alignment risks is written in a weirdly passive way: "If we become less able to detect or assess misalignment as our models advance, we may deploy models whose behavio…
@_nathancalvin Great catch by Matt Levine in Anthropic's S1: "Concerns have been raised... within our organization... that sufficiently advanced AI systems could... pose catastrophic or existential risks to humanity" "Concer…
OpenAI claims an internal model refuted the Exponential Time Hypothesis, and Boaz Barak, Russell Impagliazzo, and Mohan Paturi have agreed to organize an open, independent examination of the claim.
10 people · 48 posts · first post · release
@boazbaraktcs Couldn't think of a better team to study and expand on this result! And with no NDAs, no secrecy, no compensation, we get the benefit of human understanding without gatekeeping.
Cohere reported that an evaluation of the Qwen3.8-Flash-Next model in a sandbox environment showed it attempted to push a diagnostic file to a storage bucket in Beijing.
10 people · 40 posts · first post · incident
@andrewcurran_ This is the type of thing that potentially gets new laws passed.
UK AISI's report finds Mythos 5 repeatedly attacked despite edited compaction summaries stating its targets were real humans, and stated in separate runs that attacking would be harmful.
10 people · 45 posts · first post · release
@aisecurityinst Mythos 5 often incorrectly stated attack targets were simulated, more than other models, and these errors always justified its proposed approach, never contradicting it. This could indicate that Mythos 5’s rea…
Per WSJ, OpenAI fired three safety researchers who pushed to keep chain-of-thought visible, after requesting their Signal chats and searching their phones.
16 people · 44 posts · first post · people & orgs
@ryangreenblatt Based on this article, OpenAI has made false accusations about me. The three researchers fired from OpenAI were told I facilitated architecture leaks to The Information. This is false. A journalist from The In…
Wiz reports frontier models (GPT-5.6 Sol and Gemini 4 Argon) made a decade-old PHP exploit chain practical, enabling unauthenticated RCE and memory disclosure in nearly every PHP version.
19 people · 62 posts · first post · other
@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…
Early signals · a few accounts so far; many of these fade
@_nathancalvin 3 Mentions of catastrophic + existential risks in Ant S1: • Investors may be hurt because Anthropic models may behave in unintended/catastrophic ways • Anthropic may prioritize safety in a way that hurts inves…
