Hacker News new | ask | show | jobs
AI Risk Benchmark: GPT-5 Leads, but Misalignments Persist (substack.com)
4 points by m1chael3ma 300 days ago