ChatGPT just got a health upgrade that beat real doctors on a benchmark 👀 OpenAI's new GPT-5.6 scored 60.5 on HealthBench Professional — outscoring the 43.7 average from 260 real physicians across 60 countries and 26 specialties. It's graded on physician-written rubrics for safety, clari

OpenAI is literally paying people $50,000 to try to break ChatGPT's safety rules 👀 They just turned their "Bio Bug Bounty" into an ongoing private program and doubled the top reward to $50K — for anyone who finds a reusable "universal jailbreak" that defeats GPT-5.6's bio

ChatGPT just clocked in. 👀 OpenAI launched ChatGPT Work — a new agent (powered by GPT-5.6 + Codex) that doesn’t just answer questions, it actually does the work. It moves across your apps and files, stays on a project for hours, and hands you finished docs, spreadsheets, slides, dashboards, even

Claude just asked the whole internet for its hardest questions 👀 Anthropic (the team behind Claude) launched "Inviting Hard Questions" — a short film + open call asking people everywhere to send their toughest questions about AI: jobs, family, society, and where all this is actually head

Claude just dropped its own “Wrapped” 👀 Anthropic launched Reflect — a monthly recap that shows exactly how you use Claude: your peak day + hour, total conversations, and the tasks you hand off most. It even added wellness controls like quiet hours and break-time nudges. 🧠✨ Basically Spotify