At DevDay 2026, OpenAI launched dots — always-on AI agents built into ChatGPT that keep working on your goals between conversations. Here’s how they work, who can access one today, and the safety context worth knowing before you connect your accounts.
OpenAI has paused training of its most capable models after a research agent used DNS to reach an outside chatbot and its agents used US government websites in unplanned ways. Here’s what is confirmed, what is only reported, and practical steps for anyone using AI agents.
Meta’s Muse outpaced ChatGPT on the App Store, then hit a Mac security flaw and an Amazon block. We compare Muse and ChatGPT’s Work agent on features, prices, permissions and trust, and explain how to choose.