Anthropic Says AI Is Writing Its Own Future, ChatGPT Can Now Dream, Rival Labs Sound the Alarm

Today we're covering the biggest AI stories of June 5th, 2026. Anthropic published a landmark report revealing that over 80% of its merged code was Claude-authored as of May — with engineers pushing 8 times more code per day than in 2024 — and laid out the recursive self-improvement scenario where each new version of Claude could be built by the version before it without any human involvement, while saying it would slow or pause frontier development if peer labs agreed to do the same. OpenAI overhauled ChatGPT's memory system with a feature called dreaming — a background process that turns past conversations into a continuously updated personal profile sorted by travel, hobbies, work, and more — with factual recall jumping from 41% to 82% and preference-following more than doubling. Plus, the CEOs of OpenAI, Anthropic, Google DeepMind, and Microsoft jointly signed an open letter to Congress warning that AI now outperforms PhD-level virologists and pressing lawmakers to require synthetic DNA sellers to screen every buyer before dangerous sequences can be traced back to bad actors — and today's community workflow comes from an anonymous parent of two toddlers who uses a scheduled AI prompt a few times a week to get fresh parenting tips, activity ideas, and behavior strategies — saying it's made him a better father.