Researchers exploited a flaw in OpenAI's community forum, hosted by third-party Discourse, to gain access to internal sign-ons and eventually an OpenAI employee's ChatGPT account with GitHub access. OpenAI confirmed it had fixed the issues after being contacted. Anthropic published data showing a rapid increase in AI usage, reporting that 26 percent of research and development work was "led by" its Claude model, up from 1 percent in March. The company stated AI systems are increasingly being used to build the next version of themselves, approaching recursive self-improvement. Anthropic shared the data to help the public understand how close the world is to reaching this threshold, where AI can train and improve itself. On 90 percent of tasks, AI collaborates with a human and performs large chunks of work, though models did not yet operate fully autonomously for the research studied.

OpenAI declined to comment further. Anthropic and Hacktron did not immediately respond to requests for additional statement.