TechCrunch

The New York Times and additional news organizations filed a sanctions motion against OpenAI on Thursday in a copyright infringement lawsuit that is two years old. Court records show OpenAI conducted internal searches of its training data for journalistic works prior to the suit and maintained a database of approximately 78 million de-identified ChatGPT conversations. A data privacy engineer underwent a court-ordered deposition in April and was later compelled to a re-deposition.

Elon Musk publicly stated that Anthropic currently leads in AI with its Mythos and Fable models and that he would not cut off access in ways that would harm the company. Reports indicate Anthropic signed a multi-year power purchase agreement with a SpaceX-linked data center following the xAI merger, while earlier export controls on the models were later lifted. Musk had previously posted that Anthropic could not win in AI competition.

Meta discontinued a feature in its Muse Image AI model that permitted users to generate images by @-mentioning public Instagram accounts. The company announced the removal in a Friday blog post after an earlier rollout that week and stated the tool had missed the mark following feedback. Two sources, TechCrunch and The Verge, reported the development.

China conducted a test of an experimental net-based recovery system for the Long March 10B rocket booster on a sea platform. Multiple sources confirm vertical return and capture roughly six minutes after separation, with the vehicle delivering a satellite to orbit. Disagreements exist over recovery platform details and the milestone's place in global reusability history.