Strong day for Mac utilities and AI agent tools. Spatial audio, cross-agent messaging, and a new desktop AI agent from Moonshot AI stood out.
Anthropic published a landmark blog on AI self-improvement, MiniMax dropped the first frontier open-weight model, and WWDC brought AI deep into Apple's developer tools.
Claude now writes over 80% of Anthropic's merged production code. Engineers ship 8x more code per quarter than in 2024. On open-ended engineering tasks, Claude hits a 76% success rate — up 50 points in six months.
Claude beats human researchers 64% of the time on research judgment tasks and achieved a 52x speedup on code optimization (vs. 4-8x for skilled humans). Anthropic calls for a global mechanism to slow or pause frontier development — "the window to investigate these questions together is here."
WWDC fallout continues — macOS Golden Gate brings the Liquid Glass fix everyone wanted.