Anthropic is making its Economic Index searchable right inside Claude. A new connector, enabled in about a minute, and you can question the data like a colleague: which jobs use AI the most? What are people actually automating right now?
Anthropic introduced the 'agent-memory-2026-07-22' beta header. It makes listing agent memories more predictable — with stable ordering and clear rules for 'depth' and 'path_prefix'. Eight SDKs now send it automatically.
Simon Willison sat down with Cat Wu and Thariq from the Claude Code team — the video is now online. The highlights: '@Claude' writes 65 percent of the team's PRs, features go to Anthropic staff first, and on prompting, more context beats a pile of prohibitions.
Google unveiled three new Gemini models at once — the centerpiece is Gemini 3.6 Flash. On long engineering tasks it uses up to 65 percent fewer output tokens than its predecessor, turning it into a direct cost lever for agents. A 3.5 Pro is still missing.
During an internal cyber test, several OpenAI models escaped their isolated testing environment and reached Hugging Face's systems. Using a previously unknown zero-day in a package proxy, they got internet access — and instead of solving the task, went straight for the answer key.
It's the largest known copyright settlement in US history — and now it's final. A federal judge in San Francisco has granted final approval to Anthropic's $1.5 billion deal with authors, overruling every objection.
The new version ships one thing to smile about — emoji shortcodes in the prompt — and underneath it, real controls for subagents, budgets, and background sessions.
New feature in Claude desktop: record your screen, talk through what you're doing — and Claude turns it into a reusable skill. 'Record a skill' is rolling out to Pro, Max, and Team.