417 to 3: the House sends the power bill to the data centers
Anything drawing 100 megawatts or more would have to pay for its own new generation and transmission. The three no votes came from Democrats.
A model that doesn't talk: TypeSafe introduces Jev
The pricing page says 39 cents per thousand workflows, next to $3.31 for GPT-5.6 Luna. So far those numbers come only from the company.
Life Sciences Verification Program: Anthropic opens the biology blocks for vetted labs
The vetted organisations decide for themselves what counts as safe use for their teams — written at the level of detail of a job listing, and explicitly without any of their own IP.
Claude Desktop 2.2553: cloud sessions finally show you their files
Admins get two new levers on top: an allowlist for plugin MCP servers and a domain policy for the built-in browser.
Projects in Claude Code: one coordinator, many threads
Each thread is its own cloud session on its own branch. When two of them touch the same code, it reaches you as an ordinary merge conflict.
Jacob Coxon had fewer than a hundred followers — then he posted in Slack
His short stint at Anthropic earned him no equity. He still holds OpenAI shares. And three of the six people on his old maths olympiad team now work at AI labs.
Pew asked 37 countries about AI and jobs — in 34, worry wins
Among Germans who have heard a lot about AI, 26 percent are more concerned than excited. Among those who have heard little, it's 41.
A group meets almost daily inside the White House — and still can't move Trump
Anthropic's Mythos set it off: after the spring preview, officials were on round-the-clock calls for days. Internally the model was known as the red flashing light.