Claude Code 2.1.266, shipped Sept 9, reverts a 2.1.265 regression where CLAUDE_CODE_USE_GATEWAY alone forced Cloud-gateway sign-in and broke API-key sessions with "Not signed in to the Cloud gateway." Updating is the only fix; no config change needed.
Cloudflare added support for running Cursor Cloud Agents inside customer controlled Sandboxes, so Self Hosted Machines workers execute terminal, filesystem, and browser tools in an isolated Cloudflare account while Cursor still runs the agent loop. It joins similar Cloudflare execution hooks for Devin Outposts and Claude Managed Agents…
Bloomberg reports Anthropic finished diligence then dropped a roughly $6B deal for Decart AI, whose chip efficiency stack was meant to squeeze more training and inference from existing compute ahead of Anthropic’s IPO. The sides may still pursue a commercial partnership instead of an acquisition…
Jacob Coxon quit Anthropic on Tuesday, saying OpenAI and Anthropic are racing toward self improving superintelligence and "gambling with our lives." Alignment Science lead Evan Hubinger agreed labs earnestly fear existential risk, putting his own odds above 10% within a decade and saying Anthropic still lacks a clear plan for superintelligence alignment…
Former Chrome and Google Cloud AI developer experience lead Addy Osmani joined Anthropic to work on Claude Code, aiming to improve how developers supervise the agent in production. The hire lands after Anthropic said Claude Code passed $2.5B run rate revenue by May 2026…
At Goldman’s Communacopia, OpenAI CFO Sarah Friar said the company is pushing task specific AI into chip design, life sciences, and financial services, and testing outcome based pricing against open weight rivals. She cited OpenAI’s own models helping tape out its Jalapeño inference chip…
Spotted something we missed? Start a thread.
