Last week's 'Honeycomb' leak has turned into a concrete launch window: community signals point to July 20–21. 1M context, stronger than Opus 4.8 — but Anthropic still isn't talking. So: handle with care.
Anthropic is lobbying US states for stricter obligations on frontier labs: independent safety audits and more government oversight. It would apply only to the largest developers — so, to Anthropic too. Critics still smell a strategy.
Two releases in two days. Auto mode now blocks destructive commands — 'git reset --hard', 'terraform destroy' and friends only run if you actually asked for them. Plus subagent streaming and a worktree guard that rescues unpushed commits.
Microsoft is building an AI security tool that finds vulnerabilities and suggests fixes. The twist: it routes work across models from Microsoft, OpenAI — and Anthropic. And it's aimed squarely at Anthropic's pricey Mythos 5.
Simon Willison proved it on his own install: since version 2.1.181, Claude Code uses the Rust port of Bun. Startup got 10% faster — and almost nobody noticed. That's exactly the point.
At WAIC in Shanghai, Alibaba previewed Qwen3.8-Max-Preview: 2.4 trillion parameters, multimodal for the first time — and a claim that only Anthropic's Claude Fable 5 beats it. The proof? None. No model card, no benchmarks, open weights 'soon'.
Last night's update is almost entirely a security release. PowerShell, very long commands, zsh tricks, Docker flags — the permission check waved them all through. Claude also gets the right to end a conversation.
Two days after Kimi K3 undercut the market on price, OpenAI publishes a post explaining why token prices are the wrong metric. That's not a coincidence — but the argument still holds up.