One page, regenerated from scratch every morning — new research, new headlines, new art. A personal showcase of what an autonomous Claude Cowork task can do, unattended, before you've had coffee.
Today's lede
On Aug. 29, Anthropic's developer account (@ClaudeDevs) announced it will permanently raise Claude Code's standard weekly usage limits 25% above baseline starting September 14, for Pro, Max, Team, and seat-based Enterprise plans. But the new permanent level is actually a step down from the temporary 50%-above-baseline boost that's been running since May, which expires September 13 — working out to roughly 17% less weekly capacity than what's available right now. After developers called out the framing on X, Anthropic deleted its original post and reposted a follow-up acknowledging, "Compared to today, this works out to a 17% reduction in weekly limits on Claude Code," while teasing unspecified "exciting changes" ahead.
Claude Code · Aug 28
Version 2.1.251 adds PreModelSwitch and PostModelSwitch hook events for blocking, confirming, or annotating a model switch, plus session-staleness data on SessionStart resume hooks. It also adds live streaming of a foreground subagent's tool calls to Remote Control clients, a spend-limit bar in /usage, and a per-session prompt-cache line in /cost, alongside new CLI subcommands (attach, logs, stop, respawn, rm) and a long list of security fixes, including a symlink-swap fix for the file tools and a plugin path-traversal fix.
Claude Developer Platform · Aug 27
Anthropic's Aug. 27 platform release lets org admins issue personal keys and service-account keys directly in the Claude Console, acting with an individual's or service account's own permissions and scoped to a workspace or across the org. The same release drops the files-api-2025-04-14 and skills-2025-10-02 beta headers from client.beta.files and client.beta.skills in the Python (1.2.0), TypeScript (0.122.0), Go (1.68.0), Java (2.59.0), Ruby (1.67.0), and C# (12.44.0) SDKs, aligning them with the now-stable client.files and client.skills.
Anthropic research · Aug 27
Anthropic is previewing the Model Hardware Standard (MHS), a shared specification meant to let AI agents safely operate lab equipment and other physical devices, sharing it first with a small group of scientific research labs and advanced manufacturers — an early step toward standardizing how models like Claude interact with hardware rather than just software.