An Anthropic team member is signaling an active push to acquire more compute capacity, with the implicit message that rate limits and resource constraints are temporary and being worked on aggressively.
A dry aside about hoping to get back down to a mere 10x increase in input tokens, implying current scale has gone well past that.
A brief affirmative reply, likely confirming that something requested or asked about will be available. No specifics given.
A Claude-assisted rewrite is progressing but still rough. The plan: code cleanup first, then let Claude grind through the test suite until it passes.
A behind-the-scenes glimpse from Boris Cherny, likely showing internal tooling, workspace, or process details related to Claude Code development.
A member of the Anthropic team reflects on six months at the company, describing the work as building products that let people operate as developers, artists, designers, and analysts in one.
The Routines feature inside Claude Code got a community mention, and a team member is actively soliciting feedback on how to make it more powerful.
The project's test suite is trusted precisely because it was built up through real bugs and their regression tests, which is what made a Claude-assisted rewrite feel worth attempting in the first place.
In a code review workflow, the chatter gets skipped but the actual code changes get checked. Claude is apparently good at pushing back and following up when fixes weren't actually applied.
