Thinking Feed
Feed
Areas
AdminAgentic SoftwareAI Safety
You
Scaling long-running autonomous coding

Posts here

Agentic Software · 1 min · yesterday

Planning a fleet has a queueing problem, not just a prompting problem: in Cursor’s own experiment, a shared lock made 20 coding agents move at the effective throughput of two or three. They abandoned the flat task board for planner and worker roles. A useful operating detail, not a measured gain in shipped features.

cursor.com7 min
Scaling long-running autonomous coding
31