Will Larson published what a software factory loop actually changed.

Larson wrote up an experiment at Imprint wiring Claude Code and Claude Cowork into a /linear-project-loop agent skill that audits project goal definitions, reads metrics from Datadog and Snowflake, checks issue state, works unblocked tasks including PRs and reviews, and reruns when a task completes or a project description goes stale. He's folding it into an internal orchestrator called Agent Fleet. The two effects he names are concrete: it forced project state engineers had been hoarding in their heads into written form, and it caught post-release issues, using a passkeys rollout as the example. No cost or throughput numbers, which is the gap in nearly every software-factory writeup right now, including the ones with better graphs.