A curated stream of model, platform, research, and product updates from the AI ecosystem, reviewed for engineering relevance before publication.
A Microsoft rollout study examines adoption, retention, and output signals for command-line AI coding agents.
A large-scale repository study finds AI coding agent traces across open-source work and warns that single-signal detection misses most adoption.
A study of Claude Code adoption reports changes in commits, repositories, languages, and developer behavior after AI-assisted coding.
A research review describes how agentic tools move software engineering from code completion toward supervised delegated execution.
A Claude Code architecture study breaks down permission modes, context compaction, MCP, skills, hooks, and agent loops.