Warp outlines shift to software factories with dual engineering roles

Software engineering is evolving from manual coding and local agent steering to a "software factory" model where agents handle autonomous work across the lifecycle. Warp identifies two distinct responsibilities for engineers in this new paradigm: product engineering, which involves defining, steering, and ensuring the quality of the product, and factory engineering, which focuses on optimizing the agent systems themselves.

Product engineering remains human-centric, leveraging human taste and context to determine what to build. However, it shifts from local, private development to public, cloud-based workflows using tools like Slack and Jira. While some tasks (like simple bug fixes) can be fully automated—starting at around 20-30% for Warp—the goal is to minimize "human touches per PR" over time.

Factory engineering involves treating the agent infrastructure as a system to be debugged and optimized. Engineers use metrics such as DORA metrics and LLM-as-a-judge scorers to improve skills, context, and model selection. Warp integrates this responsibility into performance reviews, emphasizing that optimizing the factory is a core engineering skill. The company's platform, Warp Factories, supports this transition by allowing teams to build and measure their own workflows on open infrastructure.