Warp

Confidence 0.75 · 2 sources · last confirmed 2026-11-04

Warp is a developer-tools company that began with an AI-assisted terminal and in 2026 sells Warp Factories, a cloud product for running coding agents as a software factory: work enters from Slack, Linear or GitHub, agents triage, specify, implement, review and verify it, and the factory is defined in code. Its founder and CEO is Zach Lloyd, formerly a principal engineer who ran engineering for Google Docs. Warp publishes talks and demos on its own YouTube channel.

Promoted to an entity page on 4 November 2026, when the Warp channel became the author: of a second source page.

Appears in this wiki via

• 2026-10-19-lloyd-warp-what-it-takes-to-build-a-software-factory — What it takes to build a software factory (19 Oct 2026). Lloyd on factories as open infrastructure defined in code, and factory engineering: metrics, LLM-judge scoring, self-improvement and internal benchmarks. Warp’s own figures: cost per pull request from about $55 to about $15 in six weeks.
• 2026-11-02-holmes-warp-agentic-self-improvement-loop — How to build an agentic self improvement loop (2 Nov 2026). Benjamin Holmes, DevEx engineering lead, on the outer loop: store conversations, score them with rubric judges, isolate failures, let an agent propose skill changes. The open-source Skill Doctor skill packages the method. On screen: 13 of 53 proposed changes approved, and $2,755.94 in scorer cost to date.

Reading this channel

Warp sells the product its talks describe, and both talks end with a call to request access. The methods are stated in general terms and one is released as open source, which makes them checkable; the figures shown come from Warp’s own factories and are not independent.