If it's nearly free to build software now, it's nearly free to clone too. @zachlloydtweets, founder and CEO of Warp, works through what that does to a software business in "Software Engineering Is Becoming Factory Engineering," and @aiDotEngineer has it on YouTube. Zach spent over 20 years as an engineer, led engineering on Google Docs, and says he hasn't written a line of code in six months. The talk lays out what he thinks replaces hand-written code: a loop where ideas come in, agents triage and spec them, implementation and review happen with humans at specific checkpoints, and the whole thing feeds back into itself. - Open source as a moat, not a mission. Warp went open source a couple of months ago after five years building closed. It's over 60,000 GitHub stars and a couple hundred contributors now, with over 800,000 active developers using Warp, and Zach frames the move as a way to get distribution and ecosystem advantages a product alone can't buy once it's cheap to clone. - The factory loop, spelled out. Ideas in, an agent triages, a spec gets written for anything non-trivial (Warp splits this into a product spec and a tech spec), an agent implements, humans and agents both review, agents verify, a human reviews the product, you ship, then monitoring feeds straight back into the top of the loop. - as a public factory floor. Warp built a site showing every issue moving through the system and which agents or contributors are working it, as a live, if imperfect, version of the factory he's describing. - Skill loops that self-correct. Observer agents watch how a skill gets used and improve it over time. His example: a code review agent leaves comments, a senior engineer corrects them, and an observer agent uses that correction to improve the review agent for next time. - What this means for engineers. His read: you'll write less code but ship more product, and the job shifts toward building and tuning the system that builds the software, closer to process engineering than coding. - Advice for people entering the field. Adaptability, critical thinking, and the ability to still reason about underlying systems and architecture, since someone has to understand what the agents are writing. I'm working through the published talks from AI Engineer World's Fair sharing summaries and takeaways. Follow for more!