Is Harness Engineering Software's Last Breath?
Pieter Levels says YC's batch proves software is mostly dead. But a harness is software. A two-question test for what AI will absorb, and what it can't.
European Space Agency opens role for a Harness Engineer on July 1, but the job has no link to AI. Harness refers to the bundle of wires that carries power and signals between spacecraft components, preventing expensive machines from talking over themselves. Two months later, Pieter Levels analyzed Y Combinator's latest group of startups and observed an abundance of harness and hardware startups, leading him to conclude that software is mostly dead.
Hardware has always required harnesses, and in the current Silicon Valley context, harness essentially represents software. The crucial question isn't whether software survives the harness era; rather, it's which aspect of software will endure. While the author co-founded Dojo Coding that builds agent harnesses for companies and trains developers, they argue that most harness code isn't worth building and focus on the claim that nearly all software is dead.
Levels' post is composed of four logical sentences, suggesting that the most forward-thinking place in tech has stopped building software. The claim is debatable, as the most early-adopting location in technology has stopped creating software, possibly because frontier labs offer generic harness solutions. Therefore, Levels' assertion that software is mostly dead and hardware reigns supreme holds merit.
Written by urgent.news from HackerNoon's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
