Blow asks: If AI codes, why hasn't software improved?
Original: Jonathan Blow – If AI solved coding, why is my software still the same? [video]
Why This Matters
The gap between AI coding hype and real software quality is a live industry debate.
Game developer Jonathan Blow challenges AI coding hype in a video, questioning why software quality hasn't visibly improved if AI tools are genuinely solving programming at scale.
In a video posted to YouTube, indie game developer Jonathan Blow — known for Braid and The Witness — raises a pointed skeptical question about AI coding tools: if these systems are as capable as claimed, why does the software people actually use feel no better than before? Blow has previously been a vocal critic of declining software quality and the tendency of the tech industry to chase hype over craft. His argument implicitly targets tools like GitHub Copilot, Cursor, and the broader wave of LLM-based code assistants that have been marketed as transformative productivity multipliers. The core tension he identifies is straightforward — extraordinary claims about AI-driven productivity should produce observable results in shipped software. So far, that evidence is thin. The video appears to be a talk or commentary rather than a structured essay, consistent with Blow's previous public appearances where he critiques industry trends from a craftsman's perspective.