Jonathan Blow – If AI solved coding, why is my software still the same? [video]

366 points · 500 comments on HN · read original →

Points and comments are a snapshot, not live.

Despite claims that AI solved coding, software quality and complexity remain unchanged.

Jonathan Blow and Wolf Tivy discuss why AI hasn't improved software quality, arguing that the explosion of complexity in code is a human and organizational problem, not a coding one. Blow contends that today's AI cannot produce better code than people because it optimizes for speed and volume, not sound design or maintainability. He suggests that if AI truly solved coding, software would be simpler and more reliable, yet it remains bloated and fragile. The conversation points to systemic incentives favoring feature delivery over long-term durability.

What commenters are saying

Many commenters agree that software quality has declined, not improved, with some pointing to recent high-profile outages as evidence. A common sentiment is that coding was never the bottleneck; organizational culture, communication, and incentives are the real constraints. Two camps emerge: those who see AI as exacerbating complexity by enabling faster production of sloppy code, and those who find AI transformative for personal projects or bypassing inefficient teams. A top comment notes that AI solves "how to ship faster," not how to build better software.