Simon Späti · 2026-09-26 · notable
Simon Späti — the problem isn't AI code, it's that nobody knows anything
Data engineer Simon Späti argues that AI writes roughly average code, and the real problem is teams that no longer understand their system's architecture or intent. The note reached 292 points on Hacker News.
Knowledge loss, not code quality, is the cost of letting AI do the thinking.
What is it?
In 'The Problem is not the AI Code, but Nobody Knows Anything Anymore', Simon Späti accepts that AI probably writes average code and can lift a weak codebase up to average. His concern is different: 'nobody knows anything, and everyone just asks Claude. You end up with no plan whatsoever.'
How does it work?
Späti builds the note from voices he quotes. One engineer describes a company where specs, code, tests and PRs are all made by Claude Code and people work 12 to 13 hours a day 'just to press enter'. Hoyt Emerson argues data people always had to know the business from day one, so AI only removes friction for them; Späti replies that newcomers who 'prompt away' lose that advantage.
Why does it matter?
Späti calls maintainability 'the final boss': the easier it is to generate a pipeline, app or dashboard, the more there is to maintain, and that gets hard when nobody understands it. He argues humans are still needed to direct and orchestrate AI, which makes intent, taste, design and architecture key skills, and suggests teams keep hiring junior engineers.
Who is it for?
engineering and data teams