Scoring Show HN submissions for AI design patterns
Article
TL;DR
21% of Show HN submissions score 5+ AI design patterns; community debates whether that’s actually a problem.
Key Takeaways
- 21% heavy slop, 46% mild, 33% clean across ~500 Show HN submissions analyzed
- Colored left borders, Space Grotesk, gradient heroes are now reliable AI-gen signals
- Most ‘AI patterns’ predate AI — shadcn/TailwindUI normalized them first; methodology questioned
Discussion
Top comments:
-
[simonw]: Side projects are ideal for AI tools; AI-assistance shouldn’t be an insult
I expect most side-projects are being built with AI-assistance now. Side projects are typically time constrained - if AI saves you time, why wouldn’t you use it?
- [dang]: HN already built showlim to filter low-effort submissions; dip in chart explained
- [jerf]: 10k lines of AI code carries zero proof-of-work; 2016 standards don’t apply in 2026
- [nomdep]: These patterns were common before AI — blaming AI for pre-existing Bootstrap-era defaults