Rating Data Reveals the "AI Design Smell" in HN Submissions

Someone analyzed the rating data of projects submitted to HN and found that projects bearing the hallmarks of AI-generated design score systematically lower—"design slop" is being recognized and penalized by the community.

What the Analysis Found

The author collected the landing pages of a large batch of Show HN projects, tagged the typical hallmarks of AI-generated design—that familiar gradient palette, cookie-cutter layouts, rounded-corner cards with emoji, plausible-sounding marketing copy—and cross-referenced them against community ratings. The result is clear: the stronger the hallmarks, the lower the score. "AI smell" has become a quantifiable liability; seasoned users lower their expectations at a glance, and that distrust carries over to the product itself.

A Reminder for People Building Products

What's interesting is the causal chain: the low score isn't necessarily because the design is ugly—AI's default aesthetic is actually pretty tidy—but because it becomes a signal that "the author didn't put in the effort." When generating a landing page takes only ten minutes, a landing page no longer conveys any information about investment; instead, the roughness and quirks that carry a human touch have become a scarce proof of sincerity. The discussion's conclusion is solid: drafting with AI is fine, but before you ship, at least edit out that default smell—otherwise you lose the first impression before you start.

via: Hacker News