A $2,500 contest for redesigning the NeetCode site became an accidental experiment in AI-generated low-effort work.

What happened

  • Navi offered $2,500 to whoever submitted the winning redesign; results so far: "a little bit disappointing" — practically all submissions were obviously AI-generated (which he says is fine in itself) 57:57
  • The failure mode: when asked why they moved/removed buttons or content and what the pros and cons were, submitters couldn't answer, or gave vague answers. "Me looking at their site for 5 minutes, I can articulate things about their design better than they can" 58:20, 58:44
  • His diagnosis: not intelligence — effort, caring, and skill set 58:59
  • His view of UX: it's not prettiness; nobody cares how pretty a site is if they don't understand what it's for or what value it gives them 59:15, 59:37

The incentive problem (host's counterpoint)

  • A contest flips incentives: with N contestants, expected value per submission drops (10 contestants → ~$250 expected), so rational effort is one AI prompt 59:37, 59:55
  • Navi's lesson learned: he should have picked a small pool of people, paid them up front, and chosen from that work 1:00:22

The deeper point

  • Pen and paper + articulated intent would beat a raw prompt: on his own site, even AI-coded parts, he can justify every positioning choice with metrics ("this is used the most, so I want to make this prominent") 1:00:45, 1:01:05
  • Low-effort prompting produces low-effort output — the differentiator is being able to defend your choices 1:01:27