Read the original at Lenny’s Newsletter ↗lennysnewsletter.com · subscriber post
By Lenny Rachitsky · lennysnewsletter.com · @lennysan on X · YouTube · LinkedIn
Lenny Rachitsky and Kyle Poyar compiled lessons from more than 20 AI product builders on what surprised them about shipping AI features. The post argues that AI products need fresh thinking: prototype to discover what is possible, treat trust and UX as core, rely on proprietary data, and optimize speed and iteration over flashy features.
Subscriber post — summary only01Key takeaways
- Start AI projects by asking what is technologically possible and prototyping, since feasibility and quality are unclear upfront.
- Demo appeal does not equal user value; validate with real users over longer periods to see past novelty-driven churn.
- Segment users by attitude toward AI, since skeptics and embracers behave very differently and averaging them hides both.
- Good UX and onboarding teach people how to use AI, and small, invisible improvements often outperform big AI features.
- Proprietary, licensed data and speed (such as precomputed results) are durable advantages as base models commoditize.
“Demo value isn’t user value.”Joshua Xu · Lenny’s Newsletter
“The smallest (and almost invisible) features are usually the fan favorites.”Claire Vo · Lenny’s Newsletter
02Frameworks mentioned
Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.