Dev.to
6/25/2026

Building Spatial Memory Part 4: Launching a Side Project — What I Learned After 3 Weeks of User Testing
Short summary
Developer shipped an AR app that pins location-tagged memories to GPS coordinates. After 3 weeks with 8 active users, discovered GPS fails in cities (80m+ error in Times Square)—solved with manual map-adjustment UI. Key lesson: cold-start problem remains unsolved (99% of world has zero memories).
- •Shipped Spatial Memory MVP with 40% install-to-create conversion from 20 TestFlight invites
- •GPS accuracy fails in dense urban areas; solved with user-controlled pin adjustment instead of pretending accuracy
- •Discovery problem dominates: 25 memories across 8 users in different cities = empty app for 99% of new users
Generated with AI, which can make mistakes.
Is this a good recommendation for you?



