things i got wrong about building hobitat

2026-07-29

hobitat was meant to be a social network for hobbies — a place to pick up something new and slowly get good at it, in public. i built it around three layers: a private layer for your own practice, a public layer to share it, and a reputation layer to reward getting better over time.

every social product has the same chicken-and-egg problem: no content without people, no people without content. so i tried to build hooks. i reached out to friends and made small challenges that break a hobby into one step a day. read a book: day one, share a photo of the cover. day two, post your favorite line from what you read. day three, rephrase it in your own words and add an opinion. one tiny action at a time.

it didn't catch. people didn't engage. and here's the thing i got wrong — i thought the problem was onboarding, or notifications, or the size of my network. it wasn't. i even tested it on myself with a crypto-trading-learning habit, and i couldn't stay consistent either.

hobbies are sparse by nature. i'll do something today, then pick it up again a week later. that gap is fatal for a social app. i added reminders and notifications, but when i reopened hobitat after four or five days, i'd simply forgotten i was ever supposed to post there. the app assumed a rhythm that hobbies don't actually have.

so that's the mistake, plainly: i built a consistency-shaped product for an inconsistency-shaped behaviour. i don't have the fix yet. maybe hobbies need something that survives the gaps instead of punishing them.

i've open-sourced the whole thing — app and web. poke around, break it, steal ideas. if you have a take on the sparseness problem, or want to collaborate, i'd genuinely love that. it's all on my github.