build in public — shipping tiny ai products

i'm om — a 17 year old indie ai founder from india, building under the git11 label. this is the exact playbook i use to ship tiny ai products in public without a team, without funding, and without pretending to know what i'm doing.

1. keep the surface tiny

a tiny ai product does one job. one input, one output. no dashboard, no settings page, no login unless the product literally cannot work without it. if you can't explain it in one screenshot, it's not tiny yet.

2. ship weekly, not perfectly

i aim for one shipped thing per week. shipped = a real url a stranger can use. not a demo, not a screenshot, not a "coming soon". this is the whole engine of building in public — you need a shipped artifact to talk about.

3. post the ugly middle

the interesting part of building in public isn't the launch tweet. it's the half-broken screenshot at 2am, the bug you can't figure out, the metric that isn't moving. that's what people follow you for. launches are the reward, not the loop.

4. pick tools you can rewrite in a weekend

i pick boring, replaceable tools on purpose. a stack you can rewrite in a weekend means you can throw a product away without ceremony. that's how you ship 10 in a year.

5. treat the audience as a co-founder

when you build in public, the audience is your distribution, your qa, and your first customer. ask them what to build next. show them the roadmap. let them vote on pricing. this is the unfair advantage of a solo founder.

the 10-product goal

my current run is 10 tiny ai products shipped publicly. each one is small enough to fail without hurting. that's the whole point — a portfolio of small bets beats one big bet you're too scared to launch.