the pitch
software, even mini-apps, is heavy. you have a five-minute idea and suddenly you're thinking about auth, hosting, branding, retention, and a domain name. that's insane. when you have a thought, you tweet it. you don't open a blog. when you have a funny video, you post a tiktok. you don't start a youtube channel.
but when you have an idea for something that should exist as software, the lightest option is still an app or website. the easiest creation tools in history are at our fingertips, and every time they generate something, they flip you from a person with an idea into a developer maintaining an app.
so drips asked a different question. what's the smallest unit of an app that still feels like software? the answer is a session. you play it for thirty seconds. it ends. it has a feedback loop. anyone can build one from any ai they're already chatting with.
the experiment
launched march 9, 2025. shipped 60 drips in the first three days. the question was "how light can the unit of software get before it stops being software?" not retention, not monetization, not a startup. a posting-velocity experiment.
the answer it found: very light. the format works. the friction is almost entirely on the creation side, not the consumption side. that insight is now showing up in everything else i'm building.
by the numbers
- 60 drips posted in the first 72 hours after launch
- full source now public on github
why it's archived
the experiment ran its course. running an interactive supabase-backed app that nobody was paying for was costing roughly $30/month, and i had already learned what i needed to. rather than maintain infrastructure for an experiment whose hypothesis had already been tested, i archived it and moved on.
i'd rather ship six experiments cheaply than maintain one forever.