A music social network MVP, at half the estimate
Built the MVP for Catnip, a music-focused social app, in half the estimated time and cost by working AI-assisted from the start.
The problem
KMGMT wanted an MVP for Catnip, a music-focused social network: the fastest possible route from idea to something real people could hold and react to.
MVPs have a particular failure mode. Build too carefully and you spend the budget before you learn anything. Build too loosely and you produce something that cannot be extended if it works. The estimate matters, but so does what you are left holding afterwards.
What I did
I built the MVP as sole developer, working AI-assisted from the outset rather than bolting it on.
- Delivered in roughly half the originally estimated time and cost.
- Used AI tooling for the parts where it genuinely compounds (scaffolding, boilerplate, exploring unfamiliar APIs) while keeping architecture decisions and the parts that carry risk under human judgement.
- Because I understand architecture and can work across the stack, I could stand up backend pieces myself that would previously have meant bringing in a second developer.
The outcome
A working MVP, delivered at half the projected cost.
This is the honest version of what AI has changed about my work. It has not replaced the judgement about what to build or how to structure it: that is still the expensive part, and it is still where projects go wrong. What it has done is widen the range of things one experienced developer can credibly take on alone.