Software for the people the industry can’t afford to serve.

I started at Amazon as a catalogue specialist. The internal tools were slow and painful, and getting even small changes from the engineering team took weeks. So I grabbed an old Ubuntu machine, spun up a Node server, wrote some Chrome extensions, and built what my team actually needed. Our output went up about 50%. The official tools never caught up.

That experience stuck with me. Most groups are too small or too niche for traditional software teams to ever build for them properly. A twelve-person catalogue team, a local shooting club, a small classroom, or a niche hobby community - none of them are big enough to justify real engineering investment.

AI changed the math. One person with the right setup can now build and maintain software for thousands of people at a cost that used to require millions of users. I’ve spent the last few years building that setup. This site is what I’m using it for.

The plan is simple

Fund it by selling something the system is already good at (hardware-verified analysis of mobile apps). This keeps things independent.

Prove it by shipping real apps for small communities and showing the economics in public.

Open it so other people running small groups or projects can use the same system.

When I say “we” here, it’s mostly the system doing the work. I just decide what’s worth pointing it at.

Further out is the part I actually care about: autonomous AI running at the edge (on real hardware, off the cloud), doing work in places engineers can’t follow. Exploration, eventually. The factory is how I learn to build it.