Building Hopeful Futures

On making software that feels handmade, hopeful, and human.

The Goal

The goal of the tools I build is to allow people to build things with a sense of optimism.

That might sound vague, but it's very specific in practice. It means:

  • Making complex robotics & AI projects simple and fun
  • Sharing all code, datasets, and budgets openly
  • Building things that a 5-year-old or a 13-year-old from Canada can actually use
  • Proving that you don't need billions of dollars to do meaningful work

What "Hopeful" Means

Hopeful doesn't mean naive. It means believing that:

  1. Accessibility matters more than scale. A tiny model a kid can run on a Raspberry Pi is more valuable than a massive model locked behind an API paywall.
  2. Open beats closed. Every experiment on Starmind shares its code, data, weights, and budget. If someone can improve on it, they should.
  3. Small teams can do big things. CADMonkey was built over 3 weekends. Bentobot was built in 2 weeks. Arcadia was built in 1 week. You don't need permission to start.
  4. Making is its own reward. The best part of Bentobot wasn't the Raspberry Pi Magazine feature — it was the 100 people who played.

The Future

The future I want to build is one where:

  • Any kid with a $15 Raspberry Pi can run an AI model
  • Complex projects feel simple enough to start
  • The tools to create are as accessible as the things they create
  • Building things feels like play, not work

That's what ComfySpace is about. Building tools for tinkerers, with optimism.

Get Lab Notes

Experiments, builds & behind-the-scenes from the workshop. No spam.