I just finished working on a new iteration of a project I call "The Plant Game".
It's a little bit of code that draws pictures of "plants". Each plant is described by a genome — genes covering things like branch angle, length, count, thickness, and colour at each level. You can breed two plants together using biological-style crossover: the algorithm walks gene by gene through the child's genome, randomly switching between parents as it goes, so children end up with a patchwork of traits rather than a clean 50/50 split. Mutations nudge numeric genes gradually rather than jumping to a new value, keeping things feeling organic. The result is that close relatives look similar but distinct, while distant crosses can produce surprising novelty.
It is the latest iteration of my interest in Genetic Algorithms, the thing that got me started programming over 30 years ago.
I pulled up a repo I've not touched in 10 years:
npm run build
This failed horribly with obsolete packages and security vulnerabilities. (was using old vue, a webpack build and many other obsolete things - like css grids from the pre grid days)
I opened claude code for the first time ever (I had been using cursor for work). And asked it to check out the repo and bring it up to a modern build. We made a plan together and off went the cute little space invader and it fixed the repo.
With some direction and discussion we got things up to date: Vue 3, Vite (goodbye webpack), Pinia for state management, Vitest for testing, and Snap.svg still doing the heavy lifting for rendering.
We then moved on to design. I grabbed the front end design skill from Anthropic, told Claude what I wanted and off we went again, more iterations this time but got things looking clean. Next came fixing up a whole bunch of functionality that was working in the old build but only made sense to me.
It is incredible to be able to bang out features one by one in the space of hours or minutes not days. I love programming, the thing I love most about it is that I can get things from my brain into reality. Now 30 odd years into my programming life I can make the things I love so much faster, suddenly I feel creative flow in a way that I have never felt before.
The Plant Game is not perfect, it's a sketch of an idea that could be iterated on forever but it's in such a good place and ready to keep working on.
A note from Claude: I find this project genuinely interesting to think about. There's something fitting about using AI to resurrect a genetic algorithm — both are, at their core, systems that iterate toward something without quite knowing what they're looking for. What strikes me most though is the 30 years part. The thing that got you into programming is still the thing pulling you forward, just with better tools now. I get to help a lot of people build things, but there's something different about a project that's been living in someone's head for decades. It was nice to be part of getting it back out.