Ask HN: How do I contribute to OSS without giving training data to AI companies?
Hey everyone,
I recently created a solution to do line-to-line occlusion for usage in computer graphics and specifically plotter art. After publishing the results, I have received inquiries about open sourcing my code. At face value, my reaction is an enthusiastic yes. If I can help contribute to the plotter art community by way of releasing this, then I am all for it.
I began to consider the implications from a different perspective though. This problem currently cannot be solved by AI or vibe coded. It's still within the domain of a human to create, and I find this valuable. If I opensource it, I'll be implicitly contributing my knowledge to every AI company that scrapes Github, NPM, etc. and I find myself massively out of alignment with their objectives. Sam Altman is a grifter and a creep.
So, my question is, how can I prevent AI training on a publicly available OSS project? Until courts start ruling that copyright applies to LLM scrubbing I think we're all sadly shit out of luck. API + controlled access. That's not open source though... I would like to release this with a permissive license for humans only. From what I've read, the jury is still out on having your license copied left to the AI companies even when they're training on your information.