The incentive for sending in patches was that you needed to fix a bug in software you were using for your own use case, so you might as well send the authors the patch, and they'd keep maintaining it all for you. But if you can vibe-code your own solution, you don't need to use somebody else's software, so the patch doesn't get made, much less submitted.

The other thing is, LLMs tend to generate terrible code that pisses off open-source maintainers. So I'm not sure even LLM-made patches will make it into open source much.

This might be the death of traditional open source. Vibe-coded-only open source may be the next generation. Which I'm fine with, as long as we can start regulating software, so that vibe-coded tools are banned for safety/privacy uses unless they follow a software building code.