From the Summary:

1. Flathub now allows AI-generated code if disclosed and labeled with the affected parts and extent.

2. Submitters must be human; agents can't generate commits, descriptions, reviews, or replies.

3. Maintainers can still reject AI-labeled or fully vibe-coded apps if they are low quality or problematic.

While, I am still annoyed about this shift of policy. I am glad they are going to be forced to disclose what parts of the code are affected by their lying prompt machines; that projects can still outright be rejected if it is pure slop, low quality or otherwise problematic. As unfortunately, the gap of projects that refuse to use LLMs is ever decreasing. I am so fucking tired of this shit. It is getting to the point where one can't even use basic dependencies because they have gacha prompt engines writing code for them. I wish the incentives for working in open source didn't enable the wildfire spread of gassed up LLMs. I just hope the fucking bubble bursts and we can move past this shit, get back to open source being about working with other humans to create awesome shit. Instead of letting the deskiller deluxe programs in on the action.