Rather than banning AI tools in development, the correct approach is to change how submissions are handled in open source projects.

What kind of recommendation do you have here that doesn't devolve to, like, closed source?

Open source requires a community that is transparently involved in the process and have equal rights to fork the code. There's really no way around this, the license grants the right for anyone to fork as long as they adhere to the license.

Sad truth is that AI has effectively killed open source unless maintainers are willing to make a clear stance for AI:0 submissions only. Otherwise, the endless pull requests are far too hard to maintain as a single individual. And we certainly don't want code being accepted by AI systems without human certification, especially in something as fundamental as an OS kernel.

Translation "Please don't investigate how we make the AI sausage. You won't like what you see and, legally, it puts me in a bad place. I can give you money!"

I mean, I agree. "content" is extremely an airquote situation, as an abstraction. I also dislike the term, but need some kind of catch-all for describing the ever increasing demand the fediverse has for space and processing. It's a real problem, though one that many people (not me) are actively working to solve and I'm grateful for that work.

My concern about fediverse space has more to do with inter-connected federation storage issues, such as caching content from other instances. Having seen how even small instances balloon in costs, it has completely scared me away from even attempting to host a small instance. I might have this wrong though, but I've heard horror stories about space requirements due to the caching-heavy nature of the federated internet.

I'm a bit confused. Is this the OOT port Ship of Harkinian or a different port?

Size and commercialisation generally go hand in hand. After all, how do you keep something running with infinite users submitting infinite content for free?

This is a big issue for everyone, it's important that the fediverse understands that size might some day increase the importance of sensible business models or proactive funding initiatives to counter the growth. This doesn't mean harmful profit models necessarily, but it does mean keeping revenues under control for individually run nodes or for archiving services for old fediverse domains that might not exist forever.

I do kind of agree with this. It's a bit suspicious project wise that this has appeared out of nowhere with no project history.

With the way things are going, I might eventually have to request people tag AI decomps or mods with appropriate notice when applicable. I understand the free stolen man hours are going to be "appealing" to the legal-gray area of game modding and DIY, but I do think people in the modding scene should be as transparent as we are expecting big corps to be when it comes to tool-assisted programming and projects.

I saw a youtube video about the 0RWX framework and, if the scene gets too cluttered, I might need to ask people to give their best estimates on what type of project it is within that framework. I think people have the right to better know how their software is made, regardless if it's the modding scene or big tech.

Anyway sorry to chat my head off Misk but I had the same thought when I saw this yesterday and I'm wondering what the retro gaming community as a whole will be able to do about this influx of AI projects.

I guess I'll have to check it out again. Last time I used it a few years ago it just wasn't exactly right on that front. But if it's within scope to use it as an image grid, I might provide feedback and changes upstream if it's close enough to the end result I want.

You should be selling your ram for your cat.

I'd love linkwarden to have a feature for image-links. I really need a substitute for pinterest that will cache a version of the image locally but still preserve the link to the page in question (so I can find the author/photographer/whathave you).

Unfortunate how bad the last 7 or so years have been market wise for the raspberry pi project. It feels like every time the hardware on one end gets cheaper, some other tech fad squeezes another component price up.

Fair fair. Ih suppose what I mean is that it's not really reverse engineering so much as it is recollecting from training samples.

I think this is definitely a concern, yes.

I think it's very likely that these models are effectively being tricked into retrieving source code of projects from memory of seeing the actual source code in the biases. This is why some project are much easier to uncover than others.