Yeah, they did that with RTX Remix afaik? But that kept the style of the game at least, it just or mostly upscaled textures. The face filter on DLSS 5 is the weird thing for me. Especially their example on requiem put me off, I love the style of the game, the acting was good, it’s been a good mix between great graphics and realism while keeping the resi aesthetic, but the gooner AI girlfriend grace is really uncanny valley territory.
All that to say, nothing I’ve said in this thread is the consensus of the moderation team because there is no consensus among the moderation team. If there was a feature to let me remove my moderator badge from comments, I’d utilize it. There is an option to “speak as moderator”, but I think that does… kindof the opposite of what I’d want? (Like it’d make the moderator badge bigger and meaner looking or something. I dunno.)
Absent a feature to remove my mod badge on a per-comment basis, though, maybe I should either 1) try to initiate some effort to get the mods all on one page, 2) step down as mod, or 3) not comment in a way that might be interpreted as a statement of the consensus of the mod team. I haven’t thought through which one of those if any I should do.
I did not want to call you out on the mod badge, it was a legit question. Because if this community is thought/planned to be an anti-AI sentiment only, it’s fair game, and then I can just call it quits and block the community on my end, because I have zero interest in neither techbro vibe code stories nor in anything that mentions AI is bad stories.
Sorry if the comment came out wrong - I just realized there that you were a mod, and probably misinterpreted your comment as well, since the whole comment chain before has been kinda heated too.
But I definitely think that you guys (as in the whole moderation team, whoever that might be in the end) should most definitely set it clear whether this is anti AI only, no discussions, or a critical discussion of AI where, given the name, the majority is obviously anti AI, but critical discussion is welcomed and encouraged. Both are fine, but y’all need to bite the bullet and decide which it is, the current state is … weird to say the least.
Is that your personal opinion, or is that the general consensus of the moderation team? Because if it’s the former, I couldn’t give less f’s about it, if it’s the latter, then you guys probably should rewrite the community description:
“A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype.”
I’d not call something a discussion if no other or more differentiated opinion is allowed, and putting that up as a community description while reacting with a “yeah, here’s the door, you kinda definitely should leave” is a fucking joke.
You’re dropping assumptions, and nothing more. Why do you think they asked for disclosure of AI usage? People are gonna use it anyway, officially or not. They did this call for this so that it can be reviewed properly.
One where we critically discuss issues with and around AI, but it seems we’re just chanting “AI bad” in a circlejerk.
If they didn’t care enough to write the code, what makes you think they cared enough to review or test the code?
Contributors =/= Maintainers.
I’ve never told anyone how to do their activism, I’ve criticized the consequences of said activism, which I still haven’t got answered, and the lack of objective arguments for this specific instance of activism, which I still haven’t heard, except for “AI bad lol”.
I’ve been rolling with the definition of slop that’s kinda universally agreed on, that is low quality, spammy AI generated content, and I’d ask again for an example of that in vim, but since your definition is “LLM used = slop”, I don’t need to do that here I guess. Also, you’ve missed the irony of telling me to not tell people what slop is, while telling me what slop is right after.
I don’t understand how you can be so dense to call that an easy problem that’s just boom done. It’s not about compiling and aliasing it, you can do that on probably any commit of vim. It’s about the maintenance and longevity of the fork, who’s gonna support it, and will it have a proper level of maintenance that will make it productively usable in the long term? It’s been forked to a “pre-AI” state of vim, how is it known that it’s not having LLM generated (as in LLM assisted) content already, before the official guidelines have mentioned that? If all that makes evi stand out is a strict no-ai policy, how is this gonna be checked and enforced (e.g human developer uses LLM tooling on his local machine, without disclosing it)? Who are the developers behind it, greetings from xz and similar supply chain attacks? How are upstream changes and fixes handled, since you’ll use it at some point with elevated privileges or to edit sensitive files? But yeah, fuck all of that, it compiles and you can just alias it, right, so we can talk about the severe problems in the open issues - will vim script be renamed too, and we need to rename vimrc to evirc asap, and boom done.
I’ve said it here at some point already, screeching “AI IS BAD REEEE” is not helping the case, it’s discrediting the “movement” or “activism” as a whole. AI will not leave. When the bubble bursts, people will stop shoving it everywhere, but it will stay where it can be used properly. Software engineering is something where it CAN be used properly, since whatever you’re building doesn’t give a flying fuck about who’s been writing the code. It’s either good, or it’s bad. Instead of worthless decisions on principle, do better. Coach and talk with people on how to do better, how to live in a world with AI responsibly and for good. Avoid, boycott and fork the ones deciding to not do that, based on objective reasons, and build it better. That’s what activism is about, using your actions to lead to change for the better, or isn’t it? And I don’t see how a hard fork, with all the mentioned consequences and problems, for the simple reason of vim maintainers saying “disclose AI usage” is leading to anything better just for the reason of shoveling an antislop and no-ai tag into the codeberg repository.
Of course it works fucking fine if it’s a hard fork of a stable state.
What mental gymnastics? The ones you’re doing right now. You have not answered a single question from my comment. And what “problem” did you solve exactly? Has there been any issue that has come up because of the acceptance of AI in vim? What kind of “slop” is actually there that makes vim problematic for you?
People vibe coding random bullshit ideas because they now can, do indeed produce slop. A bunch of highly experienced devs working on a successful project for years using tools that are at their disposal properly is not slop. You’re lending your public voice to a split of the community and of the project for made up bullshit reasons based on no objective proof but claims of slop and out of principle.
I’d trust the original vim maintainers to decide what’s a good or bad pull, instead of a bunch of random people who simply hard forked for literally no reason.
Wait, so because vim is allowing code written with AI we are switching to a random fork? The mental gymnastics here are insane once again. Is someone assuming that the vim maintainers are gonna do agentic requests? How is this project gonna handle upstream changes into their own main? Cherry-picking only “confirmed human-only” commits? Decisions like that out of spite, with zero thoughts and just out of principle do not help against slop. You’re just adding human slop to the AI slop.
I think the face is what’s most off-putting. It looks like straight from any of the generic “AI girlfriend” mobile ads you see. The environment is alright-ish I guess, for me personally it’s a bit too much into an uncanny valley too, because it looks too smooth, too perfect. It seems realistic, but the more and closer you look it feels wrong, because you don’t have that perfection in real life either, and the AI takes all of the realism and detail out of the picture.
Try taking any picture you have and smack it into an AI upscaler anywhere, and try to upscale it to x4 or x8, and compare the pictures side-by-side directly. Faces lose any kind of wrinkle, detail, markings. The generated image loses all identity it had, and while objectively it looks “good” as in “it’s upscaled correctly, it’s smooth and looks detailed” all and any of the details that give that image something are actually lost. It just has that “AI look”, the generic, hyper realistic “perfection”.