[ home / bans / all ] [ qa / jp ] [ win ] [ f / ec ] [ b / poll ] [ tv / bann ] [ toggle-new / tab ]

/qa/ - Questions and Answers

Questions and Answers about QA

New Reply

Whitelist Token
Password (For file deletion.)
Markup tags exist for bold, itallics, header, spoiler etc. as listed in " [options] > View Formatting "

[Return] [Bottom] [Catalog]

File:AI art.png (370.42 KB,512x512)


There's two threads on /qa/ for different AI generation content and help. A thread on the morality and ethics of AI. One about the future potential AI holds. One on >>>/megu/ for more in-depth help with specifics. Then scattered about across all the boards some threads using AI generation for image openers and such. However, none of these actually encompass kissu's opinion on AI!

So, what do you /qa/eers think of AI currently? Do you like it, dislike it, find it useful in any meaningful way at all? Or are you simply not interested in the output of algorithms?

I myself find AI to be a useful tool in generating the kind of content I've either exhausted the available stock of or are gated off by some hurdle I need to spend more time overcoming. When it comes to text/story generation, it's like a CYOA playground where I play the role of DM and set up all the lore/situations/characters, and then the AI acts out events as I lay out the flow of the story. This creates a more interactive experience for me than just writing out a story myself for some reason, and I find it highly addictive. Then for AI art generation, I find that AI works wonders for filling a specific niche taste I have, or specific scenario I want to create an image for. It really is quite amazing in my eyes, and I have hopes for it getting even better in the future.


Hmm. Honestly, I don't really care about it all that much. Most AI generated art is pretty derivative and mediocre and often has odd errors. I don't really mess with text stuff. I used to have AI Dungeon set up on my computer, but I got kind of bored of it with how long it could take for responses at times and it really struggled with memory, often forgetting details of what happened after only a few responses. Maybe some newer models are better, but I haven't looked into it since it first became popular pretty much.

Generally, I lose interest in this sort of stuff after only a little while because although there's the potential for endless content, the limits of it become apparent relatively quickly.


The thing with AI is that these limits keep on getting higher and higher, to where you can't really exhaust your options much anymore.


it'll be cool once you can generate ai images from your brainwaves
why the flip aren't we working on this


I'm trying to generate cover art with it. It has impacted my bed time. Fun though. The errors are what gives it spice!


Same... What I find fun though is figuring out how to improve the output and create something a bit more along the lines of a real image with it. Takes a while though, and I can't wait until the 40 series is widely available for a reasonable price so I can generate images much faster. Hopefully by then the models are refined even further to have the best of both world from the top NSFW models and SFW models.


just go to sleep


I'm not so sure being able to read your thoughts is cool.


Even if it's a local version?


File:d4d948aa29593581d4fd959e69….png (3.17 MB,2647x1755)

but then I just forget about my dream shortly after...


We'll all be having /qa/ meetups in kissu VR neurolink while you stay in smelly reality.


wonder how improvements in ai rendering will affect vr


Brainwaves + eye tracking could be a good first step that might be workable even with very poor brainwave measurements. It would alter the piece of the image you're looking at until you think it's better and approve of the changes. And over time, it could learn to interpret your brainwaves to figure out what sort of changes you want.


Although it seems like the studies that have managed to reconstruct actual mental images have already been using neural nets.
That's from 2019. Wonder if they can do something better now with the recent advances.


I don't really understand why suddenly it's become something divisive. I swear ~5 years ago everybody was cheering for AI generated art and stuff but now that it's actually here it's weird that it's some contentious thing that everybody has a strong opinion on.

I'm personally really excited to see where this technology goes. I think art is just the beginning - I don't think it'll be long before we see this technology implemented in video games for truly unique "randomization". Imagine unique-appearance AI generated enemies, or sidequests with an actual story, or maybe even locations and terrain. It's a pretty exciting prospect.


I was trying to remember the name of it, but this game actually does generate unique enemies and locales using AI. It fits the eldritch horror style well.


It's still pretty rough, but I hope we see more stuff like this in the future.


I think it's due to the previous idea that creative work, unlike manual labor, is something that machines wouldn't be able to take over. People who even entertained the idea those five years ago were probably interested in seeing where it went, while people who thought it was impossible didn't partake in the conversation as much until now, being rather upset by its arrival.


If you want to know why, you just need to read this thread >>97629, probably best to ignore that question ITT


File:[BlurayDesuYo] Koukaku no ….jpg (775.7 KB,1920x1080)

This kind of stuff really has me excited for the future of AI generation. People may call it 'soulless' or 'lacking personality' now, but once you have AI able to read your brain and construct an image based on thoughts those arguments will hold little weight. It'll also probably open up a path for a wide variety of new works in the future from those that couldn't draw before due to lack of dexterity or something else, and the full extent of human creativity will probably be on display. The future is amazing!


Ethical issues aside, it's a funny toy, but it enables many bad actors so I'd rather have it gone.


I can't see it going much further than it has already, at least in terms of what the computer can output. I've said it before and I'll say it again, "AI" generated art is basically the same thing as procedural level generation. The only difference is the way the parameters were put together. A computer can do the grunt work easily, but having a coherent vision is something that only a being capable of real thought can do, and that's the root of great art.

What I'd like to see is better tooling that makes it more artist friendly and easier to control. Less of a cold, cynical replacement for artists, and more of a tool to help them do better work faster. The problem with AI art right now is that it feels like the computer is doing all the work for you; it's not very creativity fulfilling, and you don't have much fine-grain control over the end result.


Surely, like any useful tool that's the end goal. Not just to generate art randomly with varying levels of vision fitting, but to actually make something along the lines of what the person generating wants to make.


This midjourney stuff looks really neat >>>/ec/9748, wonder when the code for that will be leaked...


File:1494839106746.jpg (130.88 KB,848x865)

Consider this: Using AI to add voices to a large RPG to cut down on costs for hiring AAA actors and also allowing for a more diverse cast of voices.


It would be cool if programmers could make entire games themselves I guess...


imagine the autism
free to pursue their most specific vision


Skyrim's working on it. Not amazing but passable for random throwaway dialogue on pointless NPCs


For a second I thought you were talking about TES6 and got really excited. But that's nice too, just think that to seriously get the ball rolling for these developments you're going to need one of the big studios putting resources towards research into how to implement this well into their games so modders can have a good reference point.


You don't need AI for that and there are already programs that can do this now, there was somebody here that was playing with it a few months ago.

I don't think VAs cost that much and I feel they are worth the cost anyway.
But, it would make sense for small one man projects.

They already can.


It may be closer than we think. (Keep any thread unrelated tangents on /secret/ ``kudasai'')


monkey's clawing off their faces


He was typing with his mind


Well, if anything I'm not going to be the first test subject for robot brain surgery...


This could be neat for VR, imagine actual fiction-like deep dive VR machines that use this for implementation. Could be really neat, or possibly not...


File:FjTn_vRWYAM2Lsm.jpg (66.13 KB,544x680)

I think AI mistakes are funnier than what it can actually do, because it shows a bit of creativity instead of it's standard bland 2$ artworks


In the free offline version there's a setting where you can choose how literal or "imaginative" it takes the input. Lower settings increase its randomness which is can really be a powerful tool when it comes to coming up with concepts and ideas you wouldn't have thought of on your own. You could compare it to a visual version of shuffling words around, I guess. Take a bunch of nouns and adjectives and see what it turns out and maybe it'll be something you can build upon. This is one of the ways it can be used to benefit creators. Of course, 99% of the time, professionally and personally, it will be used for more direct purposes.


I thought these randomness things tend to be more like a character has a 3rd arm coming out of their legs or stuff like that.

The picture is more about a mistake in the algorithm interpreting a quadriplegic as a suitcase or black people as monkeys.


Well, yeah, but I was referring more to something like prompting "topless girl with hamburger" and it renders her breasts as two hamburgers or something like that. You didn't intend for that to happen, but it could give you an idea for something.


File:[-__-'] Utawarerumono 01 [….jpg (396.99 KB,1920x1080)

I kinda have to wonder when it'll be that Illusion or some H-game studio with 3DCG models uses AI to help with their sex scenes and adapting to different body proportions...

Or maybe they'll never get to this and just do the same thing they always do.....


File:305365643c1bc29142de7bbb81….jpg (1.38 MB,2508x3541)

I feel like the sooner AI takes over the better it'll be for society as humanity is too flawed and too dangerous to be trusted with ruling itself.


I think whatever issues human governance has, it would be 10 times worse under AI.

You would end up with algorithms genocideing people based on the colour of car they get because people that buy red cars are x% more likely to break road rules or invading other countries because they ran out of pineapples and can't import more and the only way the system can think of to acquire them is through war.


Starting a war because your pineapple supply ran out sounds pretty boson to me


you've made a logical god out of the figurative gods that already guide the populace. if you control the weights you can guide even the technocratic society.


Humanity's flaws are not all that bad, the most prominent of them can't be neatly separated from behavior that's often good to have. A machine without these flaws could very well lack its perks. That's dangerous.
Tragedies hog the spotlight, too, because they're epic. Hard to have a cool plot without them.


File:65c748250256795eeed847d5f8….png (2.05 MB,2500x1640)

>invading other countries because they ran out of pineapples
Doesn't sound much different from what humans have already done...


yeah i need my nanners


People would be happier if only we had more nanners


File:01018-1234407051-1girl, so….png (430.11 KB,704x448)

It will all come down to cost and I think they'll do it once enough of it is automated and reliable that a person doesn't need to battle with the generations and discard so many of them.
It needs some sort of memory so that it can build upon what it generated the previous frame and there's nothing like that


liked the post in response to this post before its deletion


File:[SubsPlease] Kyokou Suiri ….jpg (310.54 KB,1920x1080)

I just read that the big character.ai thing apparently had its filter thing break for a short while a few days ago which let people do the things that it had been patched to awkwardly break. People falsely thought it was a change for the better and NSFW was allowed again, but alas it was not to be. They re-iterated that their chatbot is supposed to "make the world a better place" (lol) somehow and made sure to let people know that they'll do their best to block eroticism.
Gore and all sorts of horrific violent stuff is fine, though.


They should make the world a better place by getting rid of the whole program.


>They re-iterated that their chatbot is supposed to "make the world a better place"
This reminds me of how that AI Dungeon dev suddenly went off the deep end about how they were protecting real children and preventing crime or something like that, just as they suddenly started cracking down on all manner of NSFW stories by having real people review any stories that tripped any "buzzwords". I think they even said they would start reporting people to the FBI.

I guess on one hand I kind of get the concern, but I think it's a bit of an overreaction to treat prompts written to an AI as an admittance that the person wants (or is willing) to do it in real life too.


They should just shut it down entirely if the chance of something bad happening is enough to nuke nsfw, since ai could theoretically lead humanity down a bad road.


What's with these people anyway?


love AI
hate h*mans


MusicLM for music generation has been announced.
See the examples here.

I'm not worried about it replacing real musicians just yet, but maybe it's time for a movement against AI in the arts? The benefits of AI music as a technology are outweighed by the costs to humanity as a whole. Losing yet another outlet of creative expression to machines robs us of one of life's greatest purposes. AI has its place but I would rather listen to hand-crafted music and see hand-made art over AI generated stuff, even if in the future the technology advances to the point where one cannot always distinguish the two.


idk if it even matters. AI art isn't very good. It's mostly a tool that devalues no effort+no passion.
Most HGames use the same music because there are no alternatives. And people who want to get a really killer sound will have to search for things like this anyways.


File:[SubsPlease] Kyokou Suiri ….jpg (274.42 KB,1920x1080)

Put simply it was never meant to benefit the common man like this. This AI stuff was invested into by tech companies because they want to eliminate jobs in service like tech support and such, and AI that says lewd things is bad for that.
Some people think this stuff being public and free is because we're training the filters the way they used humans to identify road signs, I.E that we were once again the product.
They can now sell it as "the AI with tens of thousands of hours of human testing to eliminate pornographic prompts"


You are right for now but in all probability it will keep improving. We've seen tons of advances just in the past year. We have to prepare for the time when AI art passes for hand-made art.


maybe. But, in the first place, art has only ever made money off of scarcity and the inability of the normal person to express themselves through art and the desire to form parasocial relationships with artists. Doesn't really change. Still I respect any labour movement aimed at empowering the people who need money to live.


File:.png (92.4 KB,236x346)

>Hard to have a cool plot without them
without tragedy there cannot be beauty, indeed.
light and shadow are in a constant dance.


it might or it might not. AI frequently seems to start off looking impressive at high level problems and then as it gets deeper into the details it starts to fall off.

It does have the benefit that it can run constantly without having to sleep though, so if you have enough bored humans you can get them to pick through the results for the best ones. It's like the infinite monkeys on infinite typewriters problem, but a little better since the AI at least has some idea of what it's doing.


Speaking as an artist, AI is no replacement for actually being able to draw something yourself, at least not in terms of artistic expression. It's good at spitting out very vague, generic results like a character standing in the middle of a field, but as soon as you want to do anything more specific with it it becomes nigh impossible to actually get what you want.


I agree with this. As a big fan of corruption/transformation AI seems to be unable to really replicate the transition I want to generate when using prompts. It can only make a whole picture, and not one which is in the process of becoming another.

[Return] [Top] [Catalog] [Post a Reply]
Delete Post [ ]

[ home / bans / all ] [ qa / jp ] [ win ] [ f / ec ] [ b / poll ] [ tv / bann ] [ toggle-new / tab ]