[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

2022-11: Warosu is now out of maintenance. Become a Patron!

/3/ - 3DCG

View post   
View page     

[ Toggle deleted replies ]
File: 1.57 MB, 1024x1024, steampunk_car_concept_design_sheet.png [View same] [iqdb] [saucenao] [google]
934276 No.934276 [Reply] [Original]

Image generation AI has quickly evolved to become an amazing industry tool with many applications: brainstorming, trying out ideas, studying design, generating textures for lookdev, etc.

However, a lot of people in the CG industry seem very hostile towards AI in general. How do /3/'s working professionals feel about this? Do you use AI at work? Have people told you not to use it? Why?

>> No.934278

hey OP
balance your prejudice on my cock

>> No.934310

I think companies profiting from models trained using scraped data shouldn't exist.
I think artists should embrace free and open models as part of their workflow if it benefits them.
I don't think there's anything wrong with using AI as long as you don't use the output directly and call it "your" art or use it to directly copy a living artist's style. I want anti-AI artists to explain to me why it's okay for some of them to make an entire career out of commissioning fanart but AI is like totally theft guys.

>> No.934311

I dont use ai because it all looks the same - bland, generic. Everything it makes, I've seen a million times before and done better.

I get enjoyment from crafting my own work. Free from any copyrght issues. Free from laziness.

>> No.934312

>I want anti-AI artists to explain to me why it's okay for some of them to make an entire career out of commissioning fanart but AI is like totally theft guys.
you can commission fanart because it doesnt take the exact work and then resell it. You cant sell AI fan art of an artist who holds copyright because you have to put their copyrighted work into the database and then remix it, something you cant legally do in another artistic, digital medium - djing and mashups

>A mashup is a style of music that contains elements or samples from songs created by other artists. In 2005, a court decision regarding the case of Bridgeport v. Dimension determined that it is possible for mashup artists to be guilty of copyright infringement even if a one second sample of music is used.

>> No.934520

AI image generation isn't a remix or mashup of existing work though. It's a new image just like if an artist made a new image. There's nothing contained in the final output that was originally from an existing work. The only thing from an existing work IS the existing work in the dataset.
You wouldn't be able to train an AI on Leonardo Da Vinci, then ask for an image by Da Vinci, and be able to pick out something like a tree and find that exact tree in one of his existing works like it was copy/pasted in there.
The AI knows what trees looks like, and it knows how Da Vinci paints them, so it's able to as well with any tree it comes up with.
So the mashup example you provided doesn't really hold any ground. Though in all likelihood, that's how lawyers are going to try and explain it to the old fucks that run the country that can't even make the distinction between "the internet" and "facebook". AI is way the fuck outside of their understanding, so they'll take anything they hear first (or charismatically enough), as truth.

>> No.934527

>There's nothing contained in the final output that was originally from an existing work
thats not what they are saying on /SD/. They say you can still violate copyright if you are deemed too similar. Cope.

>> No.934530

>Substantial similarity, in US copyright law, is the standard used to determine whether a defendant has infringed the reproduction right of a copyright. The standard arises out of the recognition that the exclusive right to make copies of a work would be meaningless if copyright infringement were limited to making only exact and complete reproductions of a work.[1][page needed] Many courts also use "substantial similarity" in place of "probative" or "striking similarity" to describe the level of similarity necessary to prove that copying has occurred.[2] A number of tests have been devised by courts to determine substantial similarity. They may rely on expert or lay observation or both and may subjectively judge the feel of a work or critically analyze its elements.

>> No.934544

The mashup example still doesn't hold ground. In your example, you're still using copyrighted material in the final output. The music contained in that mashup is still property of another entity, hence the entire term "sample". It's part of a released song by that artist.
Image generation doesn't work that way, as I've pointed out.

"Striking similarity" is something entirely different. To use your music example, it'd be like a cover that tries to sound 90-99% like the original band, instead of putting their own take on it. The audio and recording are completely original and made by some dudes, but because they try to make it super close to the original, it's borderline infringement. Those court experts would definitely be on them like flies on shit.
I'll agree that the whole "striking similarity" argument will definitely hold up in court, but not in the sense you're probably thinking. Most likely it'd be "x looks like y, which we own, but they're using it commercially in a way that makes consumers think it's us". I've got no doubt that if someone makes an output that's intentionally trying to look like an entity's copyright and attempts to use it, they'll lose a case. But in regards to your mashup example, it's still something else entirely.

As mentioned previously, AI outputs aren't chopped up bits of existing media, but they can make something that looks like it's done by an entity, and that's the gray zone for sure. I'd say it's the same gray zone as fan art, covers, and that sort of thing (as another anon mentioned). As long as you don't make a big deal of it and act like you own it, or sell things with the intent that it's a 99% replication of "x", there's nothing to worry about. It's definitely on the sketchier side of the gray zone, but it's not so much different than existing things like chinese knockoffs. Which is a good way to think about it.

>> No.934546

>How do /3/'s working professionals feel about this?
Dont really care
>Do you use AI at work?
>Have people told you not to use it?
My company has no stance on it but im still not going to use it because it obviously has copyright issues

>> No.934560

They are just afraid because people will realize that AI is cheap and fast while achieving the same as the average zbrush asset masher

>> No.934561

good point regarding the mashup. that's where i stood until i looked into how the diffusion model works at a technical level. the analogy of composite image is more accurate than calling it a mashup.
even then a composite implies that you are taking elements, when in fact you aren't taking anything at all the from the original image.

>> No.934562

being scared of this shit is Low T af. nothing in life is the "safe option" you gotta work with it or find some other source of income.

>> No.934580

>I want anti-AI artists to explain to me why it's okay for some of them to make an entire career out of commissioning fanart but AI is like totally theft guys.
Because you're not selling fanart, you're selling your skill as an artist to create what the commissioner wants.

>> No.934581

>im still not going to use it because it obviously has copyright issues
How is using AI different from using any ref you don't have the rights to? (So basically any ref ever)

>> No.934582

Why would you use refs that you don't have the rights to? Respect other artists.

>> No.934583

the only issues with copyright is the datasets containing the copyrighted images, the final model doesn't contain any of it but rather contains it in a form of memory, not very different from your own brain (which is a neural network not unlike ai)
to say the final output of the ai contains copyrighted work in some form would also be saying that all art in modern times contains copyrighted work in some form
you see countless artworks (dataset)
you think of an idea (prompt)
you create the artwork (output)
there is functionally zero difference between an ai and the human brain other than efficiency and specialty. the only issue that lies here is that the experiences and sights that you see and remember (dataset) is set into stone, it is catalogued for training

>> No.934584

AI's aren't humans. They don't have rights. You can't just mash up copyrighted materials in them.

They generate substantial similarity very very easily.
see >>934530

>there is functionally zero difference between an ai and the human brain other than efficiency and specialty. the only issue that lies here is that the experiences and sights that you see and remember (dataset) is set into stone, it is catalogued for training

we understand .05% of the human brain

>> No.934586

you could also generate substantial similarity yourself if you memorised and taught yourself enough. like i said,
>there is functionally zero difference between an ai and the human brain other than efficiency and specialty.

>> No.934588

>you could also generate substantial similarity yourself if you memorised and taught yourself enough. like i said,

i can't even remember what i ate yesterday for lunch. You dont know what you are talking about and again we know .05% of the human brain. COPE.

>> No.934589

not being able to remember yesterday's lunch != not being able to memorise
it sounds more like you are coping with your poor memory. what is there to cope about for me anyway?
saying that we only "know .05% of the human brain" doesn't mean anything. i am just stating basic facts that anybody can look up and basic conclusions anybody can come to. the brain is a neural network. ai is a neural network. our lives and experiences are our datasets and training. images are ai's datasets and training.
parallels anyone can spot. educate yourself before talking about something.

>> No.934590

You're digging yourself deeper and deeper. Goodday m8.

>> No.934591

you havent refuted anything except confidently make untrue claims
if anyone, youre the one digging yourself deeper and deeper. goodday m8.

>> No.934592

There's no use arguing with an extremist, so goodbye. Dont bother replying.

>> No.934629

>we understand .05% of the human brain
And judging by the shit you write, you understand .05% of copyright law

>> No.934634

I feel like it's more akin to drawing from memory or what you imagine a subject to look like in your head, only the AI has a really good perspective on what something should look like for obvious reasons.
I guess the question is, are memories copywritable? I'm sure at some point they will be, but right now the entire subject of AI is more of a philosophical issue rather than a clear line. I just hope that people in charge get a good understanding of it before they start drawing that clear line.

>> No.934656
File: 1007 KB, 2636x1946, 1654949045908.jpg [View same] [iqdb] [saucenao] [google]

Thanks to this thread I joined the darkside and tried prompting. Ultimately I wasted a day and I came to the conclusion that I'll never prompt again.

Prompting is like watching some streamer play a single player videogame. You get no enjoyment out of it and its really boring.

Not only can you not claim to be doing all your own work anymore but it gives you almost zero benefit since it doesnt give you a 3d model, it doesnt do animation, it doesnt do rigging, it doesnt do muscle simulation, it doesnt do groom, it doesnt do texturing, it doesnt create and mix audio....

Prompting...,like watching a streamer. Boring and devalues the whole "game" if you will.

>> No.934681

Prompting can get autistically deep, but even with an hour of fooling around you get better at it. If prompting doesn't work for ya, you could always go the img2img route and feed it something you've done partway and have the AI "finish" it.
As far as benefits go, a lot of people think that the AI should be the "end product", when its strengths really lie at the beginning of a creative process. Use it to generate concepts, refine your own concepts, or to give you a way of rapid prototyping.

Final render not quite "there"? Throw it into img2img and feed it a few different artists to see some different avenues you could explore. Having trouble coming up with something to make since you're creatively drained? Find something you like online and try to describe it as a prompt for a direction to start from. Have a project you've completed partway, but reached a point where you're not sure where to go? Put it into img2img again and try to describe the scene your going for and see some jumping off points.
That's not even thinking about training your own model on your work to generate results that are consistent to your own "style" if you have one.

If you see it as a way of creating an end result, it leaves some things to be desired, but as a tool to reach your own end result it's quite a useful thing to have in your repertoire.
I really suggest playing around with it a bit more.

>> No.934708

what is the difference between stable diffusion, dalle-2, and midjourney? Does one have to have to prompt through all of this different crap?

>> No.934711

looks like dalle 2 has terrible pricing and they will ban you over certain keywords in your prompts. This is what I was trying to tell you guys...

>> No.934712

looks like midjourney has bad pricing as well...this isnt looking good bros...only stable is free but they say the results are the worst

>> No.934746

>Because you're not selling fanart, you're selling your skill as an artist to create what the commissioner wants.
Would it be fine if you pay me $200 to draw Scooby Doo kissing a horse, but then I don't deliver the drawing to you? I still put the work and created the drawing, but you're not getting it, you're paying for the time instead. If you think this is unfair, then realize that you're actually selling the fanart; the guy paying wants to touch himself to Scooby Doo, he doesn't give a shit about your effort.

>> No.934775

I couldn't tell ya.
I just run Stable Diffusion locally on my machine. AI Gen is so finnicky, I couldn't imagine paying for generations. Especially when you can't change models or embeddings. This is one of those cases where "we have AI image gen at home" is superior to any of the services offered. At least in terms of value.
As far as I know, the local SD runs on most graphics cards, so if you've got anything with 4gb of VRAM, you should be good. I run on a 1080ti.
For quality on local, you get out what you put in (good prompts, great results), but even then you can get really good results depending on models, or your settings. The SD 2.0 model is out for example (think of models like "flavors" for the AI), and while you can't do NSFW stuff (there's other models that can though), it's still the same model that the site is using currently I think. So results won't be too dissimilar. You can get a 1:1 recreation of Novel AI's outputs with the local build, for example.

It's free to fuck around with, and somewhat easy to pick up. Then you might jump down into the rabbit hole of training all kinds of shit and becoming an amateur machine learning researcher overnight without realizing it. Not really, but you do pick up a lot.
Here's what I use
And here's a good jumping off point to get started

>> No.934783
File: 964 KB, 1845x891, 1660988050088.png [View same] [iqdb] [saucenao] [google]

I have all that going for me as well. I am on a 3060 12gb. The problem is now I cant go in and get img2img working right according to this tut


What my workflow should be is sketch, zbrush sculpt, or prompt concept, send it to img2img, get an exact copy like they did in the above video, and then refine it with the Alternative Test function. In the latest version, something has changed and it no longer follows this video. I cannot get img2img to change the result when I enter a new prompt in the img2img tab. What do I do?

>> No.934795

When I use refs, I can change the pose, lighting, whatever, so that whatever I do can't be pointed back to "this specific piece". When I use AI, most of them are almost copypasting the source image in some form, so thatll definitely get into trouble. I just want to stay out of trouble, I really dont care what tools I use otherwise (photobashing with the photos ive taken on my own etc), im not an "art purist", i just want to get paid and not get sued into oblivion.

>> No.934800
File: 380 KB, 512x512, 00000-20230124223546-363713996-calvin_and_hobbes.png [View same] [iqdb] [saucenao] [google]

>You cant sell AI fan art of an artist who holds copyright because you have to put their copyrighted work into the database and then remix it
Diffusion is not "remixing". Diffusion is a method that just statistically says how noise should be reduced with information from a text prompt. It takes white noise and passes over it several times, each time using math to figure out how the noise should best be reduced to match the prompt, based on a model. The model is basically what tells it what a square looks like, what a circle looks like, what a nose looks like, etc. But that model is not a collection of a bunch of different people's artworks, it's just a bunch of math formulas that are organized in a hierarchy. There is not a single shred of art in a Stable Diffusion model. It is NOT a database of art. There is no copy or remix of anything. None of the art that was fed into it was preserved, it was used to give the model statistical information about what things CAN look like. If you fed it a bunch of Picasso paintings it'd never be able to replicate Picasso because things like "nose" and "eye" lose all meaning when it comes to cubism and depend very heavily on context, which is something AI does not understand. It can generate generic anime waifus because they all look very similar and it's easy to statistically describe how a nose should look in anime style. Try telling it to replicate something as simple as Calvin and Hobbes, and it can't. Pic related is the closest thing I could generate, and it's still only close because it knows to associate "Calvin and Hobbes" with "stuffed animal" and "tiger". It is fully incapable of replicating the art style. Don't listen to the idiots on /g/ that think this will be the end of artists. It's a great tool for playing around and conceptualizing shit with, and I think it could genuinely harm the stock photo market, but artists really have nothing to worry about.

>> No.934801
File: 270 KB, 512x512, 00812-1398261503-a 3d render of a purple blob creature with black eyes and a blank expression staring into the distance in front of a simple gray.png [View same] [iqdb] [saucenao] [google]

Huh, that tut is "new" to me. I've kinda been out of the loop for a bit, so I've missed a lot of the new techniques and shit that have been popping up since like November/December.
I've definitely ran into that issue before where img2img pushes things too far or not enough, but my "fix" for it was largely to just turn the denoise strength down.
The post date of the vid is really fucking early though (Sept 14). I think people started checking out AI in droves around mid August. The vid is still probably using SD 1.3.
So I guess I just missed this tech the whole time.

I'm playing around with the alt function now, and I'm able to get the result to change, but not in the way I'm specifying (though I guess using Gunt for the test isn't the greatest control). I just recommend using the suggested settings and then messing around with the decode stuff.
Watching the generation, it looks like it creates a version based on your prompt, overcooks it into some noise, and then maybe some kind of fancy boolean or something happens with the data.

I'll have to play around with it more, and I recommend you do as well. If I can get it working I see it being of great use in my own workflows.
For what it's worth, thanks for pointing that out to me, even if I wasn't able to help your problem.

>> No.934802

>something you cant legally do in another artistic, digital medium - djing and mashups
Sorry just wanted to comment on this as well. The more appropriate comparison in the music world would be amp profiling like Kemper does, which has also been hotly debated, whether it breaks some kind of IP law or not. What amp profiling does is take a snapshot of a recording of a guitar through a given rig, and with math, figures out what a given input should give as an output. It's conceptually similar to Stable Diffusion, albeit way way waaaay simpler (it's not AI, it's really just basic algorithms). But, amp companies have tried to argue that this breaches their IP. The Kemper amp does not copy any circuitry or IP of any amp company. All it does is read the changes in a soundwave when a guitar is played through an amp, and replicates those changes digitally.
Also Kemper has been around for a decade and is basically a third tier company still, because people vastly prefer playing either real amps, or modelers like the Axe FX where you can just build your entire sound from scratch.

>> No.934806

Generates concepts much much faster. I can draw but I'm fucking lazy. I love AI

>> No.934807

>hurr durr it's not art if the file format is changed

>> No.934808

>hurr durr it's not art if the file format is changed
Son, if you want to talk with adults, you should at least learn how to tie your shoes.

>> No.934909

Hate it out of the principle of it being creative. Machines should not be creative and take away the joy of being creative from humans. Creativity is the biggest joy from making something. Even if you work with your hands, doing some things creatively as opposed to doing them by the manual is always more fun. We even let kids be creative for fun. AI should serve humans, not enjoy itself instead of humans, and then steal the enjoyment from people. Its like if you made AI that makes tiktoks. People would not like it because part of the fun is seeing people enjoy themselves and think of different ideas, idea sharing.
>But AI is just a tool
Did you design any of those cars? Did you have that image of the car in your mind? No? Then shut your fucking mouth, all you did is type "Steampunk car conceptart" into Midjourney. Every other artist in any other field if they made this car either as sculpture, or drawing, or digital art, or 3D model, or as a tory made of crab shells, they had this design of the car in their minds, all of them, maybe it changed a bit while they were making it, but none of them "just made steampunk car". Even normies know it and hate it, which is why AI tagged pictures get on average less then if you did not tag that picture as AI. People do not want future where creativity is outsourced to machines, that to most people sounds worse then living in Matrix Metaverse owned by Meta. The threat does not come from people preferring AI art, but from AI art becoming on average indinstingushable from human art. Not just that the average person could be fooled into believing AI art is real art, but that there will be no artists who will not be asked if it was made by AI. Normies and people in general have prejudice against bots, AI and everything artificial, especially bots in online discussions, people will always prefer human art over AI art, the problem however is that it will become impossible for people to know it.

>> No.934912

The prompting is just the beginning, a jumping off point. I used to be like you, hating ai. Now i embrace it and feel a lot better. You still have to use your 3d tools and animate, rig, etc, but now you have very strong concepts and img2img. I am in the middle of training a custom style in dreambooth for use with img2img. Thankfully i have 12gb of vram.

>> No.934915

I hate it because I can see or at least imagine more into the future then 1 year from now. AI wont be replacing any 3D artists anytime soon, no matter how much AIfags will try to convince you. Ai might be able to do ugly 3D models of bright animals, but that just replaces the sculpting part of 3D artist work. 3D artist also has to create the model from extrusion out of a cube to make good topology, and then also rig it properly. There are AIs that can help you animate better by adding physics, drag and other stuff to your keyframes, so you just make like 4 keyframes for kick animation ant the AI will make it look very organic. Those programs still suck at making rigs, because unsurprisingly they cant make drivers for shit. And none of this is again going to be automated anytime soon, this is very technical knowledge, something that the bots will strugle with. It is combination of visual learning, creativity and technical knowledge. Making a picture has no need for technical knowledge, or at least outside of knowing anatomy, you can take picture and then make variations of it by morphing it to get anything you want, you cant do this with 3D models that dont have shit rig.

but despite of all of this, I still hate AI because while it wont replace 3D artists this year, or the next one, it will do it in 5 years. It is proven that it can in theory do it if you solve some problems, and there are tests of it sort of working, all creative jobs are under the threat, because they are easy to manipulate trough computer. The only thing protecting them in the eyes of people was the myth that computers will never be creative, now we have proof that they indeed can be very easily creative.

>> No.934917

If your definition of "creative" is making superficially "pretty" pictures, than sure.
Creativity is about expression and perspective, and an AI diffusion model (or most AI of any type right now) will never have any of that. Simply for the fact that it just regurgitates what's in its network. It has no opinion on things, it has no feeling that it wishes to convey, it just is.
You could argue that what I just described is more of a hand-wavey philosophical point than something concrete, but I think the people who think creativity is all about whether or not and end result looks visually pleasing are the people who are threatened by what AI can do. Absolutely there's merit in having something look good, but there's fundamentally more to art and creative expression than the surface. Those people who are threatened by AI I think are also trying to shift attention from the fact that they don't want to confront that they themselves are just doing the same thing and aren't expressing themselves. Not all of them, for sure, but I'm sure there's more than a handful.

I really don't want the idea I'm trying to convey come out as some bullshit hippy nonsense, so I'm not really sure the best way to do that. I guess the main point I'm trying to make is that any AI these days has no self awareness, so it has nothing to impart on a work apart from quantified data on what makes something pretty. An AI that can make something truly creative would actually be an Artificial Intelligence, and not just the buzzword we use now (like "hoverboards" for those shitty scooter things).
AI can definitely make things that look good, but from what I've experienced using AI pretty much since SD first dropped the local build, and from what I've seen from others, it fundamentally lacks that spark that people bring to their work. Passion and expression show in a work, even if it's not perceptible to your eyes, your brain picks it up.

>> No.934942

>Passion and expression show in a work
sounds retarded

I see a world where nobody partakes in art, or video games, manga/anime, etc
In abundance, everything is meaningless

I don't really mind it
entertainment will die, and people will return to building communities

>> No.934959

>another AI shilling thread where the AIfag thinks people don't want consistent control over their work
>another AI shilling thread where the AIfag thinks having generative algorithms taking over our lives is going to happen soon and make the world a utopia

>> No.934971

Nobody says people who create stuff want the AI, the AI is for people who are uncreative and would normally pay artists. The machine not having any creative control is its positive since it completely alienates all artists from it and is only appealing to non-artists, which then allows people to get rid of artists as a whole. Remove the creatives from the system. If AI gave actual control over the generation, artists would not complain about it nearly as much, just another tool, but since it gives the user very little creative control, artists hate it. Only people with no imagination find it appealing, or permanent newbies who try to make something vaguely resembling what they always wanted to make and then spend hours in photoshop trying to edit it. But most permanent newbie artists already got bored of it and now its just normies who just found out how to use it and AIfags who will never get bored of it.

>> No.934972

>Nobody says people who create stuff want the AI, the AI is for people who are uncreative and would normally pay artists.

>> No.934976

>Throw it into img2img and feed it a few different artists to see some different avenues you could explore.
where is the list of artists that are available to input?

>> No.934977

It doesn't seem very useful to me. The artwork so far has seemed useless on technical level, the algorithm can't iterate fine details, instead forcing the user to gamble with keywords rather than just... well, pixelfuck precisely the detail they want to change.
Oh it can crap out middling artwork by terabytes, but who the fuck needs terabytes of noise, when one well done piece would be plenty.

I see so much parallelism with NFT fanboyism in the AI art fad and I'll give it about a similar timeframe until these "prompt artists" realise they have no footing or marketable skills to enter a competitive industry just by fucking with a morally questionably trained software some company happened to provide to public. I'm sure it'll find some niche usage here and there, but I predict it will largely fade from mainstream conciousness instead of causing any great paradigm shift.

So no, I don't use AI, I don't find it useful or interesting, and mostly don't care. I prefer references that are highly precise, and thought out, and unfortunately using random number generators and noise to produce procedural art won't cut it. A bit like a difference between Minecraft or No Man's Sky's world compared to a world of Witcher 3 for example.

>> No.934978

>Prompting is like watching some streamer play a single player videogame.
Perfect analogy.

>> No.934989

>Perfect analogy
For me it is more like watching a streamer play a competitive multiplayer game.
It's enjoyable, and also makes me want to play the game myself and git gud.
And actually, that exact same thing happened to me with AI art. It's so appealing it made me want to work on my 2d and 3d art skills so I can make similar stuff.

>> No.934990

unfortunately as an AI tard you aren't the type to engage in hard work instead of daydreaming :/

>> No.934997

I've spent much more time this year drawing, sculpting, retopologizing, and creating materials than prompting.
I sometimes like to take stable diffusion designs and try to sculpt them, understand their forms in 3d, and figure out why they're satisfying to look at

>> No.935012
File: 2 KB, 213x157, file.png [View same] [iqdb] [saucenao] [google]


>> No.935023

This guy gets it.
AI gives you a lot of shit to work with. You'd be a fool to think it is of no worth to a trained artist.

>> No.935311

>AI art fad
If you think AI isn't going to affect you, then you've got your head buried in the sand.


>> No.935320

If we don't know our history we are doomed to repeat it.


>> No.935321


>> No.935322

you can also copyright a 'style'

if you sell a comic with story about blue creatures on amazon, you get a problem.

if the style id similiar to avatar then its problematic.

even if your AI never copied anything, just generated complete new with similiar pattern.

>> No.935323

in some years ai generates daily 5x AAA games with dubbed dialoges, storys, levels, 3d worlds, cinematic cutscenes

consoomers will find it great at beginning

but when market offer is oversaturated, it will become boring over time.

its psychology of humans

>> No.935347

you realize Luddites lost not because "technology beat them", but because they got executed by the government right? what are you supporting here, cucking yourself towards the government whose main interest is to dispose of you? the only reason you exist, is because governments need a lot of humans to maintain a self-sustaining society and it's in their self interest to keep you surviving. once that's no longer needed, you're out of the question. And this isn't a "loom", it's meant to replace you as a human being from all aspects of life. use your fucking brain retard, if you have any left in your skull

>> No.935365
File: 44 KB, 850x395, lol.png [View same] [iqdb] [saucenao] [google]

I'm not cucking myself towards the government. I'm simply acknowledging that the vandalism actions were an important factor in the failure of the people in question. It's important to recognize that this situation should be learned in order to take steps to ensure that such an event does not happen again with AI. I'm not blindly supporting the government.
Although it is true that governments need a lot of humans to maintain a self-sustaining society, this does not mean that the only reason I exist is to serve that purpose. I believe that humans are capable of much more than simply fulfilling the needs of governments. We have the ability to create, innovate, and make a positive impact on the world. I believe that I can use my talents and abilities to contribute meaningfully to society, regardless of whether or not governments need me to do so.
And yes I did not write any of this an AI did.It's a brave new world kid you better start learning how to use AI to your advantage.

>> No.935377

How does this even relate to my post?
It's a cute paper and some cool tech, but procedural text to x is still procedural noise I don't give a damn about.

>> No.935456

The thing about AI is it doesn't matter if you don't give a damn about it. It doesn't give a damn about you either. Producers want lots of art for fast and cheap. If a company can replace 5 modelers with an effective AI model generator, they'll do it.

Delete posts
Password [?]Password used for file deletion.