Is AI the future of Indie RPGs?

[mondblut]

If there is one thing machine learning is good at, it is identifying pattern, in this case a pose.

Apparently hands are too difficult a pattern to identify.

 

[Sarathiour]

You should really start preparing yourself emotionally, because the floodgates on neural network technology are now open, and we are entering a phase of rapid and accelerating development.

Would you consider 10 years enough of a time frame for us to quote this back to you ?

 

[The_Sloth_Sleeps]

You should really start preparing yourself emotionally, because the floodgates on neural network technology are now open, and we are entering a phase of rapid and accelerating development.

Would you consider 10 years enough of a time frame for us to quote this back to you ?

2016 Was the Year Self-Driving Cars Took the Wheel​

And made clear they're on the way to dominance.

Sure, the California DMV could get its way and successfully kick Uber's self-driving cars off the streets of San Francisco. But if it does, it will be just one small step backward in a year that's been all about robo-rides charging ahead.
In the past 12 months, autonomous vehicles have pushed into the public consciousness like never before, thanks to technical advances and pilot projects that have revealed both that these vehicles aren't so far off as many might think, and that their potential to reshape how humans move about.
In 2016, Uber welcomed the public into its cars with a (definitely legal) pilot project in Pittsburgh; nuTonomy did the same in Singapore. Google officially launched its self-driving project as a standalone company under the Alphabet umbrella, called Waymo. Otto, the Uber-owned autonomous truck outfit, hauled 50,000 beers across Colorado with nobody at the wheel---or even in the driver's seat.

Before plunging into 2017, here's a look back at a year humans will likely remember as "that time we realized we don't have to drive anymore."

https://www.wired.com/2016/12/2016-year-self-driving-cars-took-wheel/

 

[The_Sloth_Sleeps]

This opinion was obsolete even before the video was posted. There are already multiple AIs that have solved this problem. The speed of AI development is accelerating like crazy now.

Not sure if serious, but: within the past 5 days this video became obsolete? Would like to see some of the advancements you are referring to.

Learn to read. It was obsolete before it was posted.

The guy is basing all of this off of an article or two he read, and the fact that he personally hasn't seen AI being used for professional grade art, even when it's already being used for it.

His entire video rides off of the idea that all you can do with image AIs is type in words and generate random images. And if he had even bothered to do even the barest research, he'd know that's wrong.

At the end, he even talks about creating different AIs for different art styles as if it's unlikely or difficult.

Again, WHAT improvements? The video is 5 days old, and I note that the opinion expressed is very much still accurate. Waving your hand and claiming that is ancient history....is difficult to take seriously given your unbounded enthusiasm for proompting. The reality is that AI art is destined to be relegated to a very narrow niche in the art industry.

This guy in the video an AI art enthusiast, he makes tutorials on it so that explains his gushing praise but even he is somewhat realistic about its potential unlike your fantasy about being a proompter.

What Stable Diffusion and other AI art generators do is relatively simple in comparison to that little robot car.

https://futurism.com/self-driving-industry-going-nowhere

Godfather of Self-Driving Cars Says the Tech Is Going Nowhere​

"It’s an illusion."​

There’s tons of enthusiasm over self-driving cars. But nearly twenty years and some $100 billion since the first demos, the technology just isn’t there yet. In fact, it may be further off from being fully — and safely — autonomous than we’re led to believe.

"You’d be hard-pressed to find another industry that’s invested so many dollars in R&D and that has delivered so little," said self-driving pioneer Anthony Levandowski, in a must-read new interview with Bloomberg.

As the cofounder of Google’s self-driving division, Levandowski is acknowledged by his peers as one of the key engineers that got the industry up and running.

...

The hype around this "AI" art crap reminds me of the hype of "AI" cars back in 2014.

 

[Non-Edgy Gamer]

The difference being that you can use an AI for art right now. How many self-driving cars can you drive?

 

[Sarathiour]

About as many as AI "artpiece" being used in the entertainment industry.

 

[The_Sloth_Sleeps]

The difference being that you can use an AI for art right now. How many self-driving cars can you drive?

No, I can't actually.

Even if this had an isometric tilleset training set it wouldn't work to make me isometric tiles. It won't ever be ever to take knowledge from one domain and apply to another. Thats why, as stated before, you will always need a human artist.

But instead you stare at the stylistic tricks that repeatedly fool you into thinking its something genuine and this will one day make art for your game.

 

[Non-Edgy Gamer]

Again, WHAT improvements?

I'm not the one who brought up improvements, but I would assume the various new models that have been trained and features that were unlocked since day one: inpainting, outpainting, img2img, VAE and textual diffusion etc.

The video is 5 days old

For the third time: it was out of date BEFORE it was released.

What part of that is hard for you? He's not addressing the capabilities of even vanilla Stable Diffusion shortly after its release.

This guy in the video an AI art enthusiast

A pretty poor one by the look of it.

Again, I have no idea why there's a giant AI bug up your ass. You're in every thread. You're watching videos on it. But you can't be bothered to learn about it or use it because there isn't yet a model that's suited to your specific use case of making tilesets. So why are you here?

 

[Non-Edgy Gamer]

About as many as AI "artpiece" being used in the entertainment industry.

There are already several games in development using AI art, there are large games companies using AI for their voice overs now and many writers are using GPT writing assistants.

Every facet of the entertainment industry is using AI now.

edit: Sarathiour you've been in this thread and others. You've seen the citations posted. You already have your citation and are just asshurt and trolling.

 

[tritosine2k]

It would show more promise if they applied it to the NeRF stuff (lightfields) with actual depth.
But their bleeding edge is upscaled low res at 30 sec lol (usable lightfields are 8k equivalent usually) and 80's pixar 3D.
This will eat petaflops like it's nothing.

 

[Sarathiour]

He's not addressing the capabilities of even vanilla Stable Diffusion shortly after its release.

Why are you getting mad over a video you did not even watch ?

 

[Non-Edgy Gamer]

I watched the video. I brought up things he specifically talked about. You have no argument.

 

[Non-Edgy Gamer]

It won't ever be ever to take knowledge from one domain and apply to another.

What are you even talking about?

But instead you stare at the stylistic tricks that repeatedly fool you into thinking its something genuine and this will one day make art for your game.

It's already making art for games.

https://www.reddit.com/r/singularit...us_tenebrarum_an_nsfw_video_game_made_almost/

https://store.steampowered.com/app/1889620/AI_Roguelite/

And our local Codexer:
https://rpgcodex.net/forums/threads...ow-an-ai-is-doing-my-games-art-for-me.144303/

Which is more than I can say for you as you sit in these threads all day, scribbling things in MSPaint and whining about AI.

 

[Sarathiour]

The guys took creation from the midjourney community as an example ( so july 2022 at the earliest ). So way past "vanilla Stable diffusion", whatever you mean by that.

Ha, examples at least.


Wow, would you look at that, shovelware. The future is exciting indeed !

 

[The_Sloth_Sleeps]

The guys took creation from the midjourney community as an example ( so july 2022 at the earliest ). So way past "vanilla Stable diffusion", whatever you mean by that.

Ha, examples at least.


Wow, would you look at that, shovelware. The future is exciting indeed !

I was going to say, people will will find a way to make "games" out of this crap. And he wonders why I hate "AI" and these misnomers.

The shining example of "AI" game art:-

 

[Non-Edgy Gamer]

The shining example of "AI" game art:-

The AI art is everything in the background and the image.

The human art is the 3d model.

Thank you for posting an example of AI art being superior to human art.

 

[The_Sloth_Sleeps]

The shining example of "AI" game art:-

The AI are is everything in the background and the image.

The human art is the 3d model.

Thank you for posting an example of AI art being superior to human art.

Backdrops for shovelware. Impressive.

 

[Non-Edgy Gamer]

The guys took creation from the midjourney community as an example ( so july 2022 at the earliest ). So way past "vanilla Stable diffusion", whatever you mean by that.

When did the general conversation about AI in games suddenly become about Stable Diffusion alone?

Oh right, when I posted proof that you don't know what you're talking about and you had to move the goalposts.

Wow, would you look at that, shovelware. The future is exciting indeed !

You implied it couldn't be used for indie games. I posted proof that it's already being used for them.

And not only that, but these are earlier versions of AI technology than what's available right now.

 

[Non-Edgy Gamer]

Backdrops for shovelware. Impressive.

Again, it's still better than your nonexistent game.

How many months have you been lowkey begging someone to figure out AI for you so that you could have an interface?

 

[The_Sloth_Sleeps]

Backdrops for shovelware. Impressive.

Again, it's still better than your nonexistent game.

How many months have you been lowkey begging someone to figure out AI for you so that you could have an interface?

My UI is a work in progress. I'm going to figure it out myself. Proompting is hardly going to get me anywhere now is it?

 

[Bohrain]

With my massive knowledge attained from lurking the various AI smut threads on 4chan, I've come to the conclusion that it will probably have use on certain asset production. The outputs that actually do look deceptively manmade require something like 8 lines of input tags with a decent amount of trial and error. The models usually get faces right, because it's usually in the picture focus on most of the images the models were trained with, so it's natural to assume that's what they are best at. Smaller detail that is out of focus gets routinely messed up, hands being the typical thing. Hence why the cherry picked good outputs never have the have the hands properly visible where you could see the individual fingers.
In terms of practical use I think 2D backgrounds/scenery is where it's most useful. If you look at the outputs closer, yes you can see that the forms are fucked up on some house in the distance and so on, but generally speaking people don't notice if the picture looked allright at a glance. It's also a subset of art where people aren't too keen to notice minor stylistic differences.
When it comes to character art, it becomes more tricky. The AI can do good individual portraits. But if you want to do something like a visual novel or a isometric RPG with over a dozen characters, maintaining a similar style becomes tricky and you essentially have to crop out hands from everything. Skin gradients look really similar in many AI outputs too, but if the prompter is good enough to make the style look consistent that can probably be overcome. But either way this requires enough time investment to learn that your run of the mill Unity asset flipper isn't going to do. Still, this probably does compete with the cheap artists from Fiverr and such.

 

[Non-Edgy Gamer]

With my massive knowledge attained from lurking the various AI smut threads on 4chan

Most 4chaners just spam the same half-baked images over and over.

The outputs that actually do look deceptively manmade require something like 8 lines of input tags with a decent amount of trial and error.

The real workflow for quality is initial prompt to get most things right > img2img for cleanup and modifications.

Sometimes more prompt text weakens the overall result, but it depends on the strength of whatever tokens you are using.

But if you want to do something like a visual novel or a isometric RPG with over a dozen characters, maintaining a similar style becomes tricky

It really doesn't. For visual novels, you're going to be using anime, and NAI does this very well and consistently with a variety of available styles. For general portraits, it can be enough to use the same prompt and settings.

Hands, eyes and other fine details can be fixed using img2img or inpainting.

Portraits are easy. It's complex character poses that are going to be trickier. But there are ways around this using Krita and img2img.

The time commitment to learn these "skills" is measurable in minutes to a handful of hours. The time commitment to learn how to draw things yourself is potentially years, depending on the kind of art.

 

[Dexter]

Comics:
https://www.artexmachina.org/magazine/
https://www.evilcorpgame.com/blog/2022/9/29/arc-a-tv-show-based-on-the-evil-corp-boardgame
https://www.reddit.com/r/StableDiff...ew_pages_from_my_midjourney_produced_printed/
https://www.reddit.com/r/StableDiffusion/comments/yjqxhw/the_lesson_a_free_comic_download_made_with/
https://www.reddit.com/r/StableDiffusion/comments/y6dhl6/consistent_comic_character_test_with_sd/
https://www.reddit.com/r/StableDiffusion/comments/x4cqsp/working_on_a_comic_book_inspired_by_2000ad/

Examples of various professional Artists already using it as a Tool in their workflow:
https://www.reddit.com/r/StableDiff...w_to_make_money_as_an_artist_with_a_personal/
https://www.reddit.com/r/StableDiffusion/comments/ybuw1h/thoughts_of_sd_from_an_artist_in_the/
https://www.reddit.com/r/StableDiff...rt_is_popular_and_makes_money_confessions_of/
https://www.reddit.com/r/StableDiff...a_3d_artist_and_im_using_stable_diffusion_to/
https://www.reddit.com/r/StableDiff...thesis_of_a_professional_illustrator_in_love/
https://www.reddit.com/r/StableDiffusion/comments/xshmpd/raw_output_final_artwork_oc/

Various Proompting Magazines and Guides:
https://www.promptia.art/
https://openart.ai/promptbook
https://docs.qq.com/doc/DWHl3am5Zb05QbGVs

For games we have a respective thread in General Gaming with various examples and where people, including many from the Codex, have already posted AI-generated art they're going to use in their games, and I'm pretty sure people here already know that.

You can flail around and be butthurt all you like, you're not going to stop technological progress.

 

[Maxwell's Demon]

You should really start preparing yourself emotionally, because the floodgates on neural network technology are now open, and we are entering a phase of rapid and accelerating development.

Would you consider 10 years enough of a time frame for us to quote this back to you ?

2016 Was the Year Self-Driving Cars Took the Wheel​

And made clear they're on the way to dominance.

Sure, the California DMV could get its way and successfully kick Uber's self-driving cars off the streets of San Francisco. But if it does, it will be just one small step backward in a year that's been all about robo-rides charging ahead.
In the past 12 months, autonomous vehicles have pushed into the public consciousness like never before, thanks to technical advances and pilot projects that have revealed both that these vehicles aren't so far off as many might think, and that their potential to reshape how humans move about.
In 2016, Uber welcomed the public into its cars with a (definitely legal) pilot project in Pittsburgh; nuTonomy did the same in Singapore. Google officially launched its self-driving project as a standalone company under the Alphabet umbrella, called Waymo. Otto, the Uber-owned autonomous truck outfit, hauled 50,000 beers across Colorado with nobody at the wheel---or even in the driver's seat.

Before plunging into 2017, here's a look back at a year humans will likely remember as "that time we realized we don't have to drive anymore."

https://www.wired.com/2016/12/2016-year-self-driving-cars-took-wheel/

... And now you're just deflecting. While you're having fun knocking down straw men, other people are learning to code neural networks.

About as many as AI "artpiece" being used in the entertainment industry.

That ship has sailed. You haven't been keeping track of shit for nearly five years.

 

[Sarathiour]

When did the general conversation about AI in games suddenly become about Stable Diffusion alone?

Oh right, when I posted proof that you don't know what you're talking about and you had to move the goalposts.

No, you claimed that the video was outdated.

You implied it couldn't be used for indie games.

being used in the entertainment industry.

sure buddy

Comics:
https://www.artexmachina.org/magazine/
https://www.evilcorpgame.com/blog/2022/9/29/arc-a-tv-show-based-on-the-evil-corp-boardgame
https://www.reddit.com/r/StableDiff...ew_pages_from_my_midjourney_produced_printed/
https://www.reddit.com/r/StableDiffusion/comments/yjqxhw/the_lesson_a_free_comic_download_made_with/
https://www.reddit.com/r/StableDiffusion/comments/y6dhl6/consistent_comic_character_test_with_sd/
https://www.reddit.com/r/StableDiffusion/comments/x4cqsp/working_on_a_comic_book_inspired_by_2000ad/

Glad to see that gary's mod is making a comeback.

That boat has sailed. You haven't been keeping track of shit for nearly five years.

So upscaling is now defined as "artpiece" ? Whatever.