This AI rendered Apocalyptic Blender Classroom Scene in 5 seconds

It’s called self-defense. Many people are threatened by this, so the way they feel better about it is to point out how it’s not good enough. It’s a short-sighted way of thinking.

4 Likes

While you make a good point, I would argue that it’s more short-sighted to say that “AI is going to replace artists in a year or so”, given that art is a few million years old and has survived a wide variety of similar assaults over time

I just wrote an article about this very specific subject (only within the context of music):

The issue that seems to persist when approaching any discussion about ML and art is that we tend to cobble all art together in the same category. Is the artist who is creating a logo for a law firm the same type of artist who paints works to be exhibited in an art gallery? If we are able to differentiate the two, then we might also be able to understand that the former artist’s job might disappear, while the latter is probably safe.

4 Likes

Since technological progress is discontinuous, we do not know whether AI’s imperfections today will be solved a month from now or decades from now.
I look forward to AI advances, but what is possible and what is not is a very technical issue and difficult to predict.

It’s troubling for those of us who create entertainment content.

Maybe eventually AI will generate entire movies, and then it will be possible to say, “Generate the movie I need”. For people who have had a fight with a friend, for people who have lost a child, … The characters in the movie might as well be real people.

It makes me wonder what is the point of creating and viewing a fiction that is not real.

There are many cases though, where getting the tech. to work on an overarching level is actually the easy part. The smaller things and the polishing phase come afterward, and that step is sometimes far and away the part that is the hardest and takes the longest. Right now, all of the AI-based technology that delivers useful results make use of and are guided by ‘training wheels’ generated using more traditional algorithms, like OIDN using Blender’s denoising passes.

Case in point, fully autonomous vehicles. It has been proven now that we can make cars that drive themselves from point A to point B, it has not yet been proven that these same cars can go to these same points driving at the speed limit while being defensive if needed (ie. what happens when the light turns green, but someone is about to run the same light from the other direction)?

Don’t forget personal flying machines, we have working concepts for flying cars and human sized drones, but we still have to figure out safety as well as the legalize (How high above your home does property rights extend, do we need skylanes, how will said lanes work)?

Also do not forget nuclear fusion, we have reactors that work on a technical level, but making them useful is an entirely different ballgame.

1 Like

There isn’t enough tritium on earth to power nuclear fusion for any significant length of time, and we don’t currently have a way to generate more than we use.
https://www.science.org/content/article/fusion-power-may-run-fuel-even-gets-started
You picked a good metaphor for AI-generated art, I’d say. Court systems are repeatedly and consistently ruling that AI cannot hold copyright, cannot author a patent… in short, nothing AI generated is commercially viable legally curently, even if it ever gets to a point where it’s visually acceptable in a commercial environment. That may change, but I doubt any of us will see an AI holding copyright in our lifetimes. No copyright, no commercial use. No commercial use, no widespread use. It’s a well-established pattern

Though in the case of a logo generator, this would actually be a good demonstration as to how well AI can work if it was at least guided to an extent.

What I mean by that, the executive will load in the font file and the letters to use, then have the AI make it fancy.

1 Like

For a full explanation of the whole thing…

They are going to release it open source completely democratized. Its about 30 times more efficient than DALLE-2 and will run on a consumer 5GB Graphics card… im sure it wont replace 3D artists it will be a great tool, im sure it could integrate into Blender maybe as a texture generator or a previz concept generator thing… and much more.

They took 250 TB of image data and compressed it down to 2GB, they are using the 10th fastest public supercomputer in the world.

2 Likes

All these images are impressive, I was recently thinking about it and now is the chance to say it.

That AI has the perfect answer which is “42”, and still there is not one perfect answer, but an unlimited amount of them and an unlimited amount of variations for each one.

In the end it all boils down to having the perfect question. And this means that using AI to do the manual labor, simply transfers the amount of work towards the “soft skill” sciences, that are academic and intellectual, and remove away any bit of craftmanship.

Say for example, in the case of creating artwork. You would have to transfer all of the amount of learning towards other places such as: Cinematography, Composition, Framing, Film Studies. Or for creatures: Physiology, Kinesiology, Anatomy, Biology, etc…

Once you get more mental knowledge, you will be really on point while using the A.I. to make something deep and impactful.

1 Like

We are heading towards AGI whether its sentient or not it will encroach on all human skills, but i don’t think that takes value away from human craftsmanship. In most sci-fi future worlds we imagine hand crafted items as having unique value or even more value in the future because machine made items are just so common…Art, music, furniture, watches, cars, etc traditional hand made will hold the most value.
But on the upside if we do have an AGI that can ACTUALLY do all general tasks, then we will be well on our way to a post scarcity society where nearly all goods are practically free and the standard of living rises dramatically… Nobody will be struggling.

Every artist will die starving in the street before there is enough political will to make this utopian pipe dream work. Post scarcity is post capitalism.

You cannot have billionaires without poverty.

4 Likes

Well being a billionaire would be useless when everyone has access to advanced technological goods… Think how much a modern day cellphone would have costed to produce in say the year 1990, probably billions and the internet would not even be there to enable all the things it could do… So basically the idea of the utopian pipedream is that everyone will be technologically rich, just like everyone today with a smartphone is technologically rich.

So we’ll all just eat AI generated pictures of food and sleep in AI generated pictures of houses.

I get this same line from my boomer parents. “Well I never had Netflix or a smartphone, so your generation actually has it better than mine!” Never mind that he was a home owner in his mid 20s.

I see a lot of homeless people today with smartphones. How’s that technological weath working for them? I mean, sure they can pull up imdb and get Full House episode summaries whenever they want, but they can’t get a full night’s sleep without being threatened by the police or their neighbors.

5 Likes

Until then, how do you suggest paying your bills? With dreams? Because in a market with limited demand if you have an artist who needed 5 hours to complete a task and now he takes only one to complete the same task, as a client you don’t want to pay 5 hours work, but the artist will need to find work to complete those other 4 hours he will not receive any more. I just can’t understand the point of artists applauding their own substitution. The fact is: that demand is limited.

Until the perfect dream world comes to save us from starving, there are lots of real bills to pay.

3 Likes

Just a smartphone is not the solution to all world problems on its own… And just a smartphone is not going to allow any homeless person to transform their life in that particular society. But its such a powerful tool it totally depends on the person using it and what they decide to do with it. They are still technologically rich by having it compared to people 30 yrs ago. The have nearly all human knowledge at their fingertips.

You should ask someone who has the misfortune to be homeless how helpful “all human knowledge in the world” is. Spoiler: it’s not. This is a naive take, I think you need more life experience with financial hardship before you can say things like this. As someone who has very literally been days away from starving to death because I couldn’t afford food, as someone who lived in a vermin-infested hotel (although thankfully briefly) with no money to my name trying to keep my family alive, and as someone who has done extensive advocacy work on behalf of the homeless and financially destitute, I can tell you confidently that your optimistic idealism about technology does not hold up in the real world

I’m sorry if this comes across as aggressive, I find your comment upsetting but I’m trying to temper my feelings towards it as much as possible. It’s not often that someone says something here that gets me heated, but this is a profoundly naive take that borders on offensive. There’s no way to sugarcoat that

4 Likes

To be honest im parroting futurists like Ray Kurzweil who always come out with this sort of naive idealism in their talks. Now im not at all stuck in that mindset just because i heard their talks about future technology. But i am hoping the optimistic idealism thing will actually play out because thats our best hope. If it doesn’t then the world will go down a darker path, which i dont like to think about. But the technology is coming if we like it or not.

1 Like

There’s always a gaggle of philosophers ready at any moment to wax poetic about the wonders of the future from on top of their ivory towers. The thing about ivory towers is- only rich people can afford to build them, and only the fortunate have the luxury of having time to sit around and write songs about the glories of the future. If you want to know what’s happening in the world, put your boots on the ground and talk to real people in real life circumstances. I’m not saying there’s something wrong with optimism- it’s essential to have hope in life- but you’re seeing a very small, very privileged piece of the picture here.

That aside, it’s not worth my energy to allow myself to get upset about these things, and I’m definitely off-topic now, so I’m going to mute this thread and excuse myself from further discussion on it

1 Like

Ivory towers? Philosophers?

From where I am, I’m seeing artists working on $800 Best Buy laptops and a free software creating some pretty amazing CG stuff that just 20 years ago would have sounded like pure science fiction to anyone having a conversation similar to this one.

Hell, even just 6 months ago I was trying to figure out ways to feed an image sequence into Photoshop Beta just so that I could get access to a depth map – lo and behold Resolve 18 comes out a couple of months later and gives me exactly what I need! This stuff is happening fast!!

I really don’t understand why, even supposedly technology-minded people seem to have such a hard time wrapping their noggins around the idea that innovation will continue to push the limits of what is possible further and further.

2 Likes

It’s interesting to have these debates on a 3D related forums. Because CG was seen in a similar ways compared to traditional arts. Of course things have changed a lot, and very few artists now doesn’t work at least a bit digitally. Where 30 years ago everyone was using pencil and papers.
But CG as open a lot of cool possibilities, 3D didn’t killed 2D animation, nowaydays they start even to mix each other in new interesting forms.

Probably these AI tools will definitively becomes a new tool in the creative process. It’s adding some functionality and it doesn’t remove anything.

But of course even if it does really interesting results it doesn’t really think. It just learn to reproduce and interpret in a strange way.
So yeah at some point if you want an elaborated result you must push AI result further .
The initial image with the apocalyptic classroom is awesome as a starting point. Super inspiring. But indeed you can’t do anything with it right away. You can give that to a modeler, or do a set for a movie with that as only reference.

This is were a true artist may use that as a starting point and add other layers of interesting things and coherence on top of it. And also makes real variations , changes the tables but not the wall ect…

Where a regular person can only click “generate again” until they find something they like.

So my bet is yeah AI is going to be a new thing in the toolset, but as you said , it will never replace the way we think, our creativity and our way to iterate over an idea, at least not in this form. But this is interesting nonetheless.