back
41 comments
I'm not really convinced that the story of Nvidia's origin does have the profound meaning the author seems to think it does.

> the Nvidia GPUs [...] were originally inspired by [...] the needs of computer graphics artists in the gaming industry.

I think this passage demonstrates a misread of the quoted section[1] from the interview with Huang. The RIVA 128 was a 3D accelerator / video card for gamers, not game artists. Further, it was pretty clear by 1997 that 3D graphics for games was not a fad nor by that time was it a radical new approach -- Nvidia successfully entered an established market.

[1] Huang liked video games and thought that there was a market for better graphics chips. Instead of drawing pixels by hand, artists were starting to assemble three-dimensional polygons out of shapes known as “primitives,” saving time and effort but requiring new chips.

Yeah, this history is just wrong. What really happened is as so:

Early 90s: SGI invented OpenGL to make realtime 3D graphics practical, initially for CAD/CAM and other scientific/engineering pursuits, and started shipping expensive workstations with 3d accelerated graphics. Some game artists used these workstations to prerender 3d graphics for game consoles. Note that 2D CAD/CAM accelerators had already been in market for nearly a decade, as had game consoles with varying degrees of 2D acceleration.

Mid-90s: Arcades and consoles starting using SGI chips and/or chip designs to render 3d games in real time. 3DFx, founded by ex-SGI engineers, created the Voodoo accelerator to bring the technology down market to the PC for PC games, which was a rapidly growing market.

Late 90s: NVIDIA entered the already existing and growing market for OpenGL accelerators for 3D PC gaming. This was a fast-follow technical play. They competed with 3DFx on performance and won after 3DFx fell behind and made serious strategy mistakes.

Later 90s: NVIDIA created the “GPU” branding to draw attention to their addition of hardware texture and lighting support, which 3DFX didn’t have. Really this was more of an incremental improvement in gaming capability.

Early 00s: NVIDIA nearly lost their lead to ATI with the switch to the shader model and DirectX 9, and had to redesign their architecture. ATI is now part of AMD and continues to compete with NVIDIA.

Mid 00s: NVIDIA releases CUDA, which adapts shaders to general purpose computation, completing the circle in a sense and making NVIDIA GPUs more useful for scientific work like the original SGI workstations. This later enabled the crypto boom and now generative AI.

Of course, along the way, OpenGL and GPUs have been used a lot for art, including art in games, but at no point did anybody say "hey, a lot of artists are trying to make 3D art, we should make graphics hardware for artists". Graphics hardware was made to render games faster with higher fidelity.

Author here - thank you for this. I definitely don't claim to be an expert on the history of 3d graphics, and you clearly know a lot more than me about the detailed history of NVIDIA.

That said, starting in the early 1990s is missing the whole first half of the story, no? Searching Google Books with a 1980-1990 date range for things like "3d graphics" "art" or "3d graphics" "special effects" yields a lot of primary sources that indicate that creative applications were driving demand for chips and workstations that focused on graphics. For instance this is from a trade journal for TV producers in 1987: "Perhaps the greatest dilemma facing the industrial producer today is what to do about digital graphics... because special effects, 2d painting, and 3d animation all rely on basically the same kind of hardware, it should be possible to design a 'graphics computer' that can handle several different kinds of functions." [https://www.google.com/books/edition/E_ITV/0JRYAAAAYAAJ?hl=e...]

It's not hard to find more examples like this from the 1985-1989 period.

The idea didn't spring fully formed from SGI. It was a natural extension of 2D graphics accelerators which were initially used for engineering (high value, small market) and later for business applications generally and games (lower value, large markets). 3D acceleration took the exact same path, but the utility for gaming was much higher than the general business utility.

Of course graphics hardware was also used for more creative purposes including desktop publishing, special effects for TV, and digital art, so you will find some people in those communities vaguely wishing for something better, but artistic creation, even for commercial purpose, was never the market driver of 3D acceleration. Games were. The hardware was designed for gamers first, game programmers second, game artists a distant third, and for nobody else.

The closest thing to an "art computer" around that time was the Amiga which targeted the design/audio/video production markets.

Hi Ben,

It was mostly gamers. As a gamer from that time, the hardware was marketed to gamers, hard. I don't doubt that artists had an impact, but the world had many, many more gamers, than artists and gamers spend money for the best/mostest/etc.

I mainly know this from living through the CGA/EGA/VGA/SVGA/3D add-on card/3D era.

Thank you for taking the time to delve into this. While I may not agree with your conclusions, I respect your work, and the effort put in. :)

I think we agree, just define terms differently -- video games are art! In other words, gamers are consumers of artwork, and that consumer demand for a new kind of art drove demand for the hardware to go with it. (Naturally that wasn't the only source of demand - engineering and research applications were there from the beginning too).

Edit: this discussion is interesting because I have always just taken it for granted that video games are a form of art. Clearly others don't see it that way, which is fair! Nevertheless, I think a strong case can be made: https://en.wikipedia.org/wiki/Video_games_as_an_art_form

Games are a medium for artistic expression but saying that 3D hardware was designed to improve art production, or that NVIDIA was first in market, is incorrect. The hardware was designed to improve the consumption experience of something that is a mix of programming, game mechanics (which are both math and psychology), and potentially various art forms including visual, music, and narrative. It all needs to add up to fun or it won’t find much of an audience.

Gamers aren’t primarily spending time or money for the art and neither was NVIDIA. I will grant that the hardware improvements did make the visual aspects more lifelike and detailed and that allowed for increased artistic range, but production costs generally increased accordingly.

Gamers drove the hardware, not artists. 100% agreed.

I was in IT when we started the transition from 2D cards to 2D cards with a additional 3D card, to the final form of 1 graphics card.

These 3D cards were NOT marketed to professionals, they were marketed to gamers, because gamers drove the tech.

Game creators/designers as artists - not unreasonable.
Not unreasonable, no. As plain as day, even. Not sure if you thought I was suggesting otherwise?
As an amateur artist (oil painting), coder, and user of generative AI. I find people's rejection/revulsion of the use of generated images in art odd for several reasons:

- art has always been about taking existing images and ideas and tweaking them to make them your own. As Grayson Perry says "I believe in the Chinese whispers approach to art, I see something, copy it and then I make it a little better"

- very renowned artists such as Peter Doig use 'found images' i.e. a photo of people in a boat as the prime subjects in his paintings. Yet if I generate 100 images with my own prompts then that is somehow cheating?

- as the article says, artists have always used whatever tools they can find to make better art. If a trained random number generator helps me produce the image I want to produce then how is that any different?

I can't speak for everyone, but the reason I personally find generated images completely uninteresting from an artistic standpoint is that they are low-effort. I appreciate art because a human sacrificed time and effort to make it. As a general rule of thumb (with exceptions, of course), the more time and effort an artist sacrifices to make a piece of art, the more I find that art interesting, especially if the artist is highly skilled.

I don't think that AI art is fundamentally uninteresting (I am a machine learning researcher, so I think it's rad), or that it can't be used as a tool to make great art. I just think that the final image is not the point of art. The point is that a person practiced for hundreds/thousands of hours to develop a creative skill and then spent several more hours painstakingly applying that skill just to create the image. If AI is incorporated into a high-effort, high-skill process, I don't have an issue with it.

People have varying definitions of art. Mine is more about communication and changing perceptions, I couldn't care less about time spent. I know artists from both camps.
And this is why I began my comment with "I can't speak for everyone!" Art is so subjective that it would be rather silly of me to try to make a bold, authoritative statement about its nature. I definitely understand your perspective to some degree, and wish you the best :)
Also interesting is the time spent making marks vs. thinking about it

I know Matisse kept several works in his studio for many many years pondering and struggling with what to do with them and the actual mark-making was very rapid e.g. The Red Studio painting

I suppose that's the crux of it, I'm a formalist (https://www.tate.org.uk/art/art-terms/f/formalism) whereas a lot of other people appreciate the back story or narrative more
The problem I have with the 'effort' theory of value - which I also used to use! - is that you can take it to absurd conclusions.

For example, a beginner painter will take much longer to make the same painting as a master; so judged on effort, it should be better? Or if I make a minature cathedral out of matchsticks, then it might take me decades - is that therefore more interesting artistic output than a sketch that Picasso (say) made in one smooth line that took minutes?

Of course, a reasonable answer is that it is some combination of effort, aesthetics, personal choice, and fashion. Maybe it's like a recipe where the balance of these ingredients has to be right.

To a certain extent, your argument is mitigated by considering the cumulative effort of learning to produce art, AKA "knowing where to hammer", AKA "it took a lot of hard work to make this look easy".
If we incorporate cumulative effort, then the original assertion that AI art is low-effort falls flat.

In fact, since AI is trained on a much bigger data set than humans, AI art is in fact higher-effort than human-made art (and therefore, according to the GP, more interesting).

This argument rests on your views of labor theory of property, which suggests that you "own something" by "putting labor into" that thing.

We loosely apply this as a premise for intellectual property rights, but it's contingent on the state being the referee - and many cases have ended in rights being taken away from inventors because someone else worked the legal system to take it away.

In the case of generative training data, it's largely being used as a commons resource. The more you are inclined towards labor theory of property, the more it "feels like" theft, because the original laborer would be credited as owner - the labor of the machine is derivative of that labor, and largely non-transformative since it results in imagery resembling the original, versus metadata, statistics, etc.

Labor theory of property has many philosophical issues, as outlined with the "monkey selfie" case:

https://en.m.wikipedia.org/wiki/Monkey_selfie_copyright_disp...

That seems like a pretty huge stretch to me. I don't think it's valid to count the effort in developing and training AI as part of the effort to produce the art.
> I don't think it's valid to count the effort in developing and training AI as part of the effort to produce the art.

But that is what webmaven was proposing. "your argument is mitigated by considering the cumulative effort of learning to produce art"

So do you count the training and practice a human and the AI took in order to be able to produce the art, or do you not count either of them?

But then, that effort is rarely sustained by the "Ai artist" himself. It is usually someone else that builds and trains the model.
There's also the aspect of the time it takes to learn to draw/paint well, sometimes the simplest-looking things take the longest to learn

"It took me four years to paint like Raphael, but a lifetime to paint like a child" - Picasso

I definitely like this perspective as a personal take, AI should not stop anyone from pursuing their own artistic goals. The problem comes from a societal scale. Peter Doig would get a very different response if he showed off someone else's picture of a boat to a crowd or tried to sell it. The inspiration and the painting he created are clearly two different things and are recieved differently.

If you use AI, very few people, if any, could tell if you or an AI emulating you made a certain painting. I think this is a key source of the rejection, it's saying "Even if you can't tell, you can trust my paintings were made by me and not an AI, I stake my reputation on it."

It occurred to me the other day that there are (at least) 3 different meanings when referring to the value of a work of art:

* Aesthetically pleasing - often there is a traditionalist idea that only landscapes and portraits are 'real' art, and everything else is rubbish

* Financially successful - the market sets the price of a work, and that is the value, no matter what your personal idea of it

* Emotionally affecting - if the work is moving in some way, or you get something from it, then it has value

Unfortunately, people using different definitions of value here will likely never agree when these values clash when considering a work of art.

I guess this is a long-winded way of saying that I reckon some people have just defined generated images (or partly generated) as 'not Art' so will never accept them. I know I have changed how I think about some art/artists (such as single-colour painters like Rothko) from 'this is bad' to 'I don't care for this, I see that others do'.

> * Emotionally affecting - if the work is moving in some way, or you get something from it, then it has value

I suppose I fall into this camp and why I find most AI generated images nothing even close to art. Art is truth, art is a communication between the artist and the viewer, good artists do things for their personal reasons. Generative AI is cold and unfeeling. If anything, artists that use AI should embrace that and stop trying to make so many airbrushed, Penthouse style renderings that try to mimic so much that has happened before. I mean you have the power of Gandalf, so why are you making images that look like something that's been done thousands of times over? Otherwise it's content and illustration, which both have their place in the world.

Looking at the spiral image thumbnails used in the article look like something that an artist would have done 100 years ago. I don't see an artist there. I see a content creator. If they were painted by hand I would say the same thing, but be slightly more impressed by the craft of it.

I think the problem is not for "art" artists, who will embrace it as they embraced everything, but more for "commissioned" artists, who paint or draw illustrations for money.
"Found art" is a niche that's irrelevant to most 2d-image art, as a producer or consumer. And that's all text prompting / inpainting / etc is at this point. Vague, overbroad suggestions to the AI that leaves all the interesting and meaningful decisions and choices in the hands of latent noise.

I find illustrations, amateur or professional, interesting and engaging because people's style heavily reflects what they value in the imagery they create - deliberate choices made in pursuit of depicting or conveying an aesthetic, a theme, a subject-category, an idea, a notion. Paths taken in developing their skill = style.

Current AI image generators are too blunt of an instrument to be a part of that process.

The communication capabilities / concept space of image-gen AI needs a couple more paradigm shifts before it can join the pantheon of tools for "making better art". That means pursuing what they really want to depict, not filling space on the canvas with the AI's vague notions of it. Right now "style" in AI art is just a shorthand for surface-level LORA-ripoffs of existing artists, and attempts to statistically blend them together.

I find AI art interesting and also automated mass copyright theft.

It is not settled law at all anywhere that somehow using copyrighted data as training data is legal. The companies doing this assumed it was and moved quickly as fast as possible. They invented novel legal theories (“it’s just like a human”) to justify what they did, then spread it all over the internet to shut down artists questioning the legality.

If you allow AI to skirt copyright law but humans have to live by it, then human labour becomes meaningless except as training data for the next AI model, and all profit making will become concentrated into AI, while humans make nothing or next to nothing. Any innovation made by a person will be copied by the next model, which being closed source will profit only a small number of people.

>If you allow AI to skirt copyright law but humans have to live by it

Huh? At least from the argument you're presenting you're attempting to create "The Right to Read" world by RMS. Copyright is about output, not input. If I output something too close to your copyrighted content, that's the violation. Not me getting a book and reading it and sharing it with every one of my friends. Not looking by your art and being influenced by in in my artworks.

> then human labour becomes meaningless

You're 150 years late for the storyline of John Henry.

The problem here isn't machines can do everything and make us meaningless, where you're breaking is thinking that capitalism can even begin to work under this paradigm shift.

I don’t actually think capitalism can work under this paradigm shift.

Unfortunately for me, that is a bleeding edge understanding of reality that hasn’t propagated to the billions of people who live under capitalism.

That’s the tragedy. AI is a dystopian technology under current economics.

If you change the economics AI is utopian.

But we’re electing neoliberal governments, for the most part, in the West. Million $ question: what’s to be done ?

It sounds like you're talking more about a specific art form, collage, than art generally. Regardless, I am in the camp that "art is what artists do".

I think that a lot of the rejection about generated art comes from the fear that it would become the dominant artistic form rather than just another kind of art.

I watched a YouTube video the other day that expressed a sentiment that I share and that may apply here as well. He was talking about Photoshop rather than generative art: (Paraphrasing despite the quote marks) "As an artist, I love photoshop for the power and flexibility it gives me. As an art appreciator, though, I would never hang photoshop art in my home because the aesthetic is not appealing to me."

> art has always been about taking existing images and ideas and tweaking them to make them your own

It's an open secret in the art world that modifying existing works and putting your own spin on it is a common technique. But blatant copies with no such 'spin' on the original need to be called out. Also there are places like China who do clean room reverse engineering of existing tech and make new tech in the same likeness, but the engineering is completely different from the original. It's how they avoid IP/copyright suits.

If you as an oil painter can't understand why AI generated low-effort dreck is unappealing to so many, i'm surprised. It requires next to nothing to create and anyone can do it with virtually zero effort. Even Jackson Pollock (hated on by so many non-abstract artists as creator of something that "even a chimp can do") on the other hand had to spend hours getting his hands dirty and setting up his enormous canvases. There's something about that creative work process that has appeal, which generating images by typing little bits of text absolutely doesn't.

Or let's take your example of Peter Doig, whose art I happen to like. He may use found photos but comparing this to AI renderings is absurd. He takes those and then ACTUALLY PHYSICALLY PAINTS THEM into a work of art. The difference is enormous.

Now if someone incorporated generative AI into a complex composition of mixed media to construct a creative piece of art that required thinking, planning, some modicum of effort at least, I could understand it. Generating 100 copies of a certain style with a few prompts is definitely not that though.

One big difference between IRL art and computer art is in the breadth (bandwidth?) of the process.

IRL art process is a subtle thing. Analog, fuzzy and mysterious.

Computer art process is crass and pinhole-narrow in comparison.