@CSA2D7: Discussion on Sam Spratt's Masquerade and generative vs AI art

Submit
@CSA2D7
@CSA2D7

Not a McCarren Park Bench chat, but a really enjoyable discussion with @KevinBuist and @DavidLisser about Sam Spratt's Masquerade and generative vs. generative AI (forms and shapes vs. figurative, cryptopunks, bored apes, Casey Reas, Tyler Hobbs, Dmitri Cherniak, Sam Spratt): porcupineschool Julia, the curator, talks about the inclusion of Spratt in this interview. She says, "To me, this represented the most contemporary version of how an artist might work with a generative system." But I don't really buy that. I don't buy the idea that generative art is a stepping stone on the way to generative AI image-making. CSA2D7 Yeah, I agree with you To be fair to Julia - I don't know how you negotiate that differently - do you say that you have stakeholder interests (probably the genuine answer) or do you try to bridge a gap that might just be a little too large (for those who are aware) What is curatorial integrity - is a difficult question for me to answer as I am not in that position porcupineschool I think making generative art now is a deliberate choice to use systems that are not cutting edge, that are actually kind of old. Making art with bleeding edge generative AI is actually something else. More vibes, less instructions. CSA2D7 also, I agree in this specific circumstances with Masquerade - I don't know what the future holds for linking to generative AI CSA2D7 I see, so you're drawing broader distinguishing themes, I am mostly speaking about masquerade, but I get your point about randomness porcupineschool are you saying it's in the show because the collectors wanted it to be? CSA2D7 yea, that's what I think (my opinion). I also actually like the work Liking the work is distinct from themes for me porcupineschool I'm not sure if I like the work, to be honest. It seemed misplaced, so maybe I would like it more if it were properly contextualized. CSA2D7 I like the work because I am a big network and participatory person... I am not thinking about the aesthetics so much... but I think the networked folks are a FAN of the aesthetics so the line is actually not that clear if I am admiring it for its networked aspects "Sam’s not relinquishing control 100% to the generative process" - is there a magic number on this scale between 0 - 100 where you think the link between generative and generative AI becomes stronger in your eyes? porcupineschool I'm not sure. I guess generative AI broadens the possibilities of "generative" in a way that's very weird. I typically think of generative as a kind of controlled randomness. Where you can see both what the artist controls and what they don't. There's a legibility to it. Good examples are Cherniak, Reas, Hobbs, etc. But with generative AI you can make things with waaaay less input from the artist, and the parameters are completely opaque. You're not really sure what is or isn't confining the outputs of the system. Where as with more traditional generative art, I think part of the pleasure is that you can get a sense for what the rules are just by looking at the outputs. CSA2D7 Ok - how about manual vs. automatic in a car as a framework (in your eyes) to describe this distinction? porcupineschool I think that difference is not big enough CSA2D7 Do you think you get a better sense for the rules are just by looking at the outputs (w generative art) is because generative art tends to focus on shapes and forms and abstractions vs. AI which tends to be more figurative? porcupineschool When the automatic transmission shifts I can tell what it's doing CSA2D7 Hehe, this was just another way to frame the degree vs scale framework Reas, Hobbs, Cherniak, almost all of their generative works are shapes and abstractions But when Casey uses AI its almost all figurative porcupineschool No, not really. In fact, Cryptopunks were in the show, and those are figurative but the parameters are still very legible. CSA2D7 Ah, but this is the whole pfp face portraiture thing. Very simple The face is pixels, so really when broken down, just squares porcupineschool Yeah it is simple, but that's sort of the point. The viewer can reverse-engineer the logic just my looking at it. It's a very different experience to look at generative AI output and be like "i have no idea how the hell it came up with that" CSA2D7 Agreed! Reverse engineering in your mind is easier because it is shapes! porcupineschool But the same is true for Bored Apes It works just as well with illustrations that layer and combines CSA2D7 Could we not argue that the Bored Ape is really just a simple form? porcupineschool yes, for sure CSA2D7 I think so! Its the form of the monkey, to me that is closer to shape than figurative porcupineschool i'm just saying it can work with figurative imagery (simple imagery) CSA2D7 Maybe, but its a very limited figurative image (which you concede) I think broadly speaking the shape/abstraction vs. figurative framework is useful for generative vs. AI davidlisser For me, the legibility of generative systems is part of the appeal. If someone can create a fairly simple set of rules (that I can largely understand), and generate hundreds of pleasing images that (largely) still maintain their own character, then I find that quite impressive. Simplicity is difficult. If someone is using genAI and whatever model they are using has been trained on the 5bn+ images from Laion, then I'm not really surprised they can come up with something good to look at. For me, personally, some of the appeal of generative art is the constraints and sparsity of the system porcupineschool Yes!! CSA2D7 To be fair to Masquerade - in this instance I think it was entirely trained on Sam's art? I mean the output itself looks very much like Sam's work - I immediately thought Sam Spratt when I saw it? porcupineschool I think there is room for both, systems that are spare and systems that are hugely complex. What I took issue with was the idea that the more complex systems represented the future, or where generative art was ultimatley headed. I don't think so. CSA2D7 Ah, ok - I think we take issue with Julia's selection for different reasons then

█▓▒­░CSA2D7░▒▓█ (@CSA2D7)

A good example of a 1/1 that I think plays with new ideas (and also happens to be displayed prominently in the Toledo museum) is Sam Spratt’s “Masquerade.” A 1/1 by hand, replicating a sort of painterly aesthetic, but networked and participatory. Now, we can go deeper and explore…

xcancel.com

@CSA2D7: Discussion on Sam Spratt's Masquerade and generative vs AI art

AI summary of the linked article

This is a discussion thread among CSA2D7, porcupineschool, and davidlisser about Sam Spratt's work in the exhibition Masquerade and the distinction between generative and generative AI art. Porcupineschool disputes curator Julia's claim that Spratt's inclusion represented "the most contemporary version of how an artist might work with a generative system," arguing that generative art is not a stepping stone to generative AI. Porcupineschool describes generative art as controlled randomness with legible rules, while generative AI has opaque parameters, and cites CryptoPunks and Bored Apes as figurative examples that remain legible. CSA2D7 argues that the shape and abstraction versus figurative framework is useful for distinguishing the two, and says Masquerade's output looks very much like Sam's own work.

No comments yetWhat caught your eye? Start the discussion.