Tessera
Folgen

MEAT!

@draculas.dog

and i don't blab any drab gab - i chatter hep patter (toronto-based cartoonist/film student)

30 Posts185 Follower109 Folgt
MEAT!@draculas.dog · last yr.

ive gotten into the habit of taking photos of heavy duty vehicles when i see them while walking which has been great for building up a reference journal but has the unfortunate effect of making me feel like a slackjawed moron when i do it while people are around

MEAT!@draculas.dog · last yr.

man, i have been away for a hot minute huh. check it out--a new guy has hit the jackrunner scene! and there should be a new jackrunner comic coming out some time soon, but dont let hank know i told you that. Rafi won't be in it, anyways.

BildBildBild
Dylan Horrocks@dylanhorrocks.bsky.social · last yr.

Forgive me, but the other day I saw a talk that had some AI-generated illustrations and I ended up staring at them for ages, trying to study them *as art*: their mark-making, aesthetics, surface qualities, etc. Curious to know if my observations match yours? (Sorry I don't have the images to share).

The other day I was watching a talk and the slides were illustrated using generative AI images (one that looked like a painting, the other like a drawing). These slides were on the screen for 20 minutes or more and so I found myself studying the illustrations closely (the first time I've examined an individual AI image for so long!). It was fascinating to try and work out what it was that made them recognisably AI-generated and to examine their visual aesthetics (how they actually look, regardless of how they've been made). At a glance, they were quite attractive and it’s possible many casual viewers wouldn’t realise they were AI-generated—although it was instantly obvious to me and others in the audience; they just had that look. But what is that look? What are those stylistic quirks that told me this was made by a machine? The longer I studied them, the more I felt there was something really tentative and noncommittal about the mark-making, as though the artist was constantly bluffing, trying to make it look just convincing enough—pixel by pixel—in contrast to the intentional mark-making of a human who knows what they're trying to draw or paint (even if they’re not nailing it all the time). There were whole areas that could have been anything: notably dark, smudgy, or unclear, as if to cover up that the artist has no fucking idea what to put there.Also, there was no sense of focused attention: for a human artist, parts of an image are more important and meaningful, and their mark-making is (intuitively) organised around those. Elements and details are foregrounded, rendered with care and conviction; while others provide them with support, or recede—even disappear—altogether. The AI has no real sense of what matters in the image (because there's no awareness of meaning at all) and so it's loathe to leave space empty and tends to fill every square centimetre with something, and every detail is rendered with the same level (or lack) of attention. It made me think of a (fake) mind-reader or psychic, who puts out a flurry of vague statements, watching to see which ones get a positive reaction, adjusting what they're saying based on our reactions, bluffing that they know what they're talking about even though they're really only filling a blank space with an illusion of meaning, just convincing enough to get away with it. The end result can be visually compelling and persuasive, but I'm struck by that underlying lack of conviction and awareness: the tentativeness, the sense of bluff, illusion, fakery. It's in the very surface and texture of the image. Of course, who knows what AI art will look like in a few years (I'm guessing some of the output will be better at hiding the bluff, while a lot will look even more soulless and fake—but really, who the hell knows!).