A cartoonist's review of AI art, by Matthew Inman | Hacker News Reader

A cartoonist's review of AI art, by Matthew Inman

theoatmeal.com

I don’t think that toil makes good art, but choices do. Artists make choices and those choices add up to something unique and tasteful. The choices are influenced by constraints and the desire to communicate something. There is a “why”, not just a “what”.

I find that AI art misses both of those, and that makes it feel soulless. No decisions were made by a thinking, feeling human. There is no “why”.

I find there is often very much a why with lots of the AI art I see. I often use it to try and visualize a dream I or one of my friends had. When I get the prompt just right it's sometimes striking how close I get to what I was seeing.

My friends will often use it for humorous situations that are happening to us in the moment. Overexagerations or putting one of us in a crazy situation. That's about as meaningful as using a gif keyboard, sure, but it gets a pretty good laugh out of me sometimes. There are recurring ones that have become a type of shorthand for us.

I've used it for highly personalized Christmas cards and anniversary invites for other people that they just otherwise would never have made because they don't have $500 to pay an artist and Clipart is too impersonal to bother with.

I've turned my mother's drawing into a beautiful realistic picture that she absolutely adored. I turned a baby picture from decades ago into a movie of a kid laughing which made my mother instinctively grab the phone from my hand with a big grin.

To me AI art is only about the why. It brings entertainment, comfort, hilarity and agency. I would never charge anybody for it I'm just doing it because I find it a neat set of tools to mess around with. And the tools aren't just asking chatgpt to make something. You can spend months refining the craft of it similar to learning something like Photoshop.

I am professionally interested in the narrative structure of texts, as a teacher and researcher, so I ask this out of genuine interest. I often see posts like this in discussions of LLM use and I wonder where they are coming from. They read like ad copy from a marketing department rather than a non-professional expressing personal feelings. Are you in marketing, or perhaps was this post generated by LLM?

Neither. And this is another bias I keep seeing. Anyone who actually uses the technology in the way I do must be using it to write (I don't) or a shill. Nope I just have been using it for various things since the early days and I wonder why people are so allergic to being able to express themselves clearer and with more agency. Not everything has to be about money. LLMs and image generators can just be fun.

I like that my writing is considered well structured / grammatically correct enough to be llm worthy though. I often think I make several obvious errors when I write off the cuff internet posts (and probably did).

> No decisions were made by a thinking, feeling human.

Okay, but there was a human that typed or dictated some stuff into a computer. That person doesn't think or feel, or like...?

I'm reading some literature(fictions and non-fictions) reviews these days, and realized that literature itself, is just recording our lives in thousands or even millions of different views. The events in the real life can be similar or even identical, but will finally result in many different books, some of them are bad while others are masterpieces. Then I suddenly realized that why I hate some AI work as a long time content consumer, because that the creator behind them are just utilizing AI as a "tool", to quickly generate something meanlingless only for sensory stimulation, which stays at the surface-level, becoming a sort of cheap sensationalism works.

In contrast, I must admit that there are some AI assisted creations really shine , for example, generate an AR annotated POI image with nano banana(https://x.com/bilawalsidhu/status/1960529167742853378). But sadly, there are only 1% of creations, regardless it's an image, an audio or a video, are good, inspiring and exciting as previous ones.

Before AI can get a consciousness, it's a tool, no matter how "smart" it looks like. Only the human who use the tool smartly will create outstanding works.

I do think it's, like really long, but he makes some good points.

Yeah, he kind of lied about the length.

I would let it slip, but it's obvious that the content is optimised for smartphones/tablets too, so kind of painful to scroll on a computer.

As such, I sentence this post to be ignored (by me).

The irony is that an AI would do a great job summarising it too...

AI would do a horrendous job of summarizing it; the tone/impact wouldn’t hit without the visuals and timing/pacing.

>nobody seems to want it

I've had a lot of genuine fun playing with AI art generators dating all the way back to deepdream. I love the tech, I want it as it was, as it is, and as it will become. This tech have in the last few years given me much more joy than any artist have come close to. It shares creative powers freely, a far cry from the overly commercial streaks dominating much else.

What I don't want is to see yet another has-been meme artist rehash the same anti-AI tirade that we've seen so frequently that any given LLM could re-create it verbatim due to overfitting.

All talk about "Human originality, soul, heart, the divine spark" and yet all they display is hysteria.

I’m not going to pretend AI isn’t useful, but I’m begging you to look deeper into the arts! That this tech has “given you much more joy than any artist have come close to” is a pretty devastating statement. Connecting deeply with an artist/artwork is imo one of the great experiences of being human.

So tired of pointing that there is something in between "marketeer story tellers" and "real artists".

Bots in the Hall, Neural Viz, GossipGoblin, even Joel Havers animations. All made with genAI and all undeniably creative works, that could not have been created, at least not in that time frame, by a single person without it.

I love Matt's work and often agree with him, but the "no heart" take is just too harsh.

People being able to express their imaginations in a more clear manner always seems like a net positive to me. You can use it without heart and for the exact wrong reasons for sure.

The keyboard example is telling because that's basically what a bunch of our favorite musicians now do. The drum track for that song you're listening to? That's a piece of software not drums.

And plenty of people complain about "overtuned" music with robotic drums and overcorrected vocals that drain all the imperfections out of the sound.

(Here's a musician's take on it: https://youtu.be/BDJF4lR3_eg?si=kKJVF2hqSd-TOEzX)

"Being able to express your imagination more clearly" is fine for what it is but that's not how people use it. His point about the least talented kinds of people using it to pretend to art is dead accurate. People think it's snobbishness about art being this superior untouchable thing, and that is absolutely not it. You see the same thing with people going on about how SWEs will be obsolete, too, because something something code generation.

Making anything worthwhile involves thinking and choices and experience, and that sort of person consider that too hard and boring and just want something that looks worthwhile.

The idea that you can't both use AI heavily and use thinking, make choices, and use experience is the part I think I disagree with. People sometimes use AI to make thoughtless slop, that doesn't mean everyone does.

I think we're going to come into contact with brilliant projects that people have spent years making and there would just be no way they could've done them without AI. It will be a human thinking it through making thousands of choices just like the OP said about Jurassic Park. You can and basically have to guide the AI like crazy to get any output worthwhile.

Exactly this. In the early days of image generation I once spent about 9 hours carefully remastering a generated image until I was happy with the result.

Are those hours completely without value, because I used genAI during most of them? It involved a LOT of consideration and a few pretty tough choices. But because I didn't do my 10.000 hours, and it's sad for those that did, I'm not allowed to create this way?

Whatever, I'm still putting it on my wall. It means more to me than most "legit" art does.

Indeed. I think the issue is so much of what people see is just the 'slop', which can be produced in such volume by people who really don't care much for the outcome. It distorts the perception of it a lot.

(similarly with the comment about 'it's about as hard as using google'. Yeah, if you're trying about as hard as googling something then the result is probably not going to be very interesting, but that's more effort than went into most AI generated images that you'll see on the internet)

The actual title is A cartoonist's review of AI art.

This is true, but I thought the interpreted title might be something that could pique interest. I guess I could change it, but I don't think it's gonna hit the front page anyway.

UPDATED: I asked them to change it; not that it will make much of a difference, but you've got a point.

Ok, changed now. (Submitted title, for those who care to track such things, was "A Long Screed About AI Art, by Matthew Inman".)

Thanks!

AI art mostly doesn't make use of the medium. There are things you can do with computer vision and user interaction that just aren't possible with traditional media.

I remember the game PlaneShift, which has an interesting not-for-profit development model, and which uses natural language processing to handle text. The possibilities here are really impressive! You could fully voice player characters, parse user intent with better accuracy, etc.

There's all sorts of stuff you could do. Introspecting the models themselves gave us DeepDream. Setting early models up to talk to each other caused headlines about AI inventing its own language. Leaning into the unhinged nature of the medium interests me, as does displaying "thought" graphically.

Even back in the ML days, I had high hopes for this sort of thing. Discarding those hopes because of AI slop feels like throwing the baby out with the bathwater. The whole problem is that people use an amazing (but glitchy and imperfect) technology and use it to make bad art in existing fields, instead of pioneering good art in new fields. It's herd mentality from both sides.

Lost me with the initial observation... it seems that merely being "AI art" is paramount to any feeling the art could evoke. To the author, just knowing that the art is AI generated is "gross". I feel like this bias is all too common; especially in "artist" circles.

This is mostly commentary on the state of Matthew's knowledge of AI tooling. You can use preprocessors, train your own text encoders, build pipelines to automate tasks that you might not want to dwell on so you can focus on art.

AI is just another tool. Not everyone who uses a computer is an artist, but you'd be hard pressed to find an artist that doesn't use a computer who doesn't make it a point not to.

When I think of the golden age of motion graphics (pfft), I'm not thinking about fleets of pale interns toiling in Adobe After Effects to produce choppy renders at the wrong resolutions because they're still clinging to their college MacBooks.

I'm thinking about the analog system that After Effects was created as a digital, computerized abstraction of— Scanimate. And I think about the confluence of tech that enabled that and the people who seized upon a problem and solved it. It's entirely possible to do what these tools do completely manually and traditionally, but where are the high moral purists expressing disappointment in the spiritual weaklings who rely on computers to create art? When did doing become diminutive?

I love that AI can break the individualization of struggle that some artists believe in, because what's really important is growth and evaluation. Matthew was successful, and now he's insulated and grumpy. He leaned into a strength that became his brand, and I think with how little he's left his comfort zone, this Ayn Rand phase was inevitable. He's not going to give it up and start over to prove he's a "real artist", but he'll tell people doing art in other ways (most of whom will not become successful or profitable but who still have done) that they're not legitimate for exploring differently.

Art is still hard. You can't cram for art school. You have to be able to prove yourself, continuously, to your clients, your audience, your family, and yourself. Art is a pursuit in virtuosity of the fundamentals, even if other people don't like what or how you do it. And if Matthew stands by his message, then the people struggling under the disapproval of people like him who may not help you or want to see you succeed, and don't give up are doing exactly what forges an artist's spirit.

Reply on news.ycombinator.com