Stop whining about AI art

We must resist snobs trying to gatekeep creation

A few weeks ago I launched a broadside on behalf of artists against philistine conservative chuds who don’t understand art and regularly alienate creative young people by adopting provincial and unreflective attitudes toward culture and society.

Today I’m pointing my cannons in the other direction.

Not against artists as a whole, of course, but against a certain species of snooty luddite contrarian that’s recently begun popping up on this part of Substack. These guys want you to know they have a Big Fucking Problem with AI generated content:

A Forest Rebel won’t be reading any of this due to my use of that AI-generated crying artist thumbnail above, so let me take this opportunity to say I have it on good authority that he eats his boogers and pees sitting down.

He’s also hardly alone in this view.

When I first posted my tract against conservative philistinism, I received this charming comment from one Michael Vincent Hawthorne:

More recently, Hawthorne has expressed a sentiment almost identical to that of AFR, in his beautifully titled Reject AI Bullshit, Embrace Human Jank:

I hate artificial generative intelligence.

I hate it a lot.

I hate it so much, not only have I written about the subject, but I hold a pretty firm, no AI policy on Substack.

If you use AI in your thumbnails, I don’t read your content, I don’t follow, and I certainly don’t subscribe. If I see someone I do follow or subscribe to use it, I’m out.

Like many of the Zoomers marching for Palestine, Hawthorne refuses to respect or seriously engage with anyone who can’t meet his own puritanical standard, and in doing so he castrates himself rhetorically and fatally limits his own message.

If he were more willing to reach out in the spirit of polite and friendly debate he’d have a lot more success, but with this dour and aggressive attitude he limits himself to a ghetto where he can’t reach even the folks who agree with him.

And that’s why Hawthorne will almost certainly earn more followers from this essay attacking him than he has from any of his own articles.

---

It honestly makes perfect sense this would happen to a Luddite. These aren’t strategic, adaptive, or mentally flexible people. Hell, I wouldn’t be surprised if Hawthorne instinctively associates such attributes with shiftiness and a lack of honor.

The sort of person who displays an extreme hostility toward a promising new tool is also the sort of person who is bad at solving problems and accomplishing his goals more generally. The Luddite never learns the tool himself so he can more reliably assert his will to power on the world. Instead he sits around impotently complaining about everything going in a direction he doesn’t like, and hilariously tries to position his anger as some kind of rugged righteous indignation.

But perhaps I’m being unfair.

One can tell from Hawthorne’s writing he isn’t a dumb guy by any means. He has a crisp and distinctive voice (I’m assuming a background in journalism?) and boasts a fairly extensive vocabulary. I’d probably estimate his IQ in the 115-120 range.

But he also has a lot of pretty embarrassing takes that suggest he’s relatively low in openness to experience, particularly for his IQ. The guy is certainly a disciplined thinker (and is clearly disagreeable enough to spit contrarian takes, which I obviously respect), but he is by no means a creative or adaptable one. He is rigid. Brittle.

And I think that’s a big reason his footprint is so small relative to the quality of his writing, the breadth of his publications, and the length of his tenure on Substack.

Because low openness is what makes you chose thumbnails like these:

See what I mean? Horribly off-putting. The color scheme of every image is painful to the eye, and you can literally see a watermark in the central thumbnail. It looks like something an autistic nine year-old put together at summer camp.

If he fixed this shit he’d easily have ten times the follower count. But like most low openness people, Hawthorne is actually proud of his stubborn incompetence:

My thumbnails are janky. They are fucked up. The thumbnail for my most successful article (to date) is photoshopped with a picture of Hillary Clinton that couldn’t get more pixelated. My logo is shoved in the corner, and it’s way too fucking big. It’s a “graphic design is my passion” joke of a thumbnail.

I cooked that image up, absolutely hammered trying to get the article published, and thought, “That’d be funny. Let's do that,” and I did it.

Do I put more effort into thumbnails? Absolutely.

But an actual human put those together. I found the images and put them together without some AGI cooked up by a misanthropic tech bro stealing from a hardworking artist.

Michael. A question. What do you think that fugly watermark is for?

Do you ever check the copyright status of the random ugly shit you pull from google images? If not, you might want to get on that, because if you’re currently making anything off your Substack you might have some photographers and digital media editors who aren’t especially happy with you. You are “stealing from a hardworking artist” a lot more directly than I am.

Also, I refuse to indulge the retarded idea that LLMs are stealing at all. If AI is “stealing from artists,” then so is every human artist whenever he is (deliberately or accidentally) inspired by the aesthetic sensibilities of another artist.

A Large Language Model is nothing more than a digital manifestation of what Carl Jung called the Collective Unconscious. When you say to Dall-E, “draw a chair,” this tool will produce something that basically resembles the human psychological aggregate of all images corresponding to the form “chair.”

So if it’s “stealing” to train an LLM on human art, it’s likewise “stealing” for anyone who wants to create art to consume someone else’s art and draw inspiration from it.

I put them together the way I wanted them to come out. I didn’t hope that an apathetic machine would do it for me at the expense of another living, breathing human being who worked to put their vision to paper.

No Michael.

You waited for someone else to take a photograph, someone else to edit that photograph in Photoshop, and someone else to take care of the legal shit and upload it to the internet through their big legacy media publication. You then copied the photograph from google image search and made a few shitty alterations in an image processor programmed by another “misanthropic tech bro.”

When I use an AI generated image for my thumbnail, I typically will explore a lot of different concepts, and typically need to get very creative with the prompt engineering to produce exactly what I want.

This is oftentimes very tricky because I always have to think in multiple dimensions (i.e. what best matches the content of my article, what is most engaging, what coheres with other images on my home page…)

This requires a lot more creative vision and genuine effort than anything you’re doing, but the results speak for themselves. Just look at how quickly my articles have made a splash. A major reason for this is that I use highly engaging thumbnails like this:

My approach is objectively less plagiaristic, objectively more effective, and very obviously more artistically meritorious.

---

Ultimately Hawthorne is an easy target, so I’ll let up on him. I’m honestly a lot more pissed at the actual artists I occasionally see buying into these horrific arguments.

Most of these people just seem like gatekeeping snobs who don’t actually value art as a philosophical ideal, and are just selfishly agitating against something that threatens their ability to passively coast without intelligently adapting to changes in the market. Genuinely talented artists have always been able to adapt, and there’s no reason they should stop now. Operating under dynamism and constraints makes art better.

But even if AI put every existing artist out of business permanently I would still support it, because in practice it is democratizing the act of creation to a staggering degree. It lets you implement a fantastical concept without any niche technical training or expensive gear, and this will allow millions of talented people to pursue their creative aspirations in a way that was never remotely possible before.

This is amazing for artists who are high openness, novelty-seeking “idea guys” like me. But it’s also *terrible* for low openness “craftsmen” type artists who don’t have a lot of great ideas themselves, and mostly earn a living realizing someone else’s vision.

You know who really hates AI? People who used to make a lot of money drawing furry porn or Dungeons and Dragons character sheets for wealthy but aesthetically moronic software engineers. These days the SWE chuds can do that themselves with Stable Diffusion, so a lot of hipsters with no marketable skills have suddenly lost their comfy sinecure. These are 99% of the dweebs you see bitching about AI art tools.

Actually talented and inspired artists are not threatened by AI one iota. They can easily ignore it and continue to make great art the old fashioned way if they want. But a lot of them actually love it, because it allows them to upscale their existing processes and massively reduce costs by “insourcing” parts of their art once sent to technicians.

This development was incredibly useful for me given my background as a song parodist. I am pretty decent at writing funny and compelling lyrics, but I have zero training as a composer. I’m also a pretty amateurish singer, while my talents in audio editing aren’t much more extensive than a basic knowledge of Audacity.

But thanks to AI that doesn’t matter, because I now have a robotic Sullivan to my Gilbert plus an entire orchestra at my disposal, so I can create cool shit like this:

Now maybe you don’t like any of these songs.

That’s fine—other people do, and ultimately I’m just creating them to have fun.

But if you’re going to tell me it’s somehow immoral or degenerate not to develop those other skills myself (or instead spend thousands of dollars on a composer, performers, and a professional audio editor) then I’m just going to call you a fucking asshole.

The alternative wasn’t me hiring those people, it was me not being able to create at all.

AI has democratized art in a way that is massively salient to me personally, and I take personal umbrage at anyone opposing this. I also feel incredible contempt for anyone so ignorant as to assume the machine is doing *literally everything*, which ignores in an incredibly pigheaded way the role of the artist as producer and curator.

When it comes to my AI generated songs, I usually write the lyrics entirely myself, and that is what I care most about artistically. I let the machine handle what I don’t have the technical skills to do myself, using creative and iterative prompt engineering and rigorous curation to make sure it sounds decent in the end. I’ll admit this aspect of creation puts me in the role of producer rather than artist, but the final product is a legitimate work of art and a genuine reflection of my aesthetic sensibilities.

When it comes to my essays, the art is obviously in the prose. That I write entirely by hand, and would never dream of outsourcing to Chat GPT. There is also obviously a lot of creativity involved in prompt engineering the thumbnail (as described above), but I similarly view this as more of a marketing / production activity. It’s important, and authentically creative in an entirely meritorious way, but by itself is not “art” per se. But holistically speaking, the final product to which it contributes is obviously art.

The people who bigotedly dismiss this sort of creation as inherently less valid are ignorant and dangerous. This impulse is not that of a genuinely inspired artist, who can and will adapt to any technological ecology. It is the province of bigoted chuds, gatekeeping snobs, and broke furry porn technicians.

At the end of the day, if we say it’s illegitimate for artists to use this specific tool, we’ll need to say the same thing about the camera and applications like Photoshop.

This was, of course, frequently said of such tools when they first emerged.

And today we rightly consign such opinions to the dustbin of history.