There Isn’t Time to Use Generative AI
Some thoughts about the tech world’s current bubble and a reminder that things are supposed to be hard.
Look…I get it…you don’t want to read about AI. Hell, I don’t want to write about AI. But, amidst the many current horrors of the world, AI feels like the monster from It Follows: always there in the background, inching closer and closer until it infects, and ultimately, smothers you.
I’m not going to spend time on the boring/obvious stuff: the basic CliffsNotes take on things is that the AI “industry,” especially generative AI, is an overvalued tech product that is being force fed down our throats, jammed into tools and services that don’t need it, and often, actively making them worse and unreliable (looking at you Google search). It’s also comically bad for the environment, requiring massive data centers and servers that suck up a city grid’s worth of power and water so that your aunt on Facebook can post an overly smoothed image of a soldier hugging Jesus.
It’s bleak, folks.
If you want to read more about this topic from an actual reporter, I point you to Ed Zitron’s great newsletter.
I do think that, like NFTs, this AI bubble will eventually burst. Tech companies have dumped in too much money at this point for the technology to disappear entirely, but my hope is that, over time, we will see its use start to wane. Still, though, the damage has already been done, and I’m not just referring to the resources and money wasted.
I’m certainly not the first to point this out, but the big issue with generative AI is how it’s depriving us of critical thought: kids cheat at homework, your boss uses it to write e-mails, some low-level designer generates an image instead of designing it.
I’m not immune to the surface-level lizard brain appeal of this: wouldn’t it be cool if I could produce a great painting without actually having to learn how to paint? I have ideas! I think I know, generally, what would make a pretty picture!
The problem is the idea isn’t the thing—it’s never the thing—the end product is, and that end product has to be the result of the process, otherwise it doesn’t mean anything. You have to suck before you can be good, and, yeah, sucking at something—well, sucks—but it’s the sucking that allows you to learn and become better.
As a mostly bad filmmaker, people will often “pitch” me story ideas for movies. And, when they do so, there’s always this weird moment where they think they are giving me a golden ticket to something amazing. But, if we think of creativity like an iceberg, that initial idea is just the proverbial tip peeking out of the water—the actual “thing” needs to be unveiled little by little until you finally reveal the entire gargantuan mass. AI skips over the chiseling part—the mistakes, the dead ends, the do-overs—and the end product lacks humanity as a result.
When I was teaching myself After Effects and motion design, one of the first things I learned was that it’s not the software that makes stuff good. Andrew Kramer (a legend in the indie visual effects world) would make these funny, easy-to-follow tutorials. They were so popular that you’d see the visual effects or graphics copied and pasted everywhere (guilty!)—demo reels, corporate videos, indie short films. Anybody who had a cracked After Effects license was suddenly a VFX artist! It was proof that people can follow instructions, but few could take those fundamentals to make something that was truly their own. Generative AI is a super-powered version of dumb kids like me regurgitating Andrew Kramer tutorials: yeah, we could reproduce the “thing,” but we didn’t really have the skills to understand what made the “thing” special in the first place.
As someone who is paid to make visual stuff for a living, I realize that certain types of work I do simply for paychecks is gone and probably never coming back: generic explainer videos or simple graphics work especially. And, yeah, that’s a bummer—I have a mortgage and bills to pay just like everyone else. But, I’m hoping the “hard work”—the decision making work—keeps me employed long enough that I can survive until retirement (that is, if society makes it that long).
When people talk about how “easy” generative AI can make something, I think of (what else?) a movie quote. This one from A League of their Own to be specific.
To be clear, I’m not against all uses of “AI” technology and I think that most people’s understanding of AI is, unfortunately, tainted by improper and generalized vocabulary.
For instance, Generative AI? That’s gross.
But, machine learning tools/algorithms are super useful (and have been for decades). Plugins that remove background noise in an audio file? Great! Filters that can help upscale an image? I’m cool with that!
As a practical example, it used to take me a while to mask out a photo in Photoshop or rotoscope something in After Effects and now it takes seconds due to more advanced plugins. Similarly, it’s also way easier to create subtitles or captions for videos, which I think, in general, is good for everyone. I’m okay with “thoughtless” work being streamlined or becoming more efficient via better algorithms.
But, the thinking stuff? The hit-your-head-against-the-wall, wow-writing-is-actually-really-difficult, hard stuff? That should never go away.
The hard is what makes it great.