(Featured image is taken without permission from a Bluesky post which rightfully refused to send link traffic to The Times. Using that because most of the non-copyrighted images of Ozymandias that I could find online are, depressingly predictably, AI-generated).
I touched on it the last time I went off on a rant about generative AI, but there’s one thing in particular that really bugs me about the increasingly-feverish hype bubble that we’re suffering through at the moment: nobody’s even bothering to make the case that generative AI is actually good for anything.
(Outside of a few narrow applications that involve predictable, verifiable results that don’t require actual reasoning or creativity. Which I probably should’ve acknowledged in the first version of this post).
It’s possible that there was an entire stage of it, and I just missed it. I might have simply done too good a job at curating my social media intake. After all, I just started a new LinkedIn account after years of not using it, and there’s been a jarring tonal whiplash between it and the social media I actually want to use.
On top of the general-purpose psychic damage I take every time I open LinkedIn, there’s now at least four or five new posts treating AI not just as an inevitability, but as something that’s already so firmly entrenched that if you’re not already completely committed, why are you even bothering to find employment? Contrast that with the pockets of the internet I tend to hang out in, where 99% of the people seem perfectly aware that generative AI is ludicrously overhyped and overvalued at best, an outright pyramid scheme at worst.
Whenever I see something like this article in Variety quoting Reese Witherspoon as saying that “AI is the future of filmmaking,” and that “you can be sad and lament it all you want, but the change is here,” it’s met instantly with derision and mockery.
(To be as fair as possible to Witherspoon, the extended quote tries to couch it as not being a replacement for creativity, and you can be charitable and take it to mean that she’s talking about using it as a time-saving tool for production and administrative tasks? Her examples of how she actually uses it aren’t “I used prompts to write a screenplay” but “I used AI to buy a blender,” but I can’t tell if that’s a bizarre non sequitur, or if it’s an earnest/overly optimistic take on how it can be used as a tool to support the creative process instead of replacing it. In any case, she’s financially investing in this stuff, so it’d be a mistake to take her quotes as being entirely in good faith).
The “you’ve got to get on board because otherwise you’ll be left behind!” refrain is extremely familiar, because we’ve seen it with every tech hype bubble: dot coms, web 3.0, cryptocurrency, NFTs. And as pointed out in the Bluesky thread I linked to up at the top, this is the part when the technology is spun not as a pyramid scheme for already-wealthy investors, but as democratizing technology benefitting women and marginalized people.
But again, I can’t recall ever seeing a pitch for generative AI that attempted to make a case that it was actually better. Complex? Sure. Hard for laypeople to understand how it works? Definitely. Saves time? Questionable, but let’s give it the benefit of the doubt. Easier? At first, absolutely.
I’ve seen it pitched as fun and novel: here’s your selfie as if you were a character in a Studio Ghibli movie! I’m enough of a narcissist to see the appeal, but even I wouldn’t rank the results as better than mall kiosk art, and I sure as hell wouldn’t spend any money for it.
I’ve seen it pitched as being so much better at producing convincing text and images than any other methods of procedural generation we’ve tried before. And yes, that’s basically true.1As long as you don’t acknowledge that every method of procedural generation that I’m familiar with, at least, is deliberately working towards a target, instead of “just keep trying it over and over until you get close enough to something you recognize.”
But never have I seen it convincingly pitched as being as good as the real thing.
To put it back into terms I can understand: it’s like if a Sith lord were pitching you on the dark side using quotes from Yoda. “Quicker, easier, more seductive, not more powerful.”
Even Apple, the company that has perfected the art of making a convincing and appealing case for spending a lot of something you don’t particularly need, hasn’t been able to make a strong case for the idea of generative AI.
I understand that FOMO is powerful, though, and can see why venture capitalists, along with the people riding an unpredictable mechanical bull made of billions of dollars in imaginary money, would be interest in speed-running to the “you don’t want to miss out!” part before making a convincing argument of why you don’t want to miss out. It does make it seem, however, that we’re a lot farther towards the bubble bursting than I’d previously thought. I’d expected at least another year before we started to see the flop sweat of desperate “get on board or you’ll be left behind!” messaging.
It does make all of the bubble seem suspect, though. I would’ve dismissed the people issuing public statements about how AI is going to basically become SKYNET and destroy all humans as coming from kooks. But now I suspect that they’re hucksters, pumping up the bubble by claiming that these systems can do far more than they actually can.
For the record, I’m still remaining optimistic that things will settle into narrow categories of very specific utility that’s actually useful. Almost entirely in processing data, not generating it.
The only part that I think is genuinely here to stay: everything’s going to get slower and more difficult, as we’re all flooded with loads of slop and are forced to differentiate between it and the stuff that’s actually worthwhile. It already takes more time than I’d like to determine whether a vacuous comment online was generated by a bot or by a dim-witted human.
So that’s why I’m giving you — and exclusively you, the readers of this blog — the chance to get in on the ground floor of my new venture: an unprecedented breakthrough system that detects whether something is worth your time. You’d better act fast, because you don’t want to be left out!
Leave a Reply