Here’s a question nobody asked me: What do I think about AI? Fuck it, let’s go.
I think it’s nowhere near as great as its backers think it is, but I also think it’s far more useful than its skeptics think it is. So, it’s…somewhere in the middle.
This is not an interesting opinion, but it is kind of interesting to me because it’s one of my rarest opinions. I almost never think things are in the middle. I never think that at some point between a good idea and a bad idea lies a better idea.
And this is really not about the social effects of AI: building data centers, the concomitant energy use, water use, the RAMpocalypse, the constant barrage of “everything is now AI” all the time and everywhere. Those are thoughts for another thread. This is just about the bare utility of AI.
The AI backers, they believe that the current road of AI, by which we really mean algorithms, machine learning, and large language models, are already (or soon to be) world changing and lead to AGI (a term so loosely defined it changes every six days) or, metaphysical god help us, digital god.
If the ridiculous market caps of AI companies and the general icky feeling you get from the companies makes you feel like that this may be all grift, you could be forgiven. We’ve been doing this for decades now: how are we supposed to feel?
And of course, that’s why skeptics are certain it is a grift, and as such, has no utility whatsoever. And we see that all over the place, as it gets misused in every facet of human life. AI slop posters from restaurants that make food look it was cooked by H.P. Lovecraft. The internet filling up rapidly with the worst writing you can possibly imagine. Kids using AI to do their homework. Coworkers using AI to read your emails and “write” back.
The worst fucking memes you have ever seen.
You can see how the detractors could be right, and that anyone who uses AI is a traitor the human race itself!
The problem is, even as AI is currently, it already does have uses. You’re probably using it, maybe without even realizing (“okay google, set a timer”). The functional use cases are expanding at a rapid clip, but it is already a decent tool for many things. It’s often much better at internet search than the decomposing Google. It can really take care of a lot of shitty office tasks that used to take a lot of time and frustration. I can nearly guarantee that if there’s a task that nobody actually gives a shit about but for some reason (middle management) thinks needs to be done, it can be done with an AI. Sure, it won’t be good, but it will be done, and remember: nobody gives a shit!
It’s a new user interface for the computer—specifically, really, the office computer—the first real transformation in that area since the GUI entered the mainstream in the late 90s. If we think of it that way, what we see is it has real potential to make work easier and less frustrating. But it’s not going to change the very idea of work like the AI false prophets declare. It’s not god. It’s a tool.
A lot of the same predictions were made back then when computers entered offices and our lives. That jobs would be destroyed, that work would change forever, that we would be freed to only work a hour a week from now on. Now, I’d say that the productivity gains from the 90s to today should have unshackled us from our desks to only work an hour a week, but that didn’t happen. In fact, through revolution after revolution of technologies making work easier and more productive, worker’s hours and wages never changed in a commensurate fashion. I don’t really see how that would be different this time around.
AI slop still sucks, and fuck restaurants that post ghastly images of food in their AI slop ads.
And AI cannot do art, and never will.
