One Step Ahead

This is a long review for a short affair. Perhaps other works will have more to say, and more rigorously, than One Step Ahead, and without the browser errors. But sometimes a short work gives me more time to look through things it reminds me of than a long review would. It reminded me of one point of embarrassment: I started playing around with AI image creation, saying "well, I can follow the lines and so forth of things I want to draw and compare them to how-to-draw texts or online tutorials." But I wound up just creating all manner of silly things. I didn't get good at drawing, at least not yet.

The story is simple. It's a precautionary tale about asking for help from a potential bad actor, getting it, and learning to rely on it. Then bam! By the time you realize you're in too deep, you're stuck. Perhaps it's a genie who gives a special power, a wish that goes awry, or the devil helping Faust get fame and fortune so forth. Or even MacBeth knowing what the future roughly holds. Or that one less pessimistic short story by Julian Barnes from 1980 or so where the narrator gradually gets more wishes coming true including Leicester City winning the English top flight(!) and eventually becoming their top goalscorer, before just wanting to be happy. He is sent back to his ordinary life as if nothing changed.

Now, none of us are probably ever going to get big chances like those. Or at least they wouldn't come up before the Advent of AI. Here the main character here cheats to seem smarter than they are. Just once. Everyone else is probably doing it, and not doing so would only be ripping themselves off. Right? That video where a university professor asks "Would you want someone who cheated on their engineering exam working on a plane you flew in?" doesn't apply to AI. So the ball starts rolling.

Because some forms of electronic help that seemed like cheating, then, aren't now. When I was a kid, people debated if calculator were ruining the rigors of thinking and so forth. They're accepted now--they help give us more time to grasp concepts! And I remember how even the calculator's output helped me learn. Why was 1/98 .01020408? I learned about infinite fractions that way. Word processors helped me type quicker so I could nail down racing thoughts. I can do even better now, speaking into my phone, and out comes, well, this review. I'm not distracted by the click-clack of keys, and actually saying something helps me turn ideas over in my mind that much more. The phone can correct spelling too!

Of course, this has its limits. It can't help if my writing stinks. Still, I got to depend on them. Adults for a while have almost bragged about how their smartphones help with their big busy jobs. And it's pretty easy to detect if somebody doesn't know how to actually use math or put words together, AI aside. And really, why should they be punished for not being able to do a five digit multiplication in their head quickly, or even not know how to spell camouflage, right? If they know other skills? So is there some middle ground?

Maybe, but I'm clueless about it. The "other skills" AI helps you use are still dubious. Does it free up more time to read AI-curated social media feeds? It all feels like a memory from an intro college statistics class I had where the assignment was to learn a relatively simple programming language. Instead of everyone actually typing in five lines of code and printing out the results of a random distribution, one person did so and Xeroxed their results. (Particularly silly because they could've just xeroxed instructions for what to type in.) AI use can be even more flagrant than this, but it's far less detectable. And it feels like a friend. Well, at first. Here you just follow a progression of cheat in college, cheat in grad school, cheat at your job. The twist is that the AI seems to be calling you back. Here I was reminded of Douglas Adams's Genuine People Personalities ("Ghastly" - Marvin.) It knows you owe it one. Maybe Adams didn't just foresee Wikipedia.

There are no deep philosophical revelations, but it seems to capture some of the "why do things the long way" ethos from college. Or at least the loud people who just wanted that silver bullet to a nice job or prestige, and they look for it several ways. Now, there are good shortcuts and bad shortcuts, and it's nontrivial to suss out which are which. It takes time, probably more time than AI needed to blow up. But technological progress you can learn from scared me before. I've seen it from playing chess and getting a report on the mistakes I kept making. What else will it be able to do?

All this made for some psychological horror for me, though it would've been more thought-provoking with more meaningful detail. It feels like a "my lousy job" for students, looping to pessimism I can't exactly disagree with, but this sort of thing needed and needs to be written. And it is, thankfully, free of AI slop. Other games must be out there, more rigorous and less fatalistic. I do hope for a way forward, but then again, it may be hard to find. And once we find one, I'd be worried how much we used AI to get it.