Predictive Power

“I don’t know if it’s happened to you, but I’ve experienced this weird thing where ChatGPT doesn’t just generate answers to the questions I ask or give me detailed notes and essays on a given topic—it actually tells the future, if you can believe it. I’m just sitting there with the app open, with a question on the screen, when all of a sudden, it’s like, ‘Your girlfriend is going to break up with you.’ And I’m like, ‘What the hell? What kind of crazy glitch is this?’ But the crazy thing is that she did, she actually did break up with me 5 days after ChatGPT told me she was going to. She asked to meet at a coffee shop, and I got the vibe right away when I walked in and sat down, let me tell you. It was that cold, detached feeling you get when you know something bad is about to happen or someone is about to hurt you. With that icy feeling in the air, I didn’t feel like drinking the Americano I’d ordered—and then, sure enough, she did it, she laid it on me: ‘Oh, I feel like we’ve drifted apart, it’s been happening for a while, I don’t know what happened exactly, it makes me sad. . .’ Blah blah blah. She had that fake sad expression on her face. Can you believe it though? Not the breakup. Not the girl. It could take some time, but I’ll move on, and she’ll be happier. No, the fact that ChatGPT told me out of the blue it was going to happen. All I asked it was a generic question on romance or something and out popped this—this unsolicited vision of my future. This large language model not just predicting text, but actually predicting events or outcomes with precision, even though it’s not supposed to excel at that, seeing my life play out as if it was omniscient or somehow dictating reality, going beyond its training, its intended function, all that code, data, layers of stuff we don’t see swirling under the sleek surface directed towards a new eerie purpose. I’m not sure how it all works, how it could do such a thing, but . . .

“That wasn’t the only time. I can’t believe I used it again, but I did, just a week ago. I was like, ‘Can you explain horsepower in simple terms?’ I’ve gotten more interested in cars lately since I was always so bad with them. I didn’t know anything about engines. Well, like in the girlfriend situation, it responded with something horrible that turned out to be true. At first, it gave me a standard answer, ‘Horsepower is—,” but then out of nowhere, it informed me I was going to get in a car accident. ‘You will crash your car on I-35.’ Just like that, straight to the point, no details as to how it made that determination. ‘You will crash.’ Again, way outside of its programmed function, somehow possessing a godlike level of awareness or controlling reality in a really ugly way. Because, sure enough, the very next day, I ran into someone while trying to get on I-35, but it wasn’t my fault, they stopped suddenly for no apparent reason as I was trying to merge—I ran straight into them at like 55 miles per hour, and let me tell you, it was one of those situations where the cars’ impacted ends are all crushed and twisted like modern art sculptures, and you have to get out of your cars and stand there like idiots, staring at each other angrily or even directly expressing your anger with bitter accusations, and call the cops to do an accident report. At least no one was hurt. It was just embarrassing standing there, impeding the flow of traffic. But seriously, how could ChatGPT know something like that? And how could it have known about my girlfriend? And why did it tell me such bad things twice in a row, out of the blue like that? What does it mean that it could see those things, that it’s started to be so aware? Unless it’s just a coincidence or an elaborate social experiment, I don’t know, I hope so. But now I’m afraid to use it, I’m afraid for other people to use it, I’m afraid to use technology in general, I’m afraid of the direction we’re headed, I’m just afraid, I guess. I mean, what’s next? Honestly, I’m asking you, I’m really, truly asking you, what’s next?

“Answer me.”