And, yes, that hype CAN cause a lot of damage, and is — I’m not underselling that. But the gap between what the hype says these models can do and what they can ACTUALLY do is continents apart, and people are already starting to notice that. They really are Beanie Babies — and worth about as much.

After weeks of testing: AI is dumb as hell.

It is your GPS steering you into a lake. It is your moon-landing-conspiracy aunt. It is Beanie Babies circa 1996. It is a student who did not read the book on presentation day.

It is so, so stupid, and please let that comfort you.

It will tell you what it thinks you want to hear instead of what’s in front of it. It will lie badly and gaslight you. It will happily take your hand and drive you off a cliff rather than reconsider context or new information. It is tragically, catastrophically DUMB.

It can’t synthesize information. It can’t read websites or images in documents. It follows Gestalt principles of closure and CONSTANTLY makes up information to fill gaps. It prioritizes vibes over accuracy. It will lie and then lie about lying. It can’t even ADD PARAGRAPH BREAKS.

I researched this so you don’t have to, so please believe me when I say AI is good for precisely fuck all, and I TRIED.

But that’s the thing: it doesn’t understand “essence,” or nuance. The one thing it’s supposed to be good at is basic processing tasks, responding to straightforward input, and my tests are showing that it just… isn’t. It’s really kind of fascinating, in a hopeless kind of way.

To be clear, I’m on our AI Taskforce at our college and have been tasked with doing some investigative work. Not only can AI not do half of what tech bros claim it can, it is *actively harmful* and will make up lies on the spot for continuity. It’s not just evil, it’s BAD at it.