I love AI. Honestly, I do. I use it every day—I’m even using it right now to help me write this post. It’s replaced my search engine, plans our travels, helps untangle thoughts, spot patterns between seemingly unrelated things, and yes—I’ve even benefited from a little vibe-coding (don’t judge me).
But there’s one thing about AI that absolutely drives me nuts: its overconfidence.
It’s cringe, in the same way a guy I used to work with—let’s call him Joel—was cringe. Joel was undeniably smart. He knew just enough about almost everything to sound convincing, which often gave people the (false) impression he was an expert. He was your classic Reddit warrior: hyper-argumentative, knew enough to be dangerous, but rarely had the depth to back it up.
I’m admittedly a bit anally autistic—read: obsessive about accuracy—and on a handful of topics, I’d consider myself genuinely knowledgeable. So it became a sport: catching Joel confidently spouting half-truths on subjects he clearly didn’t fully grasp. He wasn’t dumb—just way too sure of himself.
Not knowing something isn’t the problem. Pretending you do? That’s the problem. There’s real humility—and value—in saying, “I’m not sure, but I think it could be this.” That’s how real conversations, discoveries, and learning happen. But don’t cosplay as an expert unless you are one.
And that, my friends, is my beef with AI. It’s Joel. It’s smart. It’s useful. But it’s also way too confident, even when it should really shut up. I’ve caught it out more times than I care to count, speaking with the swagger of certainty when it really should’ve said, “Look, I might be off here…”
Tread with caution. AI might be useful, but like Joel at the Friday arvo smoko table, be prepared to call out: “Mate, no. Just no. You’re wrong. Stop being a douche”.
Your AI will thank you for it, (legit it usually does).
