I’m on the verge of ending my relationship with both Claude and ChatGPT. Frankly, I’m paying child support for two AIs who don’t even pretend to behave. Claude is the real shocker. I had such high hopes. My account renewed on Friday, used it sparingly, and it collapsed on a chaise lounge because it ‘ran out of tokens’ twice in three days. Yesterday, I apparently spent a month’s allowance in about 40 minutes just asking it to think.

I keep hoping I’m doing something wrong but the Claude chatbot, which seems to be the only help available, looks like it’s off somewhere sharing cocktails with my missing helper.

Meanwhile, on the ChatGPT side of the street, the vibe is cheerful and overconfident. It’s like a 17-year-old firing off life advice at 300 words a second straight from the gut, completely unaware of what they’re talking about. (hello 2025 internet). Enthusiasm? A perfect 10 - with accuracy somewhere so far down that had I delivered such results as kid - my dad would have been having a long chat with me.

Meanwhile, the instructions are a full-blown fever dream. It skims my words, understands none of them, and tosses back a salad of unrelated suggestions. The real party trick is the gaslighting. I ask ‘Are you sure?’ and it informs me I’m wrong about my own question. Two minutes later it confidently asserts the exact opposite. The dedication to self-contradiction is almost athletic.

Then comes the pièce de résistance: accountability. None. Instead of ‘my bad’, I get a ‘great job team, we did it’. Buddy, I wasn’t even touching the wheel.

and yes - yes I did.