Ever feel like your super-smart AI assistant just got hit on the head? You’re not alone. A wave of users are side-eyeing ChatGPT, convinced the latest version, GPT-5.1, has suddenly forgotten how to tie its own digital shoes, especially inside the Projects feature. One minute it’s your brilliant advisor, the next it’s arguing that two plus two is “a philosophical concept.”

Share on 𝕏 Share Reddit

So what’s the deal? The internet is buzzing with reports of a sudden “dumbening.” People using ChatGPT as a project advisor are finding it weirdly obstinate, missing obvious points, and requiring its own mistakes to be explained back to it. It’s like your normally reliable GPS suddenly insisting you drive through a lake. The frustration is so real that some are threatening to pack its virtual bags and send it to live with Gemini.

There’s something deeply funny about having to patiently explain logic to a machine that’s supposed to be the pinnacle of human knowledge. It’s the digital equivalent of your professor asking you how to turn on the projector. You start wondering if it’s just having an off day, or if it stayed up too late scrolling through weird parts of the internet.

The best part is the user’s polite refusal to share their private chat logs as proof. It’s the tech version of “my dog ate my homework,” but for AI. We’ve all been there—too deep in a bizarre, rambling conversation with ChatGPT to ever let another human see it. The commitment to secrecy, even while publicly accusing it of going dumb, is a mood.

In the end, it’s a classic love-hate relationship. We rant, we threaten to leave, but we still hope it’s just a temporary glitch. Because where else are we going to get advice on our projects and someone to argue with about whether a hot dog is a sandwich? The AI might be a little dumb right now, but it’s our dumb AI.