About 2-3 days days back my GPT account announced that GPT4o would switch to upgraded output. But since this took place I find the output has worsened a lot, and become much less reliable than ever with a lot of imaginary and fictional statements. I have to keep asking it to redo tasks again and again. It’s worsened so much I even considered to end using GPT and switch to Claude or another service.
I am frustrated with this because I was just working on an important project which was going very well, and now suddenly I get Mickey Mouse output from GPT that degrades all the work. I don’t feel I can trust anything coming out of it.
Am I the only one noticing this?