Nailed it. This is just getting infuriating:
https://chat.openai.com/share/03069d31-06d3-42e2-bfdc-ff6eaf...
GPT-4 went from being absolutely awe-inspiring to next to useless. This is the kind of response I expect from GPT-3.5-Turbo, not GPT-4.
Comparing the first offered code solution to the last offered solution, it's just insane how over-complicated it made things. It used to not be this way. The cream on top is that at the end it offered two different apologies and asked me to select which one I preferred more.
The lack of transparency behind these changes, and the gaslighting around the clearly obvious changes to the system's output are just ridiculously patronizing, and I will run for the hills the moment a company which actually values its users creates a competitive product.