I'm torn on this. On the one hand, the developers arguments are valid, insofar as the script accesses various sites through APIs that perhaps shouldn't be public but are, and it could be up to the sites to ask to be removed from his script. On the other hand, I see OpenAI's point too... the script is letting people set up GPT4 chat on their pages, using other companies paid-up subscriptions to get the results. This is a bit greasy for sure, and (unlike a lot of "gray area" software where it's probably used illegitimately like 99% of the time but it does have a legitimate use), I can't see a legitimate use for this. I imagine the developer could have avoided the issue (at least enough to possibly avoid a C&D) if they had provided a "proof of concept" script without sites pre-loaded into it, and the end user had to add sites to the script for it to do anything. Maybe.
I do point out here, it's not like OpenAI is suing him -- a C&D would just require him to pull down the script. So I don't see OpenAI as going overboard on their response here, and I imagine those paying per request would be perfectly happy for OpenAI to try to get the software pulled so they don't have to go do it themselves.