Back in March, ChatGPT worked great and I was able to use it effectively for daily coding tasks.
Then we had the May 3 update, which seriously degraded the quality and performance and I struggled to use ChatGPT effectively because it would forget what we were working on all the time.
July 19/20, solved all the issues for me and was a great improvement. Conversations were on track all the time and ChatGPT wouldn’t forget context and previous responses anymore.
But the fun didn’t last… AUG 3 update ruined the model completely. I’m getting irrelevant responses that have nothing to do with the current chat or what we’re working on. E.g. the chat is about SPFX extensions for Sharepoint online and out of nowhere it starts mumbling about Wordpress and PHP functions or replacing my Typescript with Jquery. Or it suggests a few lines of code for the script we’ve been working on the entire chat and when I ask it to add the code from the last response to the script it responds with something like:
Certainly! Based on the code snippet you provided, it seems that you are working on a SharePoint client-side application that observes buttons with a specific class and hides options in a contextual menu except for “Download” However, you didn’t specify the exact changes you want to make, so please provide more information about what you need to be modified or updated, and I’ll be happy to assist you!
I'm currently very disappointed with ChatGPT and really do not understand why they rush these updates out without any community testing or feedback. We don't have a choice and are forced to use whatever OpenAI decides on.
Long story short: PLEASE ALLOW USERS TO CHOOSE THE RELEASE THAT WORKS BEST FOR THEM.