For some reason, when prompting with the folder as context, gpt-4 starts processing it but stops before ever completing it and deletes its unfinished response.
The worst part is that even though the response never came through, it counted as a request. 
Edit: I believe the issue here is with the ‘Computing search queries.’ I really like this feature but it is not working as expected. It was working just fine earlier this week…
Edit2: This search method is consistently failing on my second prompt. It deletes itself after finishing the grey text (not sure how else to describe it, but it seems like this section is generating further detail into the prompt based on the user’s original prompt) Perhaps the pipeline is getting broken in the process?
I have seen a few other users having the same issue earlier this week (4 - 5 days ago), it seems like it is still an issue though