Considerably lower quality outputs for complex prompts since Dev Day

Has anybody else noticed a considerable decrease in quality since Dev Day?

It seems like GPT is not providing as accurate/high quality responses as it was pre-dev day.

Responses seem more ‘robot’ like where im having to be extremely extremely explicit and still not getting as great an output as it was providing pre-dev day.

Seems like alot of the context it used to draw upon to answer my prompts has been reduced some how?


Same here. For example, when I asked it to give me a list of domains (like HR, Finance etc.) it added in a few verticals (ex: real estate) into that list.