I've found that ChatGPT, even with a lot of instruction, can't figure out how to write good SQL queries.
2 out of 5 runs will result in calls to objects that don't exist.
Aside from that, building the Action to use the queries directly against the database has proven difficult. You don't get any errors back so it's nearly impossible to troubleshoot why it fails an API call.
Finally, it doesn't remember new information, so it has to learn all the tables and columns again with each new chat, which means it makes the same errors repeatedly, which means the first question on any conversation takes around 15 to 20 seconds to get anything back at all.
Overall, it pretty much sucks.
Instead, ChatGPT kept giving me solutions no matter what prompt I gave it. It seems to get "stuck" after giving a good run initially.
I'm totally not sold on this AI hype anymore, I can begin to see cracks.
As someone who doesn't use GPT-4 more than daily, I usually don't use more than $5 a month. You would need to input 500,000 tokens and output 500,000 tokens in order to rack up a $20 bill.
I can barely get anything done as is and with ~ChatGPT4 alternatives cost 100x less, OpenAI appears to be in deep trouble.