I’m so excited for this but having made a few assistants / GPTs now I can say it’s not ready for prime time. They don’t succeed very often at producing predictable results and sticking to instructions.
Agreed, it's a little premature. I think it's a good first step to demonstrate a possible developer ecosystem though, and potentially a no code one at that. I'm actively working on a client project where we'll utilize more of the Assistant API (which should be more robust than GPTs alone), potentially with AutoGen from Microsoft Research. Interested to see if we constrain the use case ofcoordinating LLMs whether we get less hallucination. It's going to be a time for exploration for a while here.
https://www.reddit.com/r/singularity/s/8hobMFZqJY
I’m so excited for this but having made a few assistants / GPTs now I can say it’s not ready for prime time. They don’t succeed very often at producing predictable results and sticking to instructions.
Agreed, it's a little premature. I think it's a good first step to demonstrate a possible developer ecosystem though, and potentially a no code one at that. I'm actively working on a client project where we'll utilize more of the Assistant API (which should be more robust than GPTs alone), potentially with AutoGen from Microsoft Research. Interested to see if we constrain the use case ofcoordinating LLMs whether we get less hallucination. It's going to be a time for exploration for a while here.