One of the things I love about the API-ification of these LLMs is that they’re plug and play.
I built https://persona.ink against davinci knowing it didn’t give as good of results as ChatGPT but knowing I could swap the model out once 3.5 came out. Today is that day, going to swap out the prompt in the cloudflare worker and it should Just Work(tm)
Is davinci actually worse than chatGPT? I know it's worse as an assistant, but in my (admittedly brief) testing the performance was the same or better for tasks like summarization, sentiment analysis, etc.
I guess it's irrelevant now because everyone will use the one which is 10x cheaper.
Yes. Asking ChatGPT to assume a persona and rewrite content, it performs significantly better than davinci in my tests. I was able to get "good enough" with a lot of prompt engineering on davinci - but ChatGPT is a shoulder above with less investment in prompt engineering.
On the flip side - I get hand swatted by ChatGPT more frequently than davinci. Davinci's moderation filters don't really pick up on much, but ChatGPT will give me a lecture instead of a translation on a lot of occasions. There are many valid use cases for rewriting/editing content that involve graphic details that davinci will gladly handle and ChatGPT will give you a lecture about.
Human existence is messy. ChatGPT doesn't like the messy.
I built https://persona.ink against davinci knowing it didn’t give as good of results as ChatGPT but knowing I could swap the model out once 3.5 came out. Today is that day, going to swap out the prompt in the cloudflare worker and it should Just Work(tm)