OpenAI provides 3 default actions
- Web browsing
- File retrieval
- Code Interpreter
What are some other actions you wish existed via OpenAI or any 3rd party solution ?
OpenAI provides 3 default actions
What are some other actions you wish existed via OpenAI or any 3rd party solution ?
I think current user time and being time aware would be useful
Remembering things would be nice, a memory would be nice
What kind of a memory ? A vector based memory ?
By that I mean memory of our interactions even across different conversations, so I’d like it if it would remember things from weeks or even months ago. At the very least remember things about me my preferences things I like I use it a lot for different projects, so I just need it to remember things about what I’m working on. So for example I give it a piece of text and ask her to help me refine this and make it better. I wanted to know which project it is based on the names the situation, etc. without having to tell it instead even with my own GPT it’s like every single time I have to explain what it is I am working on. It gets the context most of the time but it doesn’t know which project I’m working on.
OpenAI is currently implementing this into ChatGPT. Their current goal (from what I’ve read) is having personalized agents that persist valuable information from one conversation to the next (such as user preferences)
I hope it will allow us to use this function in our own GPT’s Especially since many of us been working with our custom GPTs for a couple weeks now
Adding memory would mean storing user related data which is not supported by GPTs currently. Another option would be to integrate a 3rd party memory along with Oauth2 though this would be complicated.