Enhancing ChatGPT with Continuous Cross-Session Knowledge for a More Cohesive User Experience

Hey :cowboy_hat_face:

I’ve been exploring the incredible capabilities of ChatGPT and pondering on its potential evolution. One idea that has particularly resonated with me is the concept of enabling ChatGPT to carry forward knowledge and context from one conversation to another, across separate sessions, avoiding a classic search model. This would not only streamline interactions but also pave the way for a more personalized and coherent user experience.

The core of this idea is to enable memories between ChatGPT conversations. Remembering the context and specifics of previous interactions with a user and apply this knowledge to future conversations.

This capability would allow ChatGPT to provide more relevant responses and suggestions based on a user’s history of interactions, interests, and preferences discussed in past sessions.

I’m eager to hear the community’s thoughts on this idea. Has anyone else pondered similar enhancements, or are there ongoing discussions or projects within OpenAI exploring this concept? Your insights, feedback, and suggestions would be greatly appreciated as we consider the future of conversational AI and its potential to provide even more personalized and seamless experiences.

:cowboy_hat_face: :+1:

1 Like

Hey hey, looks like you posted this just on time. :grin:

OpenAI just announced their cross-chat-memory functionality in chatGPT. Sounds very similar to what you are describing:


That seems really cool, I hope it can be used with the API’s and not just the web interface based ChatGPT. I wonder if it will be more expensive? Surely it will be as the context in use must be bigger so the token count will increase. I liked Assistants API because it had a ‘memory’ but don’t like the cost of it.

Holy crap, they sure respond quickly to user feedback! :slight_smile: lol