I’ve been contemplating the potential of a Facilitator GPT that could act as an intermediary between multiple custom GPTs. This facilitator could handle and synthesize information across various models, ensuring a cohesive flow of data and responses.
The main benefits could include:
• Improved consistency and relevance across responses when dealing with complex projects involving multiple models.
• Enhanced ability to manage and integrate diverse data streams and user interactions.
• Potential for implementing advanced chat organization and privacy controls.
I believe such a system could significantly enhance how we utilize custom GPTs for complex applications. What are your thoughts on the feasibility and potential implementation challenges of such a feature?
========
There’s been some mentions but no formal discussion on the introduction of facilitator GPT’s.
See here:
Assistants API - Access to multiple assistants - #10 by ronabarrett
Overall I’m thinking this topic might ultimately lead to a more direct conversion around implementing a general permissions model for each GPT (or even session) and perhaps a way to elevate that permission upon request/prompt from within the session. Not sure what (if anything) is in the works for something like this.