From plugins and custom gpts to a more robust chatgpt

I am reflecting, and prognosticating a little, about chatgpt -4and the introduction of the GPT store, as well as the prior plugin capability.

It would seem that the introduction of the store is meant to gather data so that a future GPT incarnation (which is trained on this data) would be able to automagically choose the most appropriate custom gpt(s) without us having to manually wade through thousands of options.

If this is a known plan by OpenAI, please let us know. I am hoping that this is the plan. I love the custom GPTs, but I see this as a natural segue to automating this.

So that, for instance, questions about math go to wolfram alpha without me having to manually coordinate this action. Or questions involving creation of a diagram go to two or three of the best diagramming options (“best” being determined through data collection of our custom gpt usage from the need to create a diagram).