Have had this idea to use LLMs to allow you to “zoom in and out” of a book conceptually, the same way you can zoom in and out on something like Google maps, and go between high-level, and detailed views of the same information.
Great idea. Something similar has already been created circa 2014. I found your post because I was trying to remember and googling. It's here, Quanta. I'm also trying to create similar things.
This looks very cool. Lots of reading materiel here. It is always amazing how much useful technology gets lost because the moment isn't quite right. Anywhere I can find out more about your stuff?
Thanks! Just below Rama's chapters his 5 min video gives a really tight overview.
I've not published anything I'm doing on AI yet. But I did create a system/database in LLMs, and then visualized it and gave a presentation in Prezi, which is one type of an ideal zoomable format for what you're pursuing as well: https://youtu.be/tfUuLrw6_fA
And in fact, just yesterday I very quickly created in chat-GPT a Prezi style infinite zoom whiteboard, so I can start creating my own experiments in there as well. Anything I prototype in LLMs I port over VS Code (w/ co-pilot) as soon as possible where I can buld it out indefinitely. Many projects on the go.
Very cool. I like this interface. I feel like infinite zoom is a great way to represent abstractions in a way that feels natural. Feel free to send over any future explorations you have on that!
That’s a really good idea !
Great idea. Something similar has already been created circa 2014. I found your post because I was trying to remember and googling. It's here, Quanta. I'm also trying to create similar things.
https://ramakarl.com/quanta2/
This looks very cool. Lots of reading materiel here. It is always amazing how much useful technology gets lost because the moment isn't quite right. Anywhere I can find out more about your stuff?
Thanks! Just below Rama's chapters his 5 min video gives a really tight overview.
I've not published anything I'm doing on AI yet. But I did create a system/database in LLMs, and then visualized it and gave a presentation in Prezi, which is one type of an ideal zoomable format for what you're pursuing as well: https://youtu.be/tfUuLrw6_fA
And in fact, just yesterday I very quickly created in chat-GPT a Prezi style infinite zoom whiteboard, so I can start creating my own experiments in there as well. Anything I prototype in LLMs I port over VS Code (w/ co-pilot) as soon as possible where I can buld it out indefinitely. Many projects on the go.
Very cool. I like this interface. I feel like infinite zoom is a great way to represent abstractions in a way that feels natural. Feel free to send over any future explorations you have on that!