Understanding and simulating the interaction with a (physical) book an interesting HCI topic. Though we have different designs of document readers, it's still difficult to mimic some interactions such as turning pages quickly and glancing at each page quickly. Our behavior is limited on a single screen that is not super powerful, and we have developed very specific technologies for this scenario. I'm wondering if we have devices supporting more natural interactions, what the hypermedia systems will be like.

It's easy to store and access data now, but this doesn't mean humans can make good use of such systems. With modern internet, we can potentially find a wide range of information but the hard thing is how to find them, and how to use them after finding. The processing power of human brain is limited and information explosion becomes an issue. AI technology might be a great helper for dealing with large amount of information. I think this is part of the reason why there's a lot of investments in this area.

I'm wondering how Bush will think if he knows the technology today. Yes, we achieved it with digital cameras and we can store the images in SSDs. It's not highly related to the technology of dark room photography and films. This is an example of the issue Bush is worried about: specialization makes it harder to bridge between different disciplines. We even don't know if there's a better solution that is simple but no one has discovered it.

Nowadays it's even more difficult for great ideas to reach "the few who were capable of grasping and extending it", not because they don't have access to the piece of information, but because less and less people is likely to have the knowledge to understand and extend it. Human knowledge is developed like a tree, with people are divided into more and more detailed specializations. However, these specializations could be very sparse in the full set of knowledge and there could be areas that no one has explored since the corresponding specialization has not been invented yet. There might be some special topics in math that is useful for AI but no one in the AI community has even heard of it. Researchers today can read papers from all disciplines, but their minds are not ready for the big integration. Improving "the power of mind" becomes an even more urgent issue now than it was in the 1940s.

In the era of AI and large models, this is still a big question. People have some understanding and measurements of physical powers, so we somehow know how they can be improved. However, there's still no clear answer what "powers of mind" are. Modern generative AI models can create contents that are similar to human work, but we still don't know whether it's able to innovate. Actually, we can't even measure what is innovation and what is an improvement of the power of mind. I'm curious if this is a topic totally within the scope of science or we should turn to occult to some degree.