This made me think of a much more interesting project. A compendium of information automatically extracted from research articles.
Essentially one totalizing meta analysis.
E.g. If it reads an article about the relationship between height and various life outcomes in Indonesian men, then first, it would store the average height of Indonesian men, the relationship between the average height of Indonesian men and each life outcome in Indonesian men, the type of relationship (e.g. Pearson's correlation), the relationship values (r value), etc. It would store the entity, the relationship, the relationship values, and the doi source.
Hyperassociativity is not necessarily apophenia. Apophenia is finding meaningful patterns in random noise, but it is not clear that the patterns being discovered were in random noise, or if the person merely saw the present patterns more clearly. That can happen too, and is not apophenia. It's something more like hypercognition
Bing chat is programmed to never be rude, but also to respond in the same tone that it was spoken to. And so, if it tries to mirror a negative tone, it will instead respond with that goodbye message, to avoid being rude.
If the book is actually good, then what is interesting about it is that it would still be about something that humans find important and relevant, due to the LLM being trained on human cultural data.
1. Start by writing down everything you know about what you want to write about in a disorganized way.
2. Read the disorganized text and figure out which information goes together, then put that information together.
3. Elaborate on the points you've now brought together.
4. Turn each elaboration of points into a proper paragraph.
5. Reread the whole document and figure out if the paragraphs make sense together, and if not, edit them to integrate them.
6. Edit out everything in your text that is besides the point you are trying to make.
Let me make an example:
1.
Trees are tall. Trees are living and have lived for millions of years. The bark of a tree is hard. Trees are immobile.
2.
Trees are tall. The bark of a tree is hard. Trees are immobile.
Trees are living and have lived for millions of years. Trees bring up nutrients from their roots and absorb sunlight in their leaves. The branching patterns of roots and branches are similar. There is a reason for this.
3 - 4.
Trees are tall living creatures that have existed for millions of years. They can be so tall because they are very hard due to the strong structures that cellulose holds. They can't be too hard, and some flexibility is necessary so the tree doesn't snap, but they are still much harder than they are flexible. This hardness makes them immobile, which requires the static form they take to be reliable and robust enough to keep them alive.
The best static form a tree can take is the one they evolved to take. The fractal branching pattern is the pattern that exists in the equilibrium of three requirements: it needs the strongest possible structure, allows for the greatest surface area of leaf coverage, and takes the least amount of volume up. These requirements lead to the formation of the fractal structure of tree branches. The fractal structure of roots is governed by two main forces: the need to fill as much volume of earth as possible with the least amount of root matter. This leads to branches being slightly different from roots, and according to these requirements, one would predict that roots have a much shorter distance-to-divergence of their branches than tree branches.
5 - 6.
How are trees so tall? Why do they branch as they do? Over millions of years, trees have evolved into the form they take. Several forces guided this evolution, and the properties of trees reflect them. What lets trees get so tall is that they are very hard and flexible due to the strong bonds that cellulose holds. This hardness makes them immobile, which requires the static form they take to be reliable and robust enough to keep them alive. What determines the properties of this static form?
The best static form a tree can take is the one they evolved to take. The fractal branching pattern is the optimal pattern that falls in the equilibrium of three requirements or forces: it needs the strongest possible structure, allows for the greatest surface area of leaf coverage, and takes the least amount of volume up. These requirements lead to the formation of the fractal structure of tree branches. The fractal structure of roots is governed by two main forces: the need to fill as much volume of earth as possible with the least amount of root matter. This leads to branches being slightly different from roots, and according to these requirements, one would predict that roots have a much shorter distance-to-divergence of their branches than tree branches.
------
As you can see, it's a simple process that allows for rapid expansion of ideas, starting from me simply dumping information about trees, to making a point about why trees take the structure they do.
Not at all. The caution about attributing human concepts to other animals is that they can have an arbitrarily different brain organization in such a way that stretches those notions beyond what we usually take them to mean. Does a tiger experience joy? If it does, what does joy feel like to a tiger? Is bird song language? What are they saying? That last question is obviously begging the question, are they saying anything? Is "saying" something that more than humans can do, or when we refer to the word "say" do we only apply it to humans expressing human concepts? We don't even know if the concepts we use to describe minds are even valid, or if they are just products of introspection. This, along with the fact that other abstract aspects of the human condition are also on shaky ground when you change the underlying substrate those aspects emerge from, means that it's not clear that animals have those same exact aspects in their species-specific condition, and if they do have some similar aspects, those aspects can be quite different than what we'd find in humans, with completely different sets of states. TL;DR We hardly know if you and I see the same red, let alone if a tiger and a human feel the same joy.
If it makes them better than other people at reading things up close, such a person could thrive in an environment that demands that. Of course, they should also be able to wear glasses when the environment demands it.
Cameras don't see colors, a wall or a texture on that wall. Cameras merely capture a grid of values. All color perception, object recognition and figure segmentation happens in the brain.
There is an even greater opportunity cost while you're not trying anything.
Even if you don't end up liking the thing you're trying, at least your life has more breadth and depth than it did before.
Essentially one totalizing meta analysis.
E.g. If it reads an article about the relationship between height and various life outcomes in Indonesian men, then first, it would store the average height of Indonesian men, the relationship between the average height of Indonesian men and each life outcome in Indonesian men, the type of relationship (e.g. Pearson's correlation), the relationship values (r value), etc. It would store the entity, the relationship, the relationship values, and the doi source.
Something like a quantitative Wikipedia.