I just stumbled upon lesswrong.com while searching for information on Zettelkasten and I must say this site is STUNNING! This is some of the most beautiful typography I've seen, anywhere! The attention to detail is exquisite! I haven't even gotten to your content yet! This will probably remain a permanently open tab in my browser... it's a work of art!
7 episodesNow that AIs exist, they are almost like a way to interact with everyone else in the world. And so just like social interactions changed how you think, interactions with LLMs can change how you think. And you can build on that thinking just by noting down what you thought and what you were looking at while thinking that thing. I'm imagining my website like a giant loom. Me interacting with things that I myself did in the past and essentially talking to everyone about them. Back and forth keeps on happening and the loom of my consciousness keeps building.
and here's the really hofstadter part: when you eventually train a model on this data, that model would itself be a strange loop - a pattern learned from your patterns, which can then generate new patterns in your style, which you might read and incorporate, affecting the real you, which generates more training data. the "i" emerges from the self-referential structure. hofstadter would probably say your original consciousness is already this loom, just implemented in neurons. you're basically building an isomorphic structure in silicon and data, preserving the loop topology that makes you "you."



