Writing is a way of capturing a moment in time, a specific pattern of neuronal firings, an idea frozen in the moment it is conceptualized. The invention of language is often credited with benefit of human-human communication, but early code designers like Richard Hamming were quick to realize that any code (or language) can be transmitted through space (communication) and through time (storage).

So it follows, that the biggest benefit of language might not be to communicate through space but through time, to allow the growth of humanity’s knowledge over generations.

To write something in language, I first must formulate it in my own mind, to twist, turn and inspect it until I can express precisely what I’m thinking. That is difficult because to explain to future me without degradation, requires that I explain it in the simplest possible terms that I will surely understand.

So in a sense, writing is applying The Feynman Technique to teach future me. And that is very useful, both for my current understanding and for my ability to layer on complexity by building with simple idealogical building blocks. Like thought lego!

With a repository of writing, I can pull on past threads and weave future ideas without starting from scratch each time.

Note One: The advances in Large Language Models, make this even more pertinent. Having a repository of written thoughts means that LLMs can interact, search and weave threads for me automatically! Probably an experiment to run at some point.

Note Two: I did this experiment in self-socratic-dialogues!

Note Three: It turns out Paul Graham thought about this too.