Taking notes is the alternative to passively consuming information. It's how I work out the consequences of my ideas and bring together relevant information. It's how I remember what happened three months ago when I tried the same thing but with twice as much cornstarch. Sometimes I take my notes in the form of Wikipedia articles (e.g.,
https://en.wikipedia.org/wiki/Drafting_film), but most of them contain original research, so more usually I take notes in Emacs in Markdown (about 1.7 million words so far), or on paper with a pencil or pen. (Markdown is pretty shitty at sketching out process flow diagrams for air conditioners or drawing circuit schematics.) Occasionally I use graphviz, org-mode, Falstad's circuit.js, or Jupyter.
When I want to remember who did that paper on implicit fractal rendering with reduced affine arithmetic, I grep my notes (Gamito and Maddock, cf. https://dercuano.github.io/notes/affine-arithmetic-newton.ht...).
When I want to remember when Putin said a third world war was imminent due to US precision-guided munitions, I grep my notes (at Sochi in 02014, cf. https://dercuano.github.io/notes/wwiii-genesis.html).
When some dumbfuck on HN is saying "Yes China is making big growth in renewables but they're making even bigger growth in coal" and accusing anyone who disagrees of "brushing off scary conclusions" because they "can't get their heads around" the reality of the situation, I can just paste my analysis of growth in the Chinese energy market from my notes, which demonstrates that the opposite is the case: https://news.ycombinator.com/item?id=29843985
I learn more from fifteen minutes of taking notes on a book or a video than I do from an hour of passively reading it. Better still is taking notes as a way of contrasting two or three different sources of information, or using them to bring into focus inconsistencies or contradictions I wouldn't have noticed otherwise, or bring my attention to the information I lack in order to understand something.
I had read about configuring an opamp as a differential-to-single-ended amplifier (say, for high-side current sensing for a power supply current limit) half a dozen times over the years. But it wasn't until a few weeks ago when I tackled the problem with a mechanical pencil in my notebook (schematics, equations, paragraphs of flawed reasoning) that I really understood how to do it. After three tries. Maybe if I'd taken better notes when I was reading about it, it wouldn't have taken me so long.
Reading over my old notes, I see how incredibly wrong and clueless I was about so many things. Without notes, people tend to forget their past errors, which leads them to overestimate their judgment. I also see how I've changed, not always for the better.
Diagrams, equations, and data tables in my notes are much easier to show to someone than to tell them verbally in a conversation, and they're generally better drawn than what I can manage to sketch out on scratch paper or a blackboard during a conversation.
It would be great to have software that enabled better integration of my notes with databases and computer models, but the more important thing is retrievability. I want to reduce the chance I'll lose my notes, or that I won't be able to find things in them.