I've kept notes badly for about twenty years and slightly better for the last three, and the improvement came from one change. I stopped saving things and started saving what I thought about them.
The old system, if it deserves the word, was collection. An article I liked went into a folder. A quotation went into a document. A book got underlined. The collection grew steadily and I almost never returned to it, and on the rare occasions I did I'd find a passage highlighted in a book I no longer remembered reading, with no record of why I'd marked it. The highlight preserved the sentence and lost the reason, which was the only part that was mine.
What I do now takes longer per item and produces far fewer items. When something strikes me I write a line about what it made me think, in my own words, and I put the source underneath it. Two sentences, usually. If I can't write the two sentences, the thing didn't actually strike me, it just interested me for a minute, and I let it go.
That filter removes most of what I used to save. I'd guess nine in ten. What survives is a much smaller set of notes that are actually about me rather than about the world, and those are the ones I use. Nearly everything I've written in the last two years started as one of those lines rather than as a topic I sat down to cover.
The old practice for this is the commonplace book, and people kept them for centuries without a productivity vocabulary attached. What I like about the historical version is that it was explicitly a book of things worth having again, not a record of everything encountered. The selection was the work. Modern capture tools have made saving free, which sounds like an improvement and removes the only mechanism that made the archive useful.
I don't have a sophisticated setup and I've stopped looking for one. A single running document, dated entries, newest at the top. No tags, no categories, no links between notes. I tried the elaborate versions and found I was spending the attention on the structure, which felt like thinking and wasn't. Searching a long document works fine, mostly because there's less in it now.
Two habits do most of the work. The first is writing the note the same day, because the thought decays faster than the reference does. I can find the article again in a month. I can't reconstruct why it mattered. The second is rereading the last month's entries at the start of every month, which takes about fifteen minutes and is where the connections happen. Two notes written three weeks apart turn out to be the same idea, and neither of them looked like an essay alone.
There's a version of this for teaching too. I keep a short file per course of things that went wrong or landed unexpectedly, written the same afternoon. Not a formal reflection, just a few lines. When I rebuild the course in the summer that file is worth more than the student evaluations, because it records the moment rather than the aggregate impression of the term.
The thing I'd say to anyone starting is that the system matters less than the requirement to say something. A note that only records what someone else wrote is a bookmark. It will sit there. A note that records what you made of it is already the beginning of your own work, and it'll still be legible to you in a year, which is more than I can say for anything I highlighted in 2011.