| I described it a couple of weeks ago here: https://two-wrongs.com/reading-slightly-more-incrementally.h... But I can try to summarise/expand. A large part of it is learning to write good flashcard prompts. That was like 80 % of the hurdle for me. It does not come naturally for me, so I needed to look some tips up and then I notice I'm getting better with practise. I started by reading https://andymatuschak.org/prompts/ and Wozniak's https://www.supermemo.com/en/blog/twenty-rules-of-formulatin.... At least for me, another 15 % came from making it convenient to create flashcards. Actually designing the flashcard prompt takes a bit of mental effort (it's hard to write good prompts!) so if it then is technically challenging for me to create the flashcard from it, it's less likely to happen. I have a script that takes in essentially Markdown sections with bullet lists where each list item becomes its own flashcard with the title given by the Markdown heading. That works for fast entry for me. I write the Markdown on my phone, and then every now and then I use the script to batch import all I've written so far into my spaced repetition software. The last 5 % for me was switching over from writing my plain old notes into writing flashcards instead. That bit was fairly easy because I was already in the habit of making notes. I stop and write notes on: - Things that connect to something else I know, even if the authors don't point that connection out. - Things that seem insightful, even if I don't yet know what the insight is. - Things that surprise or delight me, or make me sad, or whatever. Any emotional reaction. - Fundamental facts that other results build on, unless I'm already very confident that I know that fact even if someone would quiz me unexpectedly. (So I wouldn't write a flashcard for Bayes' rule because I can recite it in my sleep – but I do have a flashcard for the log-odds version of it, where posterior = prior + sensitivity - false positive rate, because I haven't yet seen/used that version as much (though clearly I did remember it just now).) - Steps or details of a simple process that I'll likely want to follow later without having to reference the book. This is probably not an ideal way to select things to remember, but I have to go by something and these are usually fairly good points at which I take a short break from reading anyway, so I might as well create a flashcard. Then comes actually studying the flashcards, but that's not what you asked about so I stop here! |
> The last 5 % for me was switching over from writing my plain old notes into writing flashcards instead.
I would caution against dropping writing plain old notes for broader level topics. I've been doing SR for over 4 years, and a few times it's happened that I'm consistently failing flashcards on a particular topic. The way for me to get "unstuck" was to admit I've forgotten key fundamentals, and go back to my plain old notes for a proper review - the way I would do it before flashcards. This gives me the bigger picture, in a logical order.
Going through atomic flashcards did not help me "relearn" something.