Spaced repetition memory system (2024)
notes.andymatuschak.org
notes.andymatuschak.org
I'm not affiliated, just a big booster. For those familiar with Anki it follows the same conventions. It has an excellent system for managing cards. Adding cards is as easy as writing a bullet point: [front of card] == [back of card]. They got the ergonomics right and clearly know the space very well; it has the right keyboard accessibility and shortcuts and navigation. It supports the basics you'd expect like cloze deletions (fill-in-the-blank), image occlusion (cover up parts of an image). It manages assets like PDFs and images. It uses FSRS (the best SRS scheduling algorithm atm).
It has the best (optional) AI integration into a product I've seen except for the usual code-generation suspects. I'm learning spanish and can type into a bullet point something like "el vaquero ==< [tab]" and have the translation automatically generated for me into a forward and reverse card. I'm learning math and can cloze-delete parts of latex equations; the AI can very frequently generate excellent and accurate latex equations, which I can make small edits to as I'd like. These kinds of bonuses make taking live flashcard-based notes during my spanish tutoring sessions and math-based parts of classes feasible.
It's less low-level configurable than Anki and more "works out of the box" with a smaller extension system. I've had enough of trying to fiddle with Anki. Overall just excellent -- I'm not affiliated in any way. Development is very fast. Release note videos are incredible, minor updates occur ~weekly. I've run into a few bugs, especially when I was traveling overseas where internet isn't strong, but overall very pleased with it.
You can upgrade to the Pro version for $10/month if you want tables, PDF uploads, and more.
The $18 version is our most expensive plan that includes AI credits as well.
As H. L. Mencken observed in 1926 [0]: "Nobody ever went broke underestimating the intelligence of the North American public."
Perhaps there should be an event next year to note the centenary of that astute observation, which seems to have aged quite well.
Mochi has great native apps on macOS and iOS (and maybe more?), the cards are formatted in markdown so I can generate them with LLMs with a custom system prompt, and I just found out today they have an API so I might try my hand at getting an LLM to push new cards on its own via. an MCP server.
- it has a clunky and complicated UI: interactions with blocks/line was clunky on mobile and web.
- The table UI for showing your cards, can also be very limiting.
- Converting blocks to cards by adding `<==>` is an ingenious idea.
- The use of "AI" is really over-rated.
I stay away from some of the AI features:
- One feature generates additional context for a flashcard as you practice. This burned up my AI credits like crazy and added nothing but distraction to my practice. - Another AI feature for PDF summarizing just didn't work. It made a claim immediately and clearly contradicted by text in the exact area I was highlighting.
The cases where it works, I absolutely love.
Seeing this did make me wonder how I might be able to get better at memorising important parts of iso/iec standards at work, but I can't see how that maps to flashcards
But of course it's possible that you almost never need the same information twice, in which case committing it to memory wouldn't be particularly useful.
Mixing the questions between both pools and studying as a unit I've found has had two great benefits: 1) I'm not focusing on Technician first and then going for General as a "bonus" and 2) it helps me see the connections between the material.
Anki is set up with a long review cycle (1 day, 1 week, 1 month, then automated) and I sit down to do my reviews about once a week. In that process, I usually end up having new ideas to make notes about based on either the randomized order the notes show up in or spotting a connection between the review note and something I've been working on lately.
[1] In practice, I let many/most of these go unrecorded - I probably average about one new note per day, but in bursts.
* Learning biology, memorize terms like "anabolic reaction" or "reverse transcriptase"
* Learning algebra, memorize major groups like S_n or GL_n
* Learning statistics, memorize the major probability distributions, their means, and standard deviations
* Preparing for math contests, remember things like "Chinese remainder theorem"
That's a tiny part of learning, but it dramatically accelerates the other parts. At that point, when you're working through texts, you'll understand what you're reading without looking things up or thinking about it. And when you're engage in complex problemsolving, you'll have the knowledge ready.
Do this either on or before the first (surface learning) pass, and once they're memorized, use them in more advanced contexts (e.g. reading research papers, teaching, complex problemsolving, etc.).
All this stuff interconnects, and SR gives a fast, cheap way to start building out the simpler parts of the knowledge network.
I am one of the least qualified people in the world to write cards for a topic I am learning. I would quite likely create cards that would help me memorize inaccurate information effectively and efficiently. I'd rather not take that risk.
I use it as an accompanying tool in a real language school (learning German). I started a new Study Set from scratch, and add new words to memorize every lesson. Liking it so far!
That only does space repetition?
* Open Source
* Cross-platform
* $0 except on iOS
* Popular enough to have a community and ecosystem around it
A lot of it down to the effort this guy has put into making the deck. https://www.asiteaboutnothing.net/w_ultimate_spanish_conjuga...