Hacker News new | past | comments | ask | show | jobs | submit login
Using Anki to remember what you read (superorganizers.substack.com)
901 points by dshipper on March 5, 2020 | hide | past | favorite | 304 comments

This article reminds me of posts about "bullet journalling" on sites like Pinterest: the optimization of the productivity process or tool (or in the case of "bullet journal" posts, the beauty of the tool) is the end itself, not a means to an end. The posts are about/by "productivity" or "journalling" enthusiasts and don't map well to people who are simply seeking a bit of an organization or productivity boost, not an extremely time-consuming new hobby.

The thing in this article that most reminded me about beauty-journalling posts was the todo list. In vanity journalling circles, the todo lists are full of meta-tasks "post to instagram; new journal layout; blog post about journalling" as well as pure vanity items like "yoga!; Breakfast: saffron avocado toast (yum!!)" Likewise, in this post his todo list includes items about writing about productivity, using other productivity tools, and other such "meta-productivity" items.

There's nothing wrong with this but it reveals that the "productivity" itself is the hobby, not just a means to an end.

I think you can go too deep into it as a form of procrastination, but I think this post is sincere. Spaced repetition is backed by science as an effective way to acquire knowledge (e.g., https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3782739/).

I'm personally more skeptical of the Zettelkasten/Roam/PKB craze as being effective, but that's only because I feel like I've had yet to find a problem where it would have been the solution. I'm constantly left feeling like it may be the solution to many problems I have related to memory and remembering thoughts, but without personal experience in it, I can't say what it's worth.

I don't think it's even possible to evaluate a productivity system without using it for a long amount of time and conducting periodic reviews on it regularly to see if it's helping with anything. I don't know what the metrics to measure are, and I don't know what a review like this would look like.

On the paper journaling thing, though, I think the answer is much more clear cut: it's more about the artistic nature of it than anything else. A paper journal is almost always impossible to search, easily damaged, easily lost, and easily customized with stickers, different pens and pencils, and other decor. It's easy to find a notebook you like with a paper you like and then justify writing in it. However, I don't think there are many practical benefits that can't be applied to an electronic system that's functionally similar, save for the effect of slowing down entry so you remember it more.

A few years ago I was looking through my last.fm scrobbles. Scrobbling is passive and it's widely supported, so I did it pretty regularly. The process of reviewing them was like tapping an area of my memory I didn't know existed. I saw the songs my daughter liked to listen to when she was three and remembered trips we took while listening to those songs, and I mean vivid memories of the drive, including the wrong turn we made that made us go through that album twice. Things no photo could ever recall for me. It was a wonderful experience that I may never have had if I wasn't keeping that record.

Going back through old notes I've made on books I've read or movies I've watched gives a similar experience, taking me back to how I thought at the time and a window into how I've changed.

I certainly don't want to study everything to death, but I wholly see the appeal of a system to contain things I've known and done.

Yes I have same experience. Been using Last.fm for 14 years and that history can sometimes reveal things I didn't have any other chance to recall. My long youtube music favorites playlist is also interesting, because I usually add only things I really loved, and keep listening to it for a while, so there is just these time fragments, when I was kind of addicted to some of the songs, but against last.fm it just cannot compete at all.

> On the paper journaling thing, though, I think the answer is much more clear cut: it's more about the artistic nature of it than anything else

A counterexample as food for thought: I do all of my task tracking in a paper notebook. It looks boring -- there's no design and I only use one black and one red pen.

What I've found in over 5 years I've used more than a dozen task trackers for various organizations across different projects and teams. Each time, I have to learn the system and adapt my process to it. Every six months the productivity community goes nuts over some new app.

On the other hand, my notebook process is just the same. Paper has been around for over a thousand years and will be around a lot longer than that. I never have to keep up with the new hot productivity system or worry about the company that makes my favorite app going out of business. Paper will never release an update that makes me redo my whole process.

Paper let's me find a system that works and stick with it. Paper gets out of the way. I can focus on being productive instead of productivity.

Totally agree. Before starting with a paper journal about 4 years ago, I was using todo.txt for tasks with some org-mode to supplement. I stuck with it for a good while (2012-2016) but needed more flexibility.

I did go through an experimentation phase trying to do what the "official" bullet journal people and the bujo "influencers" suggested. Instead of having to restart and adapt to a new app, I could try a little bit and discard what didn't work or I didn't like.

To the point that grandparent made about:

> A paper journal is almost always impossible to search, easily damaged, easily lost, and easily customized with stickers, different pens and pencils, and other decor.

I can't run a grep on a notebook, no, but I often know about when I took a note on something to find it quickly. I usually run through a single notebook for a bullet journal per year. I started with fountain pens the same time I shifted back to analog. I have a little fun with colored ink, changing colors at most once a week. I tend towards waterproof/water-resistant inks that can weather some random wetness and still be fine. Some of them are even UV-resistant/forgery proof. I'm not taking hours upon hours to create the perfect spread.

I'm not writing outside in the rain. I've misplaced my phone more than my notebook. With the way that many are paper phobic these days, I'm not in real fear that my book would grow legs and walk off.

My BS detector went off immediately when I read the link about Zettelkastens, and then I googled "Zettelkasten", and I skimmed through five different articles that were each pages long and talked over and over about how wonderful Zettelkastens are but didn't give any information about how they work or how to implement them practically. That's a pretty reliable indication that the service is the end, not a means to the end.

For the uninformed, a Zettelkasten is a searchable, unordered, labeled database. That's it.

I think your BS detector has a bug that may cause you to miss out on some good stuff! ;-) Sometimes there are ideas which are really good and very simple, but hard to explain because understanding them requires a shift of perspective which comes more easily from hands-on experience than from verbal explanations.

"Zettelkasten" is more of an attitude toward note-taking than a specific technique... it can be implemented in many seeingly almost unrelated ways, from Luhman's index-card box to personal wikis, etc. The defining characteristic is more in the way you integrate note-taking and note-review into your daily workflow and thinking process.

The best description I've found of it is this article: https://www.lesswrong.com/posts/NfdHG6oHBJ8Qxc26s/the-zettel.... Slightly long, but it completely describes how they work and how to implement them practically (on notecards). On the other hand, the articles on https://zettelkasten.de are pretty difficult to understand for the uninitiated.

> For the uninformed, a Zettelkasten is a searchable, unordered, labeled database. That's it.

I could spin up ElasticSearch or something similar and have it crawl all the files on my computer and get an 80% solution. Kinda like Apple's "Finder" functionality.

Almost, but not quite. Zettlekastenicians (Zettlekasatengineers?) try to only put succinct thoughts into their systems that they can tend and grow into larger productions of thinking or learning--there are some more specialized amounts of explicit metadata (your thoughts, ideas, inspirations, questions) that cannot be crawled and indexed by machine which will only see the implicit metadata.

If you index everything together you get something that it is merely a pile of searchable things. You do not experience the "conversation" with the slip-box as Niklas Luhmann described it. Software like http://zettelkasten.danielluedecke.de/ that implements the "digital version" of the system suffers from what I like to describe as the "Hudsucker Proxy" problem: A young idealist visionary comes out and holds a big plastic loop and mutters "You know... for kids!" and nobody has a clue what that is supposed to mean. It isn't until someone shows how to use the "big plastic loop" before anyone understands what's happening. And that's really hard to do with something as abstract as a Zettlekasten despite it just "being little notes gathered together with lots of numbers and letters pointing at the other notes" or an "unordered database".

Now I want to find someone who uses this and follow them around for a day or two asking questions.

Oh, so it's just a hipsterish new name for a personal wiki.

Kind of, yes, but from my understanding of the original system as devised/practised by Luhmann was that the real value of it laid in the connections between the individual notes and the work of actually filing things away in that web of knowledge.

It's trivial to just write everything interesting down in whatever format, but the idea behind Zettelkasten is that by examining each piece of information before slotting it into its proper place in the grand scheme of things made you re-evaluate old notes and consider how things fit together as a whole. This aided memory and improved understanding of the topics.

My gut feeling says the concept is sound. You're not just archiving information, you're piecing together a complex web of knowledge and relating new information to what you already know, which in turn might spur original thought. That might happen with a wiki, but it's not a given.

(It also pre-dates the concept of a wiki ;P)

Well, the original one isn't unordered, and isn't meant fundamentally to be "searchable" (its original "implementation" being index cards).

Here's my attempt at conciseness: A zettelkasten is like a bunch of files in a single folder with 1. a naming schema that kinda indicates what follows what (an ordering scheme, but persnickety and optional, see https://zettelkasten.de/posts/luhmann-folgezettel-truth/), and 2. links in between the files, originally just by writing the names to be manually looked for.

Some people use 2 to be able to get rid of 1, as with an indication of PREVIOUS (link) and NEXT (link)

The point is that you have one thing on each card / file that you might want to go back to on its own.

It makes a lot of sense in academia where you want to be able to trace ideas in a referenceable way, identify your own takes on bits of other people's work, and connect different sources together when you see similarities.

My very elementary (just skimmed some blog posts about it for a few hours this morning [1]) understanding of why Zettelkasten is a game-changer is that: its a system of organizing notes in a bottom-up, atomic fashion, instead of the usual top-down categorization where the notes don't lend itself well to reviewing or linking with concepts you've previously learned. Basically, a graph (where links are bidirectional) instead of a tree. The power of such a system is that it maps closer to how knowledge is stored in our brains, and that the more you use the system, the better benefit from "network effects".

In traditional note-taking systems, the category is the most obvious first-class citizen. In ZK, the little ideas take precedence, because some little idea that belongs to category A can easily end up belonging to other categories B, C, D, ... as time goes on and you study more subjects/discipline.

I'm sure the idea I'm describing is nothing new, groundbreaking, or unique to ZK - but this is the impression that I got, and I've had the pain of having very poorly managed knowledge via Evernote/OneNote/Notion/etc. before. I also dabbled with Anki and mindmapping. I believe the optimal solution is to combine all 3, since mindmaps have the "graph" structure part down; Anki only focuses on recall, but not conceptual understanding/linkage; and traditional notes, while obviously storing more information, are usually organized in hierarchical top-down structures that don't lend themselves well to reviewing or building a deliberate map of knowledge.

Another alternative to mindmaps is something like Jerry's Brain [2], where irrelevant nodes are collapsed until you explore them. Jerry has given extremely compelling presentations because he's used it for something like 20 years.

I am aware of a bias as I write this: I simply might not have had enough discipline when using traditional note-taking systems before, and am going through a bit of a (pre) honeymoon phase with this new ZK stuff. Just my 2 grains of salt.

[1]: https://nateliason.com/blog/roam

[2]: https://www.jerrysbrain.com/

They are also usually todo lists of a size that... are small. I don't need to manage a todo list of 10 items. That's not the problem. How to manage 100+ items with due dates is a much more interesting question, and productivity porn is not the place to find answers to that.

(I still read it. I love that people obsess so much about it. It's like going into an office supply store, and just lovingly gaze at organizers :)

Well, that bring memories. In my other life I had o office job with generic title & over the time I became the one of the oldest employee (first joiners, oldest not by age). So I found myself managing timesheets, monies, appointments, cars, vacations of 100+ people. I started a journal & also came across Bullet Journal. Internet was like decorate, spend time on it; but mine was purely functional.

Simple basic page with To Do Title. New Day starts after a blank line. Each task is a blank square box, task text. When done, a tick in the box & done-date in right margin.

At end of the month, all pending tasks get reviewed; no-longer-needed gets a cross X in box; needed ones get moved to Big To Do page, original entry gets a arrow in the box. Once all tasks sorted, done, cancelled, moved, the big box in front of title gets ticked with today's date.

Big To Do page has a box, task original date in dd/mm format, then task.

Some specific pages like Financial, 6-Monthly, Logs etc with templated tasks prefilled with current months data, kind of checklists. 6 Month page gets tasks required not in immediate future, but sometimes in ^ months. If available, deadline date gets mentioned in task.

All pages numbered at lower right corner, in continuous series.

An Index page with Page Title and page numbers.

I was using a A5 ring binder with two holes.

It was not pretty to see. It was readable, not chicken scratches. Occasionally some tasks get red marks, or green. All completed pages gets scanned as pdf.

It was amazingly working. 4 years 9f my daily office work, keeping up with everybody's every type of requests. I miss that now, that now I am in a job where we can't take our bags/books to desk.

People have different needs and motivations. A friend was a fancy journal keeper, and the “ceremony” around planning and executing the fancy journal stuff was part of the review process of upcoming and past tasks.

Thanks for posting this; you have expressed so eloquently my experiences and opinion on this trend.

I've used spaced repetition to help learn stuff in general. It's one of the most rewarding things I've done. Some general thoughts:

1. A lot of spaced repetition apps don't optimize for fast card creation. I like to make a card as soon as I come across something interesting, but the delay to get Anki started, and to create the card, etc just makes it frustrating. I wound up building my own app for this.

2. Cloze deletion is surprisingly effective. For those who don't know, cloze deletion involves taking a sentence, blanking the interesting parts out e.g. "The moon landing happened on _______", and trying to recall the blank parts. This is effective because it's a quick way to make cards, thus solving the problem of slow card creation.

3. I fared better with lots of small, one-sentence cards. My rule of thumb is that a card should fit in a tweet.

4. Subjectively speaking, using spaced repetition didn't just help me recall stuff in my cards, it also helped me recall stuff in general.

One of the issues Anki has, I find, is that if you end up taking a long break for whatever reason, there's no way to "reset".

I used Anki to study Chinese for a period of about 10 years. At some point I decided that I wanted to memorize the poker "outs" (probabilities of filling out a hand based on what had currently been dealt). Then I went through a time where I was really busy and didn't study the poker deck for a month (but I made time for Chinese). When I came back to the poker deck, the "spaced repetition" system was completely broken: I had a massive long list of cards that had expired, most of which I'd completely forgotten; but it just kept showing them to me in one giant loop, rather than focusing on a few to actually teach me. And I didn't even have a clean way of telling it, "Just pretend I haven't seen any of these cards at all". I ended up just deleting the deck; that discouraged me from doing anything else I wasn't willing to commit to doing every single day.

I've used Anki for about ten years, and I'm going to tell you a huge secret...

There's absolutely no need to "reset". Ever.

If you have only ten minutes to devote to Anki, then only spend ten minutes. If at some later point, you have more time, then spend that time.

Set the maximum reviews per day to something you can do most days -- for me, that's 250 cards. If you're behind, turn off 'new' cards. And eventually, you will catch up.

What if you don't do Anki for a month or two? You still don't need to reset. Anything you remember after that month will have a much longer time until you next see it.

Only use Anki as much as you have time for it. Let it figure out which cards to show you. Resetting messes with the algorithm for which cards to show.

I'm also over ten years on Anki, and I've racked up 2/3 of a million review during that time. I've taken too-long breaks several times over that time period, on a variety of subjects (decks in Anki). Sometimes years.

Each time I've gone back to a deck, I've just slugged out the few days of heavy reviews and let Anki take care of the rest. The stuff that I've forgotten, Anki will nag me with. The stuff that I've retained for years, stays retained.

One thing that I do suggest is to limit the maximum interval to 365 days, and to remove the review limit. I also tend to use the "hard" answer on cards that I've retained in decks that I've ignored for some time.

Having used Anki for 10 years, reset is definitely the best option.you fly through the cards you remember at the start, and don’t get bogged down by cards your ‘supposed’ to remember.

(SRS fails in this regards)

There are a couple ways to do it that I've found. None of them is perfect for every situation I've encountered. I think that part of the reason why there's no clean way to do it is that there are so many specific things someone might want to do:

- If you want to just push cards to the back of the queue, but remember timings and history, there's a built-in command to do that.

- If you want to change the current interval of a card, but keep your history, there's also a built-in command for that.

- If you need to do the above in bulk, the "Reset Card Scheduling" plugin can make this a lot more convenient.

- If you really do want to completely reset a card, forget all history, etc., and treat it as a new card, there's the "Remove Card History" plugin.

Finally, when returning to old decks, I found I usually get the most mileage by leaving all that stuff alone and just suspending the whole thing, and then un-suspending them at a steady pace. I neglected my Kanji deck for months, and I used that approach to get myself back up to speed by doing the catch-up review in the original (RTK) order I originally learned them rather than based on Anki's priority.

User plugins are a cool feature that let users do all sorts of things you couldn't forsee as the app developer nor want to encumber the UI with every possible feature.

But it feels so hacky when you need them for things that seem like rather elementary functionality like resetting a deck.

Calibre feels the same way like needing a plugin just to estimate page count. I feel lucky when a plugin actually works.

For what it's worth, I keep everything in a single deck. I never stop reviewing stuff and "interleaving" makes learning more effective than "blocking" (studying things in blocks). Source: Bjork et al 2013 (https://www.researchgate.net/publication/231610455_Self-Regu...)

I took a modified approach to this in university, though nowadays I mostly do the same.

For each new class, I would start a new deck, and after the course was over I would move it over to my main (huge) deck, and it's worked surprisingly well.

Typically (now that I'm out), I add small cards from time to time to my deck, but if there's a large quantity of things I am working on learning at any given time (say I'm going to add 50-100 new cards), I'll put them in their own deck for a little bit.

Ultimately, my repetition does end up being a wide array of subjects though.

This is why tagging is useful. Anything cards I added for United had the tag of the class it was from. You can then do a custom study session with just those tagged cards.

Same: single decks are nice. It’s still hard if you break your habit though...

I'm kind of new to Anki myself (started a couple of weeks ago) so I might be wrong, but I think you can limit the amount of cards you want to review on a per-deck-basis. The default-number is quite high (200 per day I think?).

If you make that number the same as the "new cards per day"-number, that might be a work around for your issue?

From a different (but really probably the same) angle, this is the frustration I've had with Anki as well.

I just want a "dumb" flashcard application that doesn't try to apply "smart" techniques/heuristics.

I was just looking for a way to do the same paper card quizzing digitally without having to actually carry decks on my person.

Every app I've tried over the years powered under the hood by Anki had this issue.

It's been a few years since I last checked it out, but an option for a dumb mode would have had me throwing money at it just to even try it out.

What's the use case for this dumb mode? Does Anki not have a "cram" mode that just goes through all flashcards with none of the fancy interval logic?

One of the wonderful things about software assisting with spaced repetition is that it automatically selects and remembers efficient intervals. It's crazy how little time is spent per flashcard with this automation.

I use Anki for that all the time

There’s a “cram seen cards” feature under custom study that’ll just give you a temporary brand new deck you can run through in random order

I don't know. I just tried to briefly check it out again. I'm sure I could be giving it a fairer chance.

Clearly, millions of people are benefiting from this app.

I'm just finding the features and interface very bombarding. And, the sheer number of screen taps just to study the flash cards in the manner that I'd like (and even then, not fully within my perceived control) is pretty repelling.

Thank you for pointing that out that feature. It might help eventually with conversion.

I don't see that there's any difference in responding "Again" to an expired card vs reclassing that card as new. When I'm catching back up on an old deck I just turn new card off and limit my reviews until I've worked through the backlog this way.

Suppose you have 50 unseen cards, and you have it set to show you 5 new cards every day, and energy for about 10 reviews.

Then on day 1 it will want to show you cards 1-5; day two it will show you 1-5 again and also 6-10; day three it will show you 6-10 again and also 7-15 (to simplify somewhat). So you're having some cards you know and some that are new.

If you have 50 cards you've forgotten, then on day 1 it will show you 1-10; on day 2 it will show you 11-20; day 3 it will show you 21-30, and so on -- all completely new. That makes it far more of a grind.

I ended up with similar problem. Started wondering if there’s some parameter you can tune to avoid this ”giant loop” issue?

I would prefer to iterate over small set of cards until I know them quite well.

Would it be possible to share the deck? I would love that (for poker).

You can do custom study of all the cards.

For frictionless card-creation I have made an org-capture template for quickly adding cards to an org-mode file via a custom Emacs pop-up frame. This is accessible from everywhere with the keybinding (Super + c + a).

Exporting to Anki from Emacs is done via the anki-editor mode for Emacs.


There is also org-drill that implements spaced repetition in org mode.

I think org-drill is just better overall since it's all stored as plain-text at the end of the day and each card is really editable and accessible. Anki in my experience has a clunky, slow GUI. It can't beat the creation and editing speed from vim keybindings and macros.

The thing I miss with org-drill is being able to do it on the phone. I have considered trying to integrate it into Orgzly but it does seem like a big project, unfortunately.

IME org-drill is quite slow to run through reviews, with a lot of lag.

This sounds like an issue with something else you have installed in Emacs. I suggest running the profiler to see what is actually taking time. Org-drill itself on an otherwise vanilla config is fast, so it could be that you have hooks that are slow or something.

Thank you for creating and sharing that, I hadn't thought to look for an Emacs to Anki connector before.

Being able to add my slow-but-steady Lisp learning notes into Anki would be very cool.

> but the delay to get Anki started, and to create the card, etc just makes it frustrating

I have this URL in my bookmarks toolbar:


It enables me to directly add new cards (just don't forget to sync)

> 3. I fared better with lots of small, one-sentence cards. My rule of thumb is that a card should fit in a tweet.

Can someone tell me if this is a crazy good idea? Social network + spaced repetition. Why should you build all your own cards? Taking a class with classmates? Create a group & create cards during lecture, review/curate cards afterwards during study session, then rate which ones were most useful after the exam (or homework).

I've head a similar idea, except not in context of a group project but of a wiki. So there's a wider audience for any particular topic and more people to create and improve the exercises: cards, cloze deletion tests, etc.

I think it's important that exercises are tightly linked to the source material, like to a specific paragraph of an article, etc. So these materials should probably be added to the system as well. * With the same starting point it should be easier to have a proper discussion. * A new person can read the source and understand cards more easily with context. * Later if they've forgotten the topic completely they can reread it and hopefully remember faster.

I also had some ideas about being able to discuss and alter every paragraph in the source. Allow it to evolve to be more clear as people come and discuss confusing points.

Memories are very personal; something that works for you won't work for your friend, even if you're learning the same thing in a class.

The actual act of creating the cards is also useful to the learning process. Even if you are copying someone else's cards, copying them by hand will lead to better retention than having a computer import them.

I used to think that. Now I’m not so sure.

not crazy, i think

but my cards are hard to read by other parties, when I optimize for my own learning

Agreed. I guess what I'm suggesting is a socialization of the act of card creation & curation so the group creates/uses them.

Since you've made you're own app, and have noticed that a good card is a good tweet: I want an app that puts my cards in my twitter feed. I check twitter way too much and don't like opening Anki to review my cards. Not sure exactly how it would work, maybe I see a few tweets, and then there's a 'card' tweet, and I interact with it as usual and then get more tweets. Just an idea.

For speeding up the card creation I've recently made a CLI application [1]. It generates cards from Markdown files.

I'd recommend having its own repository with Markdown files grouped by target domain using tags for every card. It'd help to search quickly relevant cards even in the same domain.

[1] https://github.com/ashlinchak/mdanki

Nice! I made a similar thing for myself, but without nearly as much polish as your project [1].

I use a custom text file format to allow creation of cloze deletions and reversible cards as well as basic cards. I also annotate text files on export so that I can export the file again without creating duplicate cards.

The big shortcoming of my script is that it generates .tsv files and Anki only allows .tsv files to contain one note type. They also do not incorporate media.

I'd like to be able to sync edits made in Anki back to the text files and vice versa. It'd be really cool to integrate a text file parser and alternative card editing mode into Anki with a plugin.

Perhaps I'll send some PR's your way instead of duplicating effort.

1. https://github.com/isaiahstjohn/flashbang

I have some cool decks, but I am terrible about reviewing them frequently. It just becomes another chore. Maybe I should really embrace it. I spent a few months brushing up on biology and that was really amusing but I fell out of the habit :/

Using the reminder aka notifications feature of the android app got the habit to stick for me. Total pita to figure out how to get it turned on though :)

In my experience with Anki all four of your points are spot on, especially number 3.

A big problem I see with pre-made decks is that they contain just too much information.

I just...can’t relate to living this way. Maybe you have to be a special type of person, but this whole lifestyle just seems exhausting and over-managed.

I agree. In fact, what I find the most interesting part about reading is .... forgetting about it! Basically concepts will sip thru, but learning the details is just a waste of brain cells, in my opinion.

Just leave the brain do it's job, don't read as it's a marathon, take breaks, think about that bits you've just read that was interesting, then promptly forget about it. You'll forget the details, but not the backbone of it.

Details don't matter in the end. Quite frankly the idea of knowing /by heart/ the name of the greek goddess blah blah blah he uses as an example would bore me solid. Worse, anyone knowing it and telling me about it would bore me solid :-)

> don't read as it's a marathon, take breaks, think about that bits you've just read that was interesting, then promptly forget about it.

That's basically the spaced repetition model. You need to forget a little bit, so that working memory isn't saving you, and the recall process takes effort. That's what builds long term memory. Spaced repetition systems just extend that process beyond the length of the book, in a time shorter than rereading a book.

> Quite frankly the idea of knowing /by heart/ the name of the greek goddess blah blah blah he uses as an example would bore me solid.

I agree, the names of greek gods is a bad example, unless you're a student of mythology. Spaced repetition is hard work, so probably only deploy it on things that matter. A better example might be Bayes Theorem. Incredibly powerful, but unintuitive and easily forgotten. Cards for Bayes' theorem might include the purpose: "Bayes' Theorem calculates how to adjust our prior beliefs given new evidence", as well as cards about the formula, or an intuitive visualization.

Or, maybe you want to learn more about the linux internals and spend a portion of your time memorizing the meanings of signals and errnos as a small part of a larger program of study. Sure, you can look these up in a man page, but part of the value in knowledge is knowing things exist. How many developers do you think know of SIGUSR2? Or... SIGBUS ;)

I don't see how cards would help me with mathematical ideas. You have to get them on a fundamental level, and then they're hard to forget. I'm more likely to forget the name of the theorem. But space repetition might be useful for formulas.

But I'm not too crazy about spaced repetition. I used it to learn all the capitals in the world, and after about a month I managed to do all of them without error. But after two years of barely using the knowledge, I only remember half.

Of course, I could keep reviewing it every few months to keep it fresh, but that's just not an efficient system, when you think of how many things I'm supposed to (and do) remember.

For me, at least, the best way to remember things is by tying them in as many associations and metaphors as I can, and that gives me a pretty reliable recall. It has some downsides: It's a bit more work than spaced repetition (requires creativity, for example), and it's not that good for unconnected data (but then, are random facts that useful anyway?). But I think that as a long-term method, it's much more solid.

I have had reasonable success using spaced repetition with math proofs (written in LaTeX). I create cards sparingly, often based on questions from problem sets which I shouldn't have gotten incorrect (i.e. the mistake stems from a fundamental misunderstanding of concepts, not a misstep in algebraic manipulation)

> I don't see how cards would help me with mathematical ideas. You have to get them on a fundamental level, and then they're hard to forget.

I find it too easy to trick myself into believing I understand a concept on a fundamental level. But often that "understanding" slips away, and six months later when faced with an example problem out of context I struggle to solve it.


Also notice how he mentions he remembers "trickle down economics" was a "very important idea" during Reagan's presidency. But there is no mention whatsoever about the idea itself, how it's considered today, whether it was a good or bad idea, its relation to neoliberalism, whether he understands how politically divisive the idea is, etc.

Someone remembering trickle down economics was something from the Reagan era tells me nothing about their understanding of that idea and their opinions about it. Precisely the bits I want to know!

The article not about 80's economics. He was just making a point, of course he knows the concepts he mentions with more depth than discussed in the post.

Maybe he does, maybe he doesn't.

He is arguing for memorization. I'm saying memorization of this kind of "facts" seems unconvincing at best. Memorizing a "data card" about Reagan with this scribbled on it seems like something that would only be useful for a trivia quiz. "Mozart composed this or that", "Reagan's presidency had something to do with trickle down economics", "Washington had false teeth [1]", etc.

I know this isn't your main point, but I cannot resist to comment:

> The article not about 80's economics

It still pervades political discourse in many countries of Latin America, and many believe it's the cornerstone of neoliberalism, so...


[1] I wanted to write "wooden", basing my knowledge in LucasArts' Day of the Tentacle, but apparently this is false.

Indeed, a fact like "Shostakovich composed his 5th sympohny in 1937" may have little value on itself, but its value is compounded with related facts. For example, if you also knew that Shostakovich was Russian, you could deduce with reasonable certainty that he composed this piece while living in the USSR (knowing also that the USSR was in place in 1937).

This in turn might allow you to enjoy the piece more, because you can relate the music to the time period and better understand what it depicts and what the composers intention was.

Now, you could also simply read a brochure about the piece/composer before the concert and you would know the same or more, but this would cost you time you may not have at that moment.


Anyways, I don't think he is arguing for memorization, rather, he argues that you should first learn something (by reading) and then use some tools to remind you of what you have learned in order to not forget it. It turns out that the tool he chose for this is one that reminds him of some facts once in a while. The intention is not necessarily to know these facts, but to be reminded of the learned concept through these. Whether this is an effective method or not I do not know. One could end up becoming very focued on the facts, forgetting the deeper knowledge behind them (as happens in eductation, but for different reasons).


Maybe I went a bit too in depth with this :p

I dont understand your point about my point (the article not being about 80's economics). Are you saying that because it is important, the article should have elaborated on it?

> I dont understand your point about my point (the article not being about 80's economics). Are you saying that because it is important, the article should have elaborated on it?

Not the person you're replying to, but my feeling is that if you're writing an article about a system to use to usefully remember things, if you're going to then cite an example of something you've remembered, you should focus on what is actually important to remember about that fact. Saying "I remembered that 'trickle-down economics' was 'important in the Reagan administration'" does not do that, and could demonstrate that this memorization method is actually teaching you to remember the wrong things.

It doesn't matter if the post is or is not about 80s economics, but if you're going to use an example out of 80s economics to prove that your method is good, then show that your method actually helped you remember something important about 80s economics, not a piece of trivia that not only isn't useful, but is counter-productive to learn about if you don't remember the meat of the idea, and things like whether or not it was a good idea.

It's possible that the author did actually know and remember the useful parts of those facts, but did not actually demonstrate that his method helped him remember actual useful facts... which is kinda the point of his article.

Thanks for the reply. I see what you mean, but I'm not entirely convinced that knowing trivia is really helpful, except maybe to impress your friends (I tend to effortlessly remember all sorts of crap from Wikipedia, the more useless the better recalled, which makes for fun conversation... sometimes).

As for my nitpick, it was a tangential point I couldn't resist making: that trickle down economics is not exclusively an 80s thing, but (sadly, in my opinion) remains very relevant today.

I absolutely agree that memorizing facts help you understand the world. Famously, students find memorizing dates in history tedious and useless, but it's important. Any new event you hear of can be compared with many concurrent events. I don't know all that much about Shostakovich, but hearing that he composed a symphony in 1937 triggers a sequence of questions. What was his relationship to Stalinist purges in the 1930s? Did he have contacts with the Russian post-revolution emigre artists across Europe and America. What did he think of Stravinsky or Ravel? I'm listening to his fifth symphony now and I can't help but imagine the ominousness of the first movement having to do with all this.

I kind of think it's in the same vein of the millennial trend of monetizing your hobbies. We've been told over-and-over that you have to maximize every experience in order to get ahead, so there's this compulsion to do so. I used to feel this way, but eventually I realized that I'm not in grad school, and it's ok to just enjoy things as I'm reading them.

As a recently relapsed biochemistry student, with a wide variety of classes in both biology and chemistry, I use Anki exhaustively. I made at least 10k cards last semester, and that was a relatively light semester in terms of credits.

With that said, I can't imagine using such a powerful, and honestly demanding, tool for everyday knowledge and tangential facts (e.g. Greek mythology as a software engineer). It strikes me as far too much tool for the job.

It seems to me that, if you struggle with retaining things you read, some simple, lightweight note taking strategies, especially handwritten, would be sufficient. I believe that's actually a recommendation that his book recommendation, Where Good Ideas Come From, makes.

How exactly do you end up with ten thousand cards on Anki on a light semester? Are you using it for problem sets?

It was a light semester in terms of credits, but it still included classes in biochemistry, physical chemistry/quantum mechanics/spectroscopy, and molecular biology.

It's also important to make useful cards, and the most useful cards are almost always very simple and can be answered relatively quickly. So, for example, a complex, multi-step DNA repair mechanism (of which we needed to know quite a few, and they differ for eukaryotes and prokaryotes) for a molecular biology class becomes quite a few individual cards. In the end, there is just a lot of stuff that either just has to be memorized, or that benefits from improving recall speed for taking a 60-90 minute exam.


I mostly read for pleasure (and occasionally for work). Even for work, I don't tabulate what I read. I read a lot of paper books too, where uploading stuff is not practical without a lot of fuss (OCR?).

I love reading. And yes, I forget a lot of stuff. But if I had to go through all this whenever I pick up a book, I'd simply stop reading.

You don't have to do this with all the books. Only with the ones where you want to commit the important stuff to memory.

Ok, but I understood the person mentioned in the article does this with every book regardless of subject. And like the OP said, I just can't relate to living like that.

Maybe OP only reads books that he wants to remember, so doing that for every book is valid. On the other way around, maybe some people wouldn't be able to relate to reading a lot of books when you don't want to commit what you learned to memory

He mentioned in the article that he first reads a bunch of book summaries to decide which books to read in full.

So he only reads books he want to remember, right?

I feel exactly like you and when I read stuff like this one - I wonder whether those people understand they're doing stuff wrong and that writing about it won't magically make them better. It's as if they're trying to hack through their own shortcomings by throwing computer applications to the mix and writing an article about it makes it work.

To me, this article is just more white noise.

The exact methodology in which he manages his life is probably less important than the fact that he's making a wide-sweeping effort in the first place, and maintaining consistency.

The opposite would be never writing anything down, never knowing what you're going to be doing on any given day, never estimating how long a task is going to take, reading books while making no effort to retain any information, etc.

We all dedicate energy toward transforming the natural chaos of life into order, even if it's just forcing ourselves to work a certain amount of hours per day, or re-reading paragraphs in books because we weren't paying enough attention the first time, or making grocery lists before we go to the store. These are things most of us do deliberately because we believe they make our lives better.

It's likely that there are certain things that he could be doing better, but I have no reason to believe that the structure he's set up isn't better than what he was doing before, and I think it's probably far more effective than the structure (or lack thereof) of most people's lives.

and yet he is only one whiteboard failure of inverting a binary tree away from being a total failure

Yep, there's this quote from Maya Angelou

"At the end of the day people won't remember what you said or did, they will remember how you made them feel"

This is basically how I approach reading as well. I think it's fairly hopeless to have a huge organizing system that assists in trying to keep information sort of floating around in the conscious mind.

Human minds aren't really fact gathering machines so I think this is largely futile and the number of things to remember is too large anyway.

I think there is more sense in trying to absorb what you read, hope that it leaves an impression of some sort or gives you a cue when it becomes relevant and just go with it and using active time for creative things.

Yeah, my initial thought was "why are you building all of these systems to remember random factoids?"

Me neither. Especially as one of the best parts of my day is spent having a drink to forget the day’s nonsense.

I'd love to join you for a drink after work but I can't do real life stuff like that right now, I need to get home to memorize some anki cards.

I recently started reading "Clean Code" by "Uncle Bob" Martin and I've decided to incorporate adding notecards on what I've come across. I've created a few cards already.

But in terms of making cards on books I read for pleasure, I agree in that it seems like overkill. One of my favorite biographies is "Peter the Great" by Robert Massie, and as much as I love reading it attempting to memorize names and details and jam pack my brain with facts would fry my brain.

Zettelkasten has been making the rounds here this year. His process appears similar to the org-mode process using Jethro Kuan's org-roam. I've started using it and while it can be a bit tricky to set up and there's a couple mildly annoying bugs (with workarounds), for a solo project it filled the niche for me perfectly. I use it for all note taking and reading now and it is very satisfying.



Also checkout out org-brain [0]. Similar idea - linking concepts, but a completely different implementation. A much more mature project.

[0]: https://github.com/Kungsgeten/org-brain

I also did not find org-roam super easy to set up, but after a second try I have it running and am enjoying it.

Edit: I tried it a month ago - before installation was more than just (load)'ing the file. Now there are nicely documented use-package instructions and a layer for spacemacs

Same here, org-roam just kind of "clicked" things into place for me.

I've modified a template for single-subject documents called "tags" so they're stored in ~/Org/tags/. I've also configured org-journal to store files in ~/Org/Journal/

One thing I'll be interested to discover, in the future, is how I archive or age out information. Most work stuff right now is related to a single project, so what happens when that project ends? Will I need to move things outside of ~/Org/ to remove clutter, or will I just get used to it?

Split tasks from your knowledge base. Archive tasks/projects aggressively.

Anki usage is huge in medical school [1]. There are a handful of huge premade decks students drill hours a day.

I think flashcards are hugely underrated for software developers. Memorization is commonly dismissed as unnecesary when we have searchable documentation and Stack Overflow, but the same argument could be had over touch typing. There is enormous value in moving data from disk to L1 cache.

If nothing else, having solid, open source, premade decks for the core web technologies would ease the path into web development for boot campers and other newcomers.

The first step in that direction is creating a text file format similar to Markdown or TOML for creating, editing and sharing flashcards. I created a hacky solution for myself [2], and another project was also shared in these comments [3].

1. https://www.reddit.com/r/medicalschoolanki/ 2. https://github.com/isaiahstjohn/flashbang 3. https://github.com/ashlinchak/mdanki

Why not use techniques memory champions use though? Never heard of one of them using Anki. Mnemonics, imagery and memory palaces on the other hand... It's like all the amateurs swearing that steel bikes are the best when all the professionals use carbon fiber ones.

Anki and mnemonics are complementary tools. Use them both.

If Anki’s UI/UX turns you off, I’ve built an alternative called Mochi (https://mochi.cards/) that uses markdown.

Incidentally it also supports Zettelkasten like note taking like described in the article.

Looks great! I have a few questions.

- I have a huge Zettelkasten note archive (I manage it through Sublime-Zk, which is getting a bit slow to operate it). Can I somehow directly import all my Zettel notes?

- I tried to import my Anki cards, but it seems like while I could go through them, I couldn't assign them to any particular deck/reclassify them. They show up as part of a "Not Found" deck and I don't see any option/drag-and-drop method to move them anywhere else.

I'll definitely be giving Mochi a try. Thanks a lot for creating it!

> I have a huge Zettelkasten note archive (I manage it through Sublime-Zk, which is getting a bit slow to operate it). Can I somehow directly import all my Zettel notes?

You can import the markdown files as notes, but it won’t maintain the linking. If you send me a sample I can look into adding a Zettelkasten specific importer.

> I tried to import my Anki cards, but it seems like while I could go through them, I couldn't assign them to any particular deck/reclassify them. They show up as part of a "Not Found" deck and I don't see any option/drag-and-drop method to move them anywhere else.

Sounds like the importer derped a bit and couldn’t find the deck associated with those cards. If you send me the .apkg file I can take a look. Anki’s data model is a little weird and there’s probably some edge cases I haven’t accounted for.

Hey Mochi looks great, really nice work :)

I created something similar which I no longer work on (but keep running on a maintenance-type basis). Feel free to be inspired/take features or just compare with another approach.

Good luck dude!


Where does it save the data locally?

I'm interested in the longer form note style but it be thing that keeps me in anki-land is the Android app. Is there a way to export to Anki? I hope understand that that question is counter to you the app's model.

No export to Anki, but the web app is mobile responsive, and I'm working on native mobile apps at the moment.

Thank you. This exactly fits my needs. I will switch to Mochi.

This is a joy-sparking product, as I've been turned off by Anki's UI/UX especially when it comes to code snippets, and I'm very accustomed to Markdown.

Really love the look and feel of Mochi.

Would love to see sample decks for languages. Like learning basic 1000 words & sentences in different languages.

I'll try using this for the next weeks but having used it for the last hour, I definitely like the feel of it a lot more than Anki.

This app is truly a savior. I use Anki for my language learning and for other things, I use Mochi.

Which spacing algorithm are you using?

I found the answer here: https://mochi.cards/faq.html

You have a distracting typo on your title page - "Spaced Repetion" :-)

Just curious whether this has any connection with Notion?

can you export export/save the cards?

Yeah you can export decks. The export is .mochi format, but it's just a zip file with some plain text data and attachment files (like images, audio, etc).

interesting. I'd found the app before but couldn't tell if there was an export feature. I'll try it out

Was just taking a break from my Anki session to check HN.

I'm a regular user, though I'm terrible at consistent, daily reviews. I somehow manage to keep on top of it, even after skipping weekends or entire weeks. You can tweak the settings to make it easier for less-than-perfect users.

Lifehack: Anki + exercise. Combining Anki with an elliptical or stationary bike improves both my review and exercise frequency and duration.

I've made my peace with the terrible UI/UX. Like many folks here I tried to roll my own SRS app at one point. But my review sessions would always get derailed with brainstorming new features or dealing with bugs. Personally, I'm not sure that the productivity gains from improving the UI/UX are worth the development time (especially when the whole point of using Anki is efficiency).

With bluetooth headphones I've moved to doing everything I can while exercising. Even sitting down to read a book feels like precious time that could be spent combating a sedentary lifestyle like listening to the book instead.

I find it difficult to do anything intellectual while exercising at a relatively high intensity. What intensity are your workout sessions? But perhaps review is different than trying to absorb a new concept?

Personal anecdote, but there's some optima for information retention at low intensity exercise. This is around moderate to vigorous walking pace or a very slow jog.

Increase intensity beyond that point and your brain's information processing ability drops below baseline.

I've found this consistently true. At lower intensity cardio for example, I can watch educational youtube videos or follow along a movie. At high intensity, I can't even watch an action movie and follow the plot. I have to switch to music and/or a sports action video (no meaningful dialog).

The more brainless the activity, the easier it is to do parallel flashcards. An elliptical or stationary bike session will be easier than an equal intensity run where you have to worry about your gait. Flashcards are little atomic chunks for you to process independent from one another. When you do flashcards the only thing you need to concern yourself with is the short prompt on that particular card in front of you. So even if you are working out pretty hard, you can still work your way through a deck, it's just going to be at a slower rate. This differs from reading a textbook or journal article, where you have to keep a certain amount of the previous passages in memory to understand the current line. You will hit a point where flashcards stop working but the ceiling is higher than with other types of learning activities. This will differ per exercise type, per intensity, and the subject you're learning.

Some Anki-like alternatives that people may like and not have seen (mostly focused on learning languages):

- Memrise - https://www.memrise.com/ - Has a very extensive selection of decks for language learning

- Supermemo - https://www.supermemo.com/en - Has a (disputed) claim that its algorithm beats Anki, various pre-made decks

- Drops - https://languagedrops.com/ - Mobile-only, has a higher reliance on pictures and audio over written translations.

There is also SuperMemo 18 (https://super-memo.com/supermemo18.html) which is from the same people as supermemo.com except it's a desktop app with way more features but also a much higher learning curve.

If you use SM and you use anki both for a significant amount of time it's not hard to see that SM leaves you with way less reps. Anki is based on the very first version of the SuperMemo algorithm (from ~1990 or so which was the first SRS algorithm ever) known as SM2. Current version of SM uses SM18.

Too bad SM is Windows-only and desktop-only. The beauty of Anki is that you can review anywhere.

On linux you can use SM via vmware/wine and on os x it's quite usable with parallels. If you're using SRS just for language learning then being able to use on mobile can be helpful but I think for those cases supermemo.com can be a decent alternative. More generally, I think the platform specificness of SuperMemo is more than made up for with its incremental reading functionality (alongside better scheduling algorithm)

> On linux you can use SM via vmware/wine

To add to this, there's a set of Winetricks recipes for running SM on Linux.[0] (Disclaimer: I haven't tried them, though I've been planning to, for a while.)

[0] https://github.com/alessivs/supermemo-wine

You might also try my website (and soon to be app) Seedlang: http://www.seedlang.com. It is a site for learning German, and later this year Spanish, French, and English. It uses video flashcards to create different learning experiences, and ties them all together with a Anki-like review deck where any word or sentence encountered on the site can be saved.

Isn't Anki based on Supermemo?

yep, it's based off the first supermemo algorithm, SM2

>Imagine this:

>You’re an 18 year old with just a high school degree. You immigrate to a new country that speaks a different language, and start work with some of the brightest engineers in the world.

>Soon after, you’re thrust into management. Now, you’re leading teams of people who are 10 or 20 years older than you, working on one of the fastest growing internet companies of the last decade.

I can't bring myself to. It's too sweet a daydream, a fantasy that not a soul on this planet would have deigned to make reality for a black boy like myself, credentials be damned.

His story is extremely similar to my second boss - a Nigerian man who was my CTO.

It happens. Where I work, the manager of programmers is from Nigeria.

if you think that's a race thing you're mistaken.

It's a mistake to assume opportunity is limited by race in this field and credentials aren't as useful as they used to be.

> “Whether you think you can, or you think you can't, you're right.”

Henry Ford


>>> It’s nothing crazy. Here’s an example from a very real non-work day:

Followed by a pretty long list.

How does he do ? There are 23 items on the list. None of which include daily tasks such as : dishes, sport, talking to family, to friends, eating, cooking, washing clothes, having a social life. All of these tasks take a lot of time, esp. if I consider the time to switch between them (which can be quite long since, hey, it's a non work day, so I'm slower...).

We're not equals.

Do do the same thing this guy and do all the things you mentioned. I create cards on sundays only and i do anki every morning before i start working. It can be accomplished easily, sans children.

but that list is >8hrs of back to back chores (according to his estimates) for a non-work day without any of the ones your parent post mentions. I don't think that's realistic.

I imagine that the subject of the article is one of those genetic mutants who only requires 4-6 hours of sleep a day. They do exist, but are definitely statistical outliers.

Not needing 9 hours of sleep is the only real way to add time to your day.

Not only that, I also require "offline" time to relax otherwise I become very weird emotionally and I start dreading everything.

i missed that and i now agree.

For Japanese learners, I built two iOS apps that cater to the special needs of the language.

Manabi [0] is a flashcard app with the same algorithm as Anki but nicer UI.

Manabi Reader [1] collects a bunch of short-form reading materials, lets you tap words to look them up, and tap to add a flashcard. It tracks the individual words you read and charts your progress word by word and kanji by kanji. This app has gotten pretty popular so I have been improving it substantially.

[0] https://manabi.io

[1] https://reader.manabi.io/

That looks really interesting, I'll definitely give this a try. How did you decide on the reading materials to use? There's obviously so much to choose from, I'm wondering what criteria you might use.

Let me know what you think!

The app is a (very heavily) dressed-up RSS reader. I maintain a list of reading sources and add them so long as they're not going to have overly sensitive material (eg someone suggested an anonymous blog with short and colorful posts from contributors reflecting on their lives, but some of the posts talk about self-harm) or be too niche. I'm most interested in feeds that get regularly updated with new content, or where they have a trove of existing content. Also always looking out for content that's good for absolute beginners.

Some RSS feeds require additional work in the app to transform their content to make them work nicely with the reader mode, so I also take that into account.

Ah, bummer that Reader is Apple only. It looks really awesome!

I will be bringing it to macOS desktop at least!

I will check these out! I use Anki for Japanese right now - mainly because I found Anki decks for each chapter of Genki 1 which I am working through right now.

Would love any feedback! The Reader app is really unique - no other app provides that kind of kanji-by-kanji reading progress functionality (probably because it took a tremendous amount of work under the hood to get it right).

This looks really neat. Do you have any plans to make it available on Android or a web app?

It's a native iOS app so I would have to start from scratch on Android. I do plan to port it to macOS via Catalyst sometime after the upcoming WWDC which will help some Android users.

I'll note that this is a big pain point for me: although Android users don't pay as much for apps, they help boost word-of-mouth marketing substantially. I see some similar but (tbh) subpar apps get more word-of-mouth online even in recommendations to iOS users just because of the huge signal boost from Android users.

This will probably get down-voted by the hacker news crowd. Information aggregation from diverse sources to construct a comprehensive knowledge graph in our heads and hence further and deepen our understanding of this world and the context in which things happen can be a goal. But at some point, the methods we use to accomplish them can take more importance than the actual goal. The method is not the goal.

This is similar to how there are many rituals in every religion that people might follow without understanding the underlying concepts. The rituals are not the religion but we still see people mindlessly doing things that their ancestors did.

I would much prefer that we build systems that automatically create that knowledge graph objectively (like the wikipedia) and layer an AI system on top that will create new connections, even with brute force, that our minds in a lifetime cannot. This will not stop us from having opinions. People will still come to different conclusions based on the same underlying data but atleast the data and the surrounding context will be available to all.

The point of Anki is to learn while making the cards, because making them takes mental effort and requires enough comprehension of the concepts. It would be harder and less motivating to use something premade, that maybe gives priority to concepts that don't really interest you.

Anki (and spaced repetition) is awesome! Some things I've had used it for to great success:

1. Memorizing Japanese vocabulary, pitch-accent, and basic grammar rules. It would probably be useful, at least to some degree, for any language like this.

2. Any certification which required strict memorization. All the basic ComptTIA certs were like this, and the CCNA:R&S cert (unfortunately) required memorizing commands and their syntax.

3. Verses in the Bible (though this is pretty basic in comparison, just Address <-> Text).

In all honestly I think one of the best things you can do if you need to memorize something is make the flash cards yourself, whether with Anki, another app, or even just index cards. This forces you to think about what it is that you're trying to memorize and phrase the text of the cards in a way that you understand. After that, reviewing is just kind of "maintenance" in my opinion. There was definitely a marked difference in my retention when I was using pre-made Anki decks vs. creating them myself. It is also easier to create cards using the desktop app.

It's a side note, but I also don't agree with Anki's pricing model. The app is free on Android[1] but $25 on iOS[2]. I think I heard (I don't have a source) that the developer's justification was that they needed to make money from the all the time and effort they spent creating Anki, plus hosting costs, etc - so why not do a cheaper price on both Android/iOS, or do a free-to-download app with a subscription model? For what it's worth you can use the web version on iOS but the app is a better experience IMO.

[1] https://play.google.com/store/apps/details?id=com.ichi2.anki

[2] https://apps.apple.com/us/app/ankimobile-flashcards/id373493...

> I also don't agree with Anki's pricing model. The app is free on Android[1] but $25 on iOS[2]. I think I heard (I don't have a source) that the developer's justification was that they needed to make money from the all the time and effort they spent creating Anki, plus hosting costs, etc - so why not do a cheaper price on both Android/iOS, or do a free-to-download app with a subscription model

People with money overwhelmingly buy iPhones. $25 is a pittance for anyone who can afford an iPhone. Anybody who uses Anki seriously gets far, far more value than $25 out of it. I used it for well over 200 hours before I stopped and I know I’ll go back to it again.

The Android version is not maintained by the developer of Anki. The iPhone app is. He chooses to charge for the iPhone app, which enables him to make a living making tens of thousands of people’s lives better.

If you don’t want to spend the price of two pizzas on an app that the modal user will use for over a hundred hours don’t.

> People with money overwhelmingly buy iPhones. $25 is a pittance for anyone who can afford an iPhone. Anybody who uses Anki seriously gets far, far more value than $25 out of it. I used it for well over 200 hours before I stopped and I know I’ll go back to it again.

I used Anki's web interface from my phone for about 6 months before I bought the iOS app. Then about 4 years after using the app almost every day, I sent them another $25 donation. I got way more than $50 worth of value out of it over the 10 years that I used it.

(I've now written my own study tool for Chinese which fixes some of the issues with using flashcards for language learning. Maybe at some point it will show up on "Show HackerNews"; but it's slow going when you've only got a few hours a week.)

I agree with you - I definitely got more than $25 of value from the iOS app and was happy to pay the price. I guess my thing is that I wonder if the developer would be better off trying to solicit more users at a lower price (or a small, monthly/annual subscription) than a one-time purchase.

Fwiw, two pizzas in US buy you up to 20 meals in a developing country. In a developing country, people often get cheaper iPhones (second hand or previous generation). The apps, however cost the same.

(co-author of the article) I wrote a longer list of categories [1] a while ago (some of which aren't mentioned in the article), here's the list:

  * Basic information about countries e.g. population
  * Ingredients and dishes from restaurant menus I didn't know
  * Important people and places
  * History facts (typically from Kindle highlights)
  * Conversions between units (e.g. lbs to kg)
  * Season for various vegetables and fruits
  * Keyboard shortcuts for vim, readline, etc.
  * Learning words and terms I don't know from Kindle/Instapaper highlights
  * Useful statistics
[1]: https://news.ycombinator.com/item?id=17847031

Seems people are taking issue with my gripe about the pricing. To be clear:

I bought the iOS app, was happy to do so, and got a lot more than $25 of value from it. In my opinion, the developer might be better off with a subscription model rather than a one-time purchase.

Developing for iOS is a lot more expensive than Android - you need to pay for Apple hardware and there are ongoing fees for "notarizing" your app. It's quite normal that an iOS app would ultimately cost more.

I agree, Anki on iOS is significantly underpriced relative to the value it provides ;)

If I were to guess, it's because Android offers many cheaper phones, so the platform is a proxy for the wealth of the user.

I use it for interview prep. I know some people advise against memorizing stuff in programming but it's been effective for me so far. You can spend 3-5 minutes in an interview trying to reason about implementing the partition method of a quicksort or translating 2D array indexes into 1d. With anki you can just regurgitate it out of your deep memory in 30 seconds whether you use it in your day to day or not.

Do you have an example of how you structured your cards for such a question? It seems impossibly hard to remember an algorithm without committing it to muscle memory so to speak.

Would an intervier be more likely to pass you if you were clearly recalling a memorized answer? That sounds so worthless... If I was interviewing I would switch to a different question.

In a world where you are expected to solve 2 leetcode medium-hard questions in 30 minutes with no syntax errors and test cases.... I doubt it. The questions come in different forms with optimizations, it's not about memorizing answers it's "pattern matching" and being efficient in implementation.

A partition method would be part of a larger implementation you still have to reason through and talk through. You just remember that the quicksort of the top half starts at pivot+1 instead of pivot, etc. Essentially you eliminate the little mistakes people tend to make when they are nervous

can you share your interview decks?


A bit of a self-post, but I've compiled a list of Anki tips I've learned and found. My deck is about 14,000 cards over the last few years - so I've learned a bit from my own mistakes :)

Using Anki is less important than keeping relevant and digested notes. I use Anki daily but I don't really need to memorize everything. It's enough to be able to fast retrieve previous collected info. In the end I just don't want to repeat myself. For that I simply use one searchable text file (and a github public repository - to maintain some quality - for digested notes/cheatsheets).

I sometimes wonder if tools like Anki aren't just another form of procrastination. The time invested could probably be better spent.

Really? I have a rule with Anki where the question is one line and the answer is no more than 3. No cards take longer than a minute to write, or more than 10 seconds to remember.

If it's anything longer, then it's probably not a subject you should be using anki to learn or you're trying to remember too large a bloc of information at once.

I have absolutely no idea how this could be a form of procrastination, or how I could more efficiently memorize information that must be memorized.

> I have absolutely no idea ... how I could more efficiently memorize information that must be memorized.

Outside of a schooling/testing context, memorization alone is rarely beneficial. You need to be able to practically apply the knowledge, which happens through experience; looking details up when you need them forms a natural kind of spaced repetition anyway that’s tuned to what you actually need to know.

I don’t agree. I was for a long time of the school that understanding is the only things that matters, and the details can always be looked up. Then during my post doc ( physics) I worked with a couple of Russian guys who had been forced to memorise a lot. What they gained was an enormous speed. They could try six different ideas in their mind while I still looked up the details necessary to try my first. So don’t underestimate root learning.

There’s a third category here, and that’s functional skill. It’s quite a separate thing from either understanding or rote learning. They all feed into each other, of course, but the point I tried (and apparently failed) to make is that it’s functional skill that’s useful in real-world contexts.

Sometimes the thing holding you back from improvement is a lack of facts, and sometimes it’s poor understanding of theory. In my experience, however, the fastest way to get better at doing something is almost always to practice doing that thing; most of the time a sufficient collection of facts and theoretical understanding will come along as a side-effect of that work.

How much of their prior training was memorizing specific facts, and how much of it was drilling the mechanics of solving typical problems? The former is what most people use spaced repetition for and what I believe is of limited utility; the latter is incredibly valuable and I never meant to imply otherwise.

Well, as a counterpoint, imagine how much time could be saved if you were able to recall from memory not just the most-used functions that you need for a problem, but the next level down of sometimes or rarely used ones. Or for patterns, or for other such things that can be simplified down into memorizable / recallable blocks.

Regularly looking things up is ok, but actively re-experiencing the thing you're trying to learn is a better way to make it stick.

In many cases less time that what I spend looking it up. I regularly look up things I haven't needed before, a few minutes of reading and I know it. Many of those things are something I expect to never need again in my lifetime. The few seconds to memorize all those things is greater than the time saved. Particularly since I don't know what I will need next week and so I'll be spending a lot of time learning things I turn out to never need.

> actively re-experiencing the thing you're trying to learn is a better way to make it stick

On this we agree, but to me this means practical application in various contexts, rather than call-and-response memory drills.

For programming, memorisation doesn't matter that much. Not knowing the argument order for a function isn't a big deal. You use an IDE or you get a good offline docs viewer (Dash, DevDocs etc.) or you learn to Google efficiently.

Language learning is an obvious use case.

Also: law and medicine. Having knowledge mentally 'to hand' is pretty important if you've got a patient under general anaesthetic, or a judge asking you a very difficult question.

> Language learning is an obvious use case

Memorizing word pairs can certainly be done with spaced repetition, but it’s unclear how much that translates to actual language ability. Second-language acquisition appears to be primarily dependent on reading (or listening to) the target language for content, and most words are learned via seeing them in context instead of being looked up in a dictionary.

I have no experience with law or medicine, but I expect the story is similar: practical knowledge is what you need to hand, and not book knowledge. Book knowledge is what gets you through the exams and into the practical part of your training.

That depends entirely on how you implement spaced repetition.

Word pair is the easiest way to do it in anki, but it's not the only way to implement it.

Beyond the very basics necessary to extract some meaning from a second language, I’d be shocked if time invested in spaced-repetition drills of any design had better returns than reading the target language for pleasure.

Anki has two benefits: memorization and understanding. Memorization is what comes from going over the cards, which is what you allude to. Understanding comes from the process of creating the cards, where you think hard about the atoms of information in the text, and their interrelationships, and cast them into simple questions. Then memorization gives you a much deeper understanding than if you hadn't ankied the text.

I concur, what you describe is a very effective use of spaced repetition. It is different from what ‘ProstetnicJeltz described, where cards should take less than a minute to write — that describes memorization without first putting in the effort to understand.

Writing the information should take less than a minute once you formulated what you want to write. Learning takes more time.

It is worth considering the cost, yes.

My time isn't so optimised that 15 minutes each day on Anki is otherwise going to be spent on deep work.

http://augmentingcognition.com/ltm.html This writeup estimates his usage of Anki is 4-7 minutes of review time for each card over 20 years. So, if it's worth taking 10 minutes to remember a fact, it's worth putting in Anki. (This writeup also explains it's not worth putting things you don't care about into Anki).

Anki seems well suited for domains where you need to quickly access facts from a wide domain, e.g. vocabulary for language learning.

It's harder to see as important for programming. If you're programming you'll be able to recall the things you have experience using.

I use it for programming. My general rule is anytime i need to keep looking up some command, i just stick it in anki. Here are two python cards I added this week:

What method do you use to format a datetime object in python:


How do you create a new virtual env using python 3

python3 -m venv <NAME>

Those may seem stupid to you, but I was looking them up and now i do not need to. I also use them for other concepts when I want to know a language better. For example

What are the three prototypical methods of a js promise?

then, catch, finally

What does the constructor of a promise take?

and executor function

How many parameters does a executor take?


What are the two parameter of the executor function (constructor of a js promise)

resolve() and reject()

Things like this just help me as a programmer. I agree it's not for everyone, but I have the time an enjoying doing it.

Those are much closer to what I'd put in if I added Anki cards for programming than other things I've seen online. Nice and short.

Why not just `new Promise((res, rej) => ..)` instead of intermediate facts like "an executor takes 2 arguments"? I don't think I'd be able to recall what's meant by 'an executor' without first recalling that snippet.

I noticed from using Anki so much for everything that when I include answers in other questions titles I learn the material better. Your comment would obviously be faster because it is less cards to add, but I think my brain would be more likely to forget the concept that way. Personal preference I guess.

For math/programming, it's better to keep it short in my experience. But for other things, like knowledge from books, I use both methods, here are some examples:

---- What happen to slaves in Africa after the slave trade was abolished by the America/UK in the early + mid 1800s?

simply led to a redeployment of the slaves, who were now used within Africa rather than in the Americas

“So the abolition of the slave trade, rather than making slavery in Africa wither away, simply led to a redeployment of the slaves, who were now used within Africa rather than in the Americas. Moreover, many of the political institutions the slave trade had wrought in the previous two centuries were unaltered and patterns of behavior persisted. For example, in Nigeria in the 1820s and ’30s the once-great Oyo Kingdom collapsed. It was undermined by civil wars and the rise of the Yoruba city-states, such as Illorin and Ibadan, that were directly involved in the slave trade, to its south. In the 1830s, the capital of Oyo was sacked, and after that the Yoruba cities contested power with Dahomey for regional dominance. They fought an almost continuous series of wars in the first half of the century, which generated a massive supply of slaves. Along with this went the normal rounds of kidnapping and condemnation by oracles and smaller-scale raiding. Kidnapping was such a problem in some parts of Nigeria that parents would not let their children play outside for fear they would be taken and sold into slavery.”

Excerpt From: Daron Acemoglu. “Why Nations Fail.” iBooks.


What are two reasons why large scale wars no longer occur ?

1. Price of war has gone up because of Atomic weapons. 2. Weath is no longer is physical goods (gold,etc) but in the minds of the citizen (Silicon Valley).

Scholars have sought to explain this happy development in more books and articles than you would ever want to read yourself, and they have identified several contributing factors. First and foremost, the price of war has gone up dramatically. The Nobel Peace Prize to end all peace prizes should have been given to Robert Oppenheimer and his fellow architects of the atomic bomb. Nuclear weapons have turned war between superpowers into collective suicide, and made it impossible to seek world domination by force of arms.

Secondly, while the price of war soared, its profits declined. For most of history, polities could enrich themselves by looting or annexing enemy territories. Most wealth consisted of material things like fields, cattle, slaves and gold, so it was easy to loot it or occupy it. Today, wealth consists mainly of human capital and organizational know-how. Consequently it is difficult to carry it off or conquer it by military force.

Harari, Yuval Noah. Sapiens (p. 372). Harper. Kindle Edition.


Basically, I create cards with short answers but provide the context (usually the paragraph I found it in). This method works extremely well for remember everything you read. I could write a more technical blog post on this if people actually care and find this interesting.

That depends on what knowledge you're trying to retain. When I majored in Spanish in college I was extremely efficient compared to my classmates. Whereas most would memorize lists of words and sentences and dedicate tens of hours to it, I would average less than 5 hours of preparation time per exam.

>> When I majored in Spanish in college I was extremely efficient compared to my classmates.

Did you use spaced rep?

Yes. After class, I would review my notes and add new items. Reviewing for exams amounted to those 5 hours average, spread over 3-4 weeks.

I tried to use Anki to study for a tech certification.

In the end, I got a lot more out of doing the research to create the cards than I did from the actually looking at the cards. I might even go so far as to say that after creating the cards, they were practically worthless.

This would be less true if the exam was further from the creation of the cards, though. It was within a month.

Someone else's cards might have helped a little, but not nearly as much.

>This would be less true if the exam was further from the creation of the cards, though. It was within a month.

Anki is basically meant for longer term studying. The criteria for when a card matures in Anki is when the interval becomes 21+ days.

I think that many people could think that fair assessment, but really it's just about what you value in the world.

Is becoming more knowledgable just "procrastination?" Is reading non-fiction outside your specialty procrastination?

I feel like for much of human history we've held people who have read an entire library's-worth of books in high esteem. Were they just procrastinating?

I still see where you're coming from. I used to think that people who meditated, or hiked the Appalachian Trail were "selfish," because they were doing something that benefitted only them. And I still find it hard to understand someone who decides to learn Mandarin at the age of 93. But people should be allowed to enlighten and actualize themselves however they wish, and it seems absurd to think that a planet full of more knowledgable people would be a bad thing.

It depends what you're doing. I mainly use it for learning vocabulary.

I'm learning German and Anki is nothing short of amazing. I'm easily able to learn ~20 words a day primarily just practicing on my phone during bathroom breaks.

On the other hand I'm also learning Russian and Anki just doesn't work. Words won't stick in my head the way they do with German.

I've heard very numerous accounts of it being great for Japanese.

>On the other hand I'm also learning Russian and Anki just doesn't work. Words won't stick in my head the way they do with German.

Because (for English speakers) Russian is harder than German. You need to reduce the number of words learned per day. Japanese is even harder to learn, but you have to be persistent. I spent more than an hour every single day for several years straight to learn the vocabulary.

Also there's a unique challenge with Russian is that nouns, verbs and adjectives are all conjugated, and the rules are quite complex. The noun cases would be completely foreign to non-Slavic language speaker. Even if you learn the vocabulary, using it in sentences is not trivial at all.

That's a good idea, I'll give it a try with fewer words. I did notice that I'd just come back to a word I'd seen 30 seconds ago and not recognize it, maybe I need to reduce the number of switches in between so it has more opportunity to stick.

> The noun cases would be completely foreign to non-Slavic language speaker.

Picking those up was actually fairly straightforward. Cases may be completely foreign but the grammatical concepts behind them aren't. For example Russian has nominative, accusative and dative but English has subjects, direct objects and indirect objects and I learned all those in school.

I think somewhat ironically, the biggest grammatical problem I had with Russian was the tenses. English has 12 and Russian has ~5 so it was hard to figure out how to express a given English sentence in Russian.

But I think the biggest problem is phoenetics. On the surface it looks simple (unstressed o -> a) but there are so many extra rules if you want to speak properly (e.g. voiced consonant before a voiceless consonant becomes voiceless).

Just going to be a bit nitpicky here. Noun declension is a feature of various non-slavic languages, including Latin which English speakers have a decent chance of having been exposed to.

Modern Russian has 6 cases (with a couple more that pop up very rarely) with 3 genders and singular/plural endings. It's a total of 36 possible combinations, which is not that bad. Masculine and neuter also share most endings so there aren't actually 36 unique ones

Probably the most difficult thing about cases when I was learning Russian was remembering which verbs took a different case than I would expect, e.g. dative instead of accusative for what seems like a direct object.

Right: but one difference (which you sort of hint at) is that German is consistent in why something takes a particular case compared to Slavic languages (e.g. direction vs location), and at least the preposition is consistent. Is it "nad morze" or "nad morzem"? No way to know without the entire sentence. And "morzem" is the "tool case" -- the form you usually use to describe that something happens with the help of a tool even though the sea (morze) is not at all a tool in that sentence. Oh, and you only use "nad" with bodies of water by the way. Sorry!

Another incredibly tricky one for Slavic languages is imperfective vs perfective aspects. In most other languages, perfect vs imperfect is just a standard construction. In Polish (and Russian, though my examples here are Polish), you change the verb itself. How? Well, sometimes you put "po" in front of it, like rozmawiać / porozmawiać. Sometimes it's "z" (jeść / zjeść), sometimes "u", "na" or "wy". And sometimes you just give up, like oglądać (but obejrzeć in the perfective), widzieć (zobaczyć), mowić (powiedzieć).

(For this reason I have found it very hard to progress in Polish without conversation with native speakers who aren't too polite to correct me.)

Yes, perfective and imperfective aspects are very tricky, particularly how they interact with various other linguistic features: imperatives (and the negative imperative), the subjunctive, and verbs of motion come to mind, each of which modifies the use of the aspects in it's own way.

> Modern Russian has 6 cases (with a couple more that pop up very rarely) with 3 genders and singular/plural endings. It's a total of 36 possible combinations, which is not that bad.

36 isn't that bad but Russian also has 253 irregular verbs [0], each of which adds another set of combinations. A lot of them are just small tweaks and there are some that follow a set of patterns (e.g. идти with its various prefixes) but it still adds a lot of overhead.

[0]: https://en.wiktionary.org/wiki/Category:Russian_irregular_ve...

Irregular verbs are a conjugation problem, which is different from noun declension.

>I've heard very numerous accounts of it being great for Japanese.

There's at least great pre-made Japanese vocab decks. I'm about two months into one. Maybe less so for kanji (I at least don't have one), so that's probably something you need to do outside.

Saves me a ton of time in medical school vs trying to brute force the facts in. Plus it keeps things I learned over a year ago fresh so it has been very helpful for preparing for USMLE Step 1. Medicine might be an ideal case for this kind of learning though.

If I don’t absolutely hate doing it then I know I’m not learning.

I'm starting out on a similar Zettelkasten style database/wiki thing using org-roam and org-journal on Emacs. I find the entire concept fascinating and I'm surprised at how enthusiastic I am with this subject.

I'm still refining the workflows and discovering how it interacts with my own idiosyncrasies. The basic setup is:

1) Take daily notes in a hardback A4 sized square dotted journal. Mark shorter tasks/notes as completed when done.

2) Sweep "stuff that needs to be remembered" into org-journal daily entries, once or twice a day. Also add interesting web page links with a short description. And "nice to have" ideas, things like that.

3) Link from those daily notes to other planning or project org-mode notes using org-roam. Write quick outlines of documents.

4) At the beginning of the doc, those planning/project org docs have links to various single-subject docs. Those links are considered "tags" so a taxonomy is slowly being built up.

Org-roam shows you backlinks from docs, daily notes, etc, so you can begin to see what is connected. There's also a graphviz output you can generate to get a more graphical view of the links.

Things I'd like to be able to do, but haven't yet coded or learned the Lisp to do so:

1) Sync/Import bookmarks and their tags from Firefox. I've got a lot of bookmarks that I've tagged, it would be nice to be able to link those 2 sets of tags.

2) Figure out some sort of end-of-week-review process or multi-doc view. Something for me to look at all the stuff I added that week, and a way for me to push stuff into next week.

I used a lot of space repetition previously, but I've found a lot of issues with it

- Most of what I learned is not useful knowledge in anki. I made many variations of different decks, some for programming and some for other concepts

- Space repetition - the concept itself is built in many different places. For instance, google will remind me of photos that I took last year on my phone. I can check up instagram and facebook to see things my friends are doing and be reminded of things we've done. I can message people and be reminded of things too, or throw parties and likewise be hit with space-repetition if done frequently enough.

don't get me wrong though, space repetition is useful especially for learning a topic that requires a lot of vocabulary. The best way to learn something is just to be fully immersed into it, if you practice it daily you basically are applying the space-repetition algorithm in practice.

If you want to learn a new language, move to the country that speaks said language.

There's a lot of wasted time and effort in making cards and determining what is "Fringe value" - cards that expire and offer no values long term

If you want to remember things about people you care about, just message them pictures of things you;ve done together. No need to make a rolodex CRM like some of my friends do, its a waste of effort and time imo.

What anki really shines though again is language learning, or learning something with a lot of barrier-to-entry vocab. This is especically true for medical industry based applications

I practice what I use everyday and immerse myself in many programming communities and I've found that I've learned and retained information far more effectively when it's shared b/w many different perspectices and contextes

Reminds me of the open-source app Polar[1], which allows you to create Anki flashcards while you read.

[1] https://getpolarized.io

Is Anki a trademark? If someone wants to make a tool called AnkiSuper, is that allowed? I'm not a big fan of the UI/UX. So I am thinking of scratching an itch.

Is the format open source?

Here is a description of the format: https://github.com/ankidroid/Anki-Android/wiki/Database-Stru...

If you'd ask me it's a perfect example of how not to structure a database. It's not even in the first normal form.

I'd love to create an Anki clone that has better UI/UX and a better database format and I think that would remedy some of the biggest pain points about Anki. But since most Anki users (me included) use several plugins that are quite essential to their Anki experience, an Anki clone would have to support those plugins, too.

The desktop app is GPL and it's all backed by a Sqlite db so it's possible to reverse engineer.

That requires a lot more effort, and it’s error prone.

You don’t want to get it wrong because you may corrupt people’s data

Ad "Zettelkasten" and how you "converse" with it (that's how I [mis]understood it):

Luhmann started out by doing these things:

* He had sequ. numbered cards, that were written on until they were full and therefore could contain multiple topics - and one topic could therefore also be on multiple cards (no. 1,7,15), if he later added to it

* He had a topic index to map card numbers to topics

* Most important: he used this system with slight variations continuously over years and it was a one person system

The benefit (that you can get with a personal wiki too, if you follow the system with Plugins) is, that when you get all cards on a topic, you also get the information on the topic you wrote down right before and after thinking about the topic you are researching...

As information organisation and knowledge is a personal thing, retrieving info from the cards with the other topics you thought of right before and after, might lead to subconscious connections that

- and here is the "conversation" -

then lead to your brain making connections that go beyond what you wrote down on the cards... at least that is how Luhmann described it... could anyone follow my little mental jump session?

Is someone here running an Anki sync server[1]?

I would be keen to know about the experience of setting it up and maintaining it. The github project doesn't seem active, which I always find fishy for server applications.

[1] https://github.com/ankicommunity/anki-sync-server

Has anyone ever created an anki deck from their college coursework that would have allowed them to pass verbal exams 10 years later?

Depends on the course, the exam as well as how long you used spaced repetition. The next time Anki will ask me what "å sløse" means is in over five years and I'm pretty sure I won't have forgotten it by then. For exams that rely less on memorization this might be different. I'm pretty sure that even if I remember all definitions from my measure theory course I still wouldn't be able to pass the exam in 10 years.

Spaced practice and retrieval practice, both of which are used in SRS, are two learning techniques for which there is ample evidence that they actually work [0].

Of course, You still need to decide what is worth remembering.

[0] https://www.learningscientists.org/

I'm using Anki for language learning sparingly (I know it defeats the purpose, but still...) What I haven't figured out it how to automatically create Anki cards from the words I'm looking up in a dictionary. I'm at a point where I rarely stumble upon words I don't understand, and they are usually rare enough to not meet again for a long time, long enough to completely forget them.

What I want is something that could catch me opening dictionary by double-tapping in CoolReader and add the shown entry to my Anki. Doing it manually breaks the flow and immersion so I can't really do it while reading.

Does your dictionary have a search history? That's how I add new cards. Every day before reviewing, I check the list of terms I looked up and then create cards for those I think will be useful to remember.

My problem with Anki was years ago and was how hard it was to programmatically create/modify decks. The database was a pain and not documented anywhere I could find. Took some work and just digging through it the hard way. I just wanted to take a deck and lookup one bit in a dictionary and add a new field for that looked up bit to a card all automagically. Not an easy thing to do. Unless you want to learn the GUI bits, and maybe not even then. It was easy enough to create a new deck, but hard to bulk add new information to an existing deck that you've been studying.

Reminds of when I was starting to learn Japanese! https://github.com/miyamoto-san/hiragana

Does anyone know if Anki or a similar app can handle importing from a text document or spreadsheet?

I have a ton of things I need to learn for my job, and I already have a text document full of notes and definitions. I'd like to assemble everything on my PC and then have flashcards on my phone. I also need to control the syncing because I'd be making cards with the company secret sauce and can't sync to anything but our corporate google drive/docs account.

Having an app that pulls the cards from a google sheets doc would be awesome.

The Anki app does indeed import from CSV files, and you can choose what delimiter character it will use to separate fields.

I recently tried instead of remembering stuff with anki, trying to get to the root cause / why of what I try to remember, it's usually a tree meaning the outcomes of what I used to enter to anki was way too many things, and the causes of these items were much fewer, once I get the causes of what I try to enter to anki, I can derive the anki cards just by deducing them. Not perfect but better than having hundreds of anki cards to remember.

One anti-feature makes me go away every time I return to give Anki another try: lack of a "dismiss once and forever" button. Many decks I'd like to study contain too many cards I already know. Yet I can only choose how easy it was so it would decide how often to repeat it, it won't let me choose "it's easier than saying my own name, don't repeat it ever".

I use the suspend card (or note) command for this. Press @ when your reviewing and it will suspend the card and you won't see it again.

It's been a while since I used Anki, but I'd suggest marking or suspending the notes you don't want to ever see again. Then you can keep them suspended, or if you marked them, you can delete them completely. Should do the trick?

It's a little hacky, but you could just suspend the card

Why would you want to remember everything you read? That sounds like overtraining.

It's not clear that managing and drilling on so much information offers a significant improvement over what is likely the most important intervention mentioned: reading widely and with purpose.

Sure flashcards can help you remember things you need to memorize. But it's not clear they should become a way of life and an end in their own.

> It's not clear that managing and drilling on so much information offers a significant improvement over what is likely the most important intervention mentioned: reading widely and with purpose.

I think you're in agreement with the original article. The individual is reading widely and with purpose. He uses Anki to assist in remembering. And that's the general guidance with Anki (or spaced repetition generally): learn first, recall later. You read, develop an understanding (at least partial), build your deck, and study with the deck while continuing to read the later material and building the deck up further.

And you don't try to remember everything, again from the article:

> When I come across something in a book that seems useful to understand and remember, I will highlight it on my Kindle and add a note to it with the text: “.flash”.

He doesn't highlight everything, just things that seem important. The importance is that many things are outside your particular field of work. But you have to interface with people who work in those domains (or want to), or you just want to have more knowledge to draw on. If you aren't using the information frequently, it won't come to mind when you need it. So with Anki you can have a relatively lightweight way to keep the jargon of another field re-callable, or to have the history of a place or people at hand.

I wrote a quick script today to use Airtable and Python Flask to do a simple Spaced Repetition. Need to add the following fields in Airtable: Front, Back, Scores (long text), NextReviewAt (date with GMT time).


I would better endeavor to digest the content and let it be forgotten than to spend time organize it and remember it repeatedly.

Generally speaking, I don't believe reading more books helps your thinking better than going back to smaller amount of books, unless you just need the knowledge to be used for now. Memorizing is not good enough when reading a good book; it requires more time thinking than reading to grasp the essence of a good book.

Anki is rather old, SuperMemo in latest versions is generally better. They're related.

The unanswered question here is why and how to make these methods of training automatic, as in that you do not need a tool or a calendar or a papet database to run them. There are some other mnemonic methods available but efficacy is unproven.

And in 2020 I stil struggle to export my Kindle highlights from sideloaded books (not bought from Amazon). And even if I did, I haven't found a note system that lets me centralize all the notes I gather from web, documents and my Kindle.

Maybe one of you can come up with a piece of service and make me tear up a little.

Have you tried https://readwise.io? I export my kindle highlights from sideloaded books through an iphone app and send them to import@readwise.io. That's better then downloading and uploading the txt file. Web is also covered. Dunno about your docuemnts though.

I had the problem where it felt too time consuming making the cards and I just wanted to start off learning a few cards at a time, so I made my own solution. Its free https://www.flashcardmax.com/download

A paper notebook would be so much worse. You would have to carry it with you physically (which is clearly an unconscionable imposition), it can't show you relevant ads, and it cannot send every glyph you scrawl to Google, Microsoft, and the NSA. How could people ever use tech like that?

For a companion post, there was an Ask HN post yesterday [1] asking people why they don't use spaced repetition.

[1] https://news.ycombinator.com/item?id=22484060

Does anybody know how to get the LaTeX stuff working on MacOS? I've been wanting to memorize a number of mathematical formulae, but I haven't been able to find clear step-by-step instructions for building LaTeX cards.

Do the instructions here not work for you? https://apps.ankiweb.net/docs/manual.html#latex-support

If you have Latex installed on your machine and have used it before, the process of getting starting should be pretty straightforward.

Why not use a picture of the formula ?

I get a feeling that these little jewels of knowledge that we come across get store in our subconscious. And that my brain will pick it up when I need it, although I can't list them all if asked to. Maybe I'm wrong :)

I’ve used Anki for months, trying to learn the Greek alphabet. I’ve failed. I don’t think I’ve learned more than one or two of its letters.

Is it possible I’m just not compatible with the way this works? Any tips on alternatives?

My hunch is that Anki is mostly useful for stopping you from forgetting things you have already learned. The actual learning is better done outside of Anki.

Agreed - I've found that Anki really helps to recall underlying facts, but if I didn't grasp a concept well outside of Anki, it's not helping with A-HA moments.

Using it to learn an alphabet effectively probably requires more upfront work (setting up and improving the cards) than you’d expect. It’s not just 1 letter = 1 card.

You want to memorise the entire alphabet as a sequence not just as individual letters. So you need cards of little overlapping sequences, like ABC, BCD, etc. The supermemo essays cover this scenario. With that work in place you’d learn it quickly.

As just separate letters divorced of any meaning, not hooked up with any other neurons in your brain, you’d struggle to ever recall it quickly.

This is a detail that people overlook and then wonder why it’s not working. (Not picking on you for doing that: I did the same before eventually reading way too much about it)

The supermemo article is here:


...by the creator of supermemo who basically formulated the whole idea.

In Point 1 he explains you shouldn’t use spaced repetition to learn. Use it to remember.

In Point 10 he covers how to remember “enumerations” (with an alphabet example)

You have to go back to the classics when you want to do something properly.

Spaced repetition and the learning curve is quite a bit older than supermemo.

Yeh, I hear ya. That’s why I used the word “formulated” rather than “discovered” —- because he really did the work to not just model it, but really implement a system with a very specific formula which he devised and which was the basis of the field we see today.

I think you'd be able to learn the Greek alphabet in at most an hour, just by writing it out and saying it repeatedly. Making it stick long term would be a matter of a couple of minutes a day for a week or so, then every few days. It's small and simple enough that flash cards aren't necessary.

For more involved facts, Anki seems great in principle, but in practice I've found it hard to create and maintain the habit of doing the cards.

How are you using it?

Anki is "just" a tool to prompt you with flash cards. Designing them is up to you and extremely flexible. You're free to add text, images, audio, whatever you want.

Some ideas of how you could use Anki to learn the Greek alphabet:

- Front of the card has the upper- and lower-case glyphs and the back has their sounds, either English, IPA or an actual recording (Wikipedia has them for everything in IPA and the Greek Alphabet page has a mapping from letter to IPA). You should also do this in reverse.

- Front of the card has a Greek word with a missing space, back of the card has the complete word. You can use audio recordings here as well if it helps you (e.g. have the front read the entire word).

There are also a few pre-made decks that people have shared: https://ankiweb.net/shared/decks/greek%20alphabet

Maybe Anki isn't a tool for you. I personally don't use Anki because it bores me to death. For me it needs to be interesting. I think there are other people out there who even don't use Anki because it's a boring way to learn something.

Spaced repetition really clicked for me when I realized that it isn’t about learning, but about keeping memories fresh. If you don’t have an emotional attachment to each card in the deck, reviews become a terrible slog. Mine is a mix of random insights I’ve had, photos of places I’ve been, summaries of things I’ve learned, etc. The questions are simply there to force me to engage my memory, which drags up the whole web of mental associations— there’s no need to add every detail explicitly; that’s what notes are for.

Yes yes, I agree. The way I tackle the threat of detachment from my cards is to aggressively delete cards as soon as I see boredom creeping in. I switched from Anki to Supermemo, though. Much better for managing context and even keeping organized notes that may or may not be used for active recall at some later point.

Can't reply to child for some reason but the reason that SM is so great for managing content is a feature called incremental reading (https://supermemo.guru/wiki/Incremental_reading). To sum it poorly, it lets you go through hundreds of articles in parallel and slowly convert them to active recall material over time. It sounds terribly complicated in practice, and learning it is indeed a pain, but once you know it there's really no way of learning more enjoyable. I can promise you that.

I don't understand how anki users are able to make cards for non-language things,I personally absolutely hate making cards while reading anything. It's hard to explain the exact process but incremental reading by contrast makes it much easier to take what's really important from an article and make cards from it.

Same. After about a year of using Anki the need to aggressively delete became apparent. With good culling and time to reflect, review can be almost a meditation.

How does Supermemo help you keep keep context?

> keeping organized notes that may or may not be used for active recall

I use a separate notes app (Bear) for this. Is there much advantage in your experience to integrating active recall notes with Evernote-style reference notes?

Absolutely. Keeping it interesting should be one of the first priorities when learning something.

For the record, Anki actually works really well for me. But interest in the subject should come first.

I've tried several times and I agree, it's boring to me. But a lot of people use it daily and see real progress.

If you want to learn the Greek alphabet try leaving Modern or Ancient Greek on Duolingo. If there’s an Anki deck with modern Greek, translation and romanization that would almost certainly work better than trying to associate a sound to an arbitrary symbol with no scaffolding to help it stick.

Writing systems are something you need to sit down and spend at least a few minutes at a time writing out.

Anki is better for inputting vocabulary with definitions and example sentences. That also puts the letters that you spent time writing into practice.

When I went to Greece I learned it in a few days just by trying to read all the signs I could see. I guess the trick is not to learn isolated letters but learn to read common words.

I, too, did this and it’s not effective. More effective is learning Greek words and memorizing the collation order of the alphabet.

Try to just learn one letter per day. Don’t work ahead and don’t try to learn more than one every day.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact