Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Learn Exponentially

saveall.ai

Post reply on HN
Print view View original

101–110 of 154 posts

Re: Learn Exponentially

#101

Post by AlchemistCamp »

detuneattune wrote: post #95
Qualadore wrote: post #83

Earlier quoted context omitted.

They're a lot better than Save All's builtin decks, such as only including 1,000 of the most frequent words of a language, and only nouns, and not even including the word's gender. Rarely with these SRS services do you see actual high-quality decks that outdo public Anki decks, which is a shame because it would be a great way to add value.

Instead of there being competing spaced repetition programs and services, I'd much rather companies just go down the route of making well-curated, frequently updating Anki decks and putting them behind a paywall instead. Refold, a company focused on language learning, does exactly this [0], and having tried their JP1K deck for Japanese for a while, I can say without any hesitation that it was shocking just how high q…

Fluent Forever started with this approach: https://fluent-forever.com/shop

Re: Learn Exponentially

#103

Post by teddyh »

motohagiography wrote: post #32

Spaced repetition turns reading into a physical activity with feedback. If you had an activity that required the knowledge you were hoping to acquire, you would learn at the same rate or faster by practice. Most lessons are encoded really poorly. I think spaced repetition is great, but practice at something is better. A hacker is just someone who has practiced learning independently and has become exceptionally good…

“One will weave the canvas; another will fell a tree by the light of his ax. Yet another will forge nails, and there will be others who observe the stars to learn how to navigate. And yet all will be as one. Building a boat isn’t about weaving canvas, forging nails, or reading the sky. It’s about giving a shared taste for the sea, by the light of which you will see nothing contradictory but rather a community of love.”

— Antoine de Saint-Exupéry, Citadelle, 1948

Re: Learn Exponentially

#104

Post by gratitoad »

motohagiography wrote: post #32

Spaced repetition turns reading into a physical activity with feedback. If you had an activity that required the knowledge you were hoping to acquire, you would learn at the same rate or faster by practice. Most lessons are encoded really poorly. I think spaced repetition is great, but practice at something is better. A hacker is just someone who has practiced learning independently and has become exceptionally good…

Reading this comment had me cheering “yes, a thousand times yes” but then it occurred to me how often I turn to an O’Reilly book instead of a concrete project and I wondered why.

I think I subconsciously avoid new projects to learn something programming-related because of how often they end up unbounded in time due to one of the fundamental attributes of writing software: unknown unknowns. The kinds of sticking points in a project that could take you minutes to figure out, or days, or even weeks.

I have limited free time and I want only so much of it to be _more_ programming. (I’m a professional software engineer) If I can’t be certain I can timebox a new pet project with _some_ degree of successful outcome, I think I’m low key anxious enough to avoid starting it.

Does anyone else struggle with this? How do folks overcome it and work on new things without them taking over?

Re: Learn Exponentially

#105

Post by darkteflon »

allenu wrote: post #66
anon2020dot00 wrote: post #6

My idea is that there is a market for community-curated spaced repetition decks. Many people want to learn the same things such as a foreign language or a programming language. The difficult part is creating a deck and crafting the answers and questions. Because usually this is a time-consuming process. So if it was a community-effort then it would be a win-win. This is probably not an original idea and if anyone kno…

> This is probably not an original idea and if anyone knows already where to find such decks, that would be cool. This is something I've wanted to do with my app (Fresh Cards). I ended up defining a simple text file format for the flashcards[1] to help make it easier to share and import cards. You could post flashcards as simple text that someone could drag and drop into the app to import. (Formats like Anki's .apkg…

I hadn’t heard of Fresh Cards before, but this looks lovely.

I’ve used Anki over the years to great effect with second language learning and am fully bought into the paradigm, but I do find Anki quite clunky as a piece of software. Fresh Cards seems like it’s designed specifically in response to that.

Does this use a similar repetition algorithm to Anki? Are there any obvious limitations versus Anki? What sort of UX differences would a long-term Anki user need to get comfortable with?

Would also be keen to hear from anyone here that’s used this.

Re: Learn Exponentially

#106

Post by carapace »

p-christ wrote: post #57
carapace wrote: post #30

Earlier quoted context omitted.

It's a good piece, and excellent as far as it goes. You should investigate hypnosis. E.g. you can just remember things, w/o spaced repetition or anything, just transfer a piece of knowledge directly to long-term memory immediately. The tricky part is second-order effects. Once you move naturally-autonomic subsystems into voluntary control you now have responsibility for them. The result is a kind of change-of-being l…

Very interesting, how would i learn about something like that? Is there a resource you'd recommend?

A couple of books: "TRANCE-formations" by Bandler and Grinder, and "Monsters and Magical Sticks" by Steven Heller, both are written for the therapeutic context but the material is universal.

Re: Learn Exponentially

#107

Post by allenu »

darkteflon wrote: post #105
allenu wrote: post #66

Earlier quoted context omitted.

> This is probably not an original idea and if anyone knows already where to find such decks, that would be cool. This is something I've wanted to do with my app (Fresh Cards). I ended up defining a simple text file format for the flashcards[1] to help make it easier to share and import cards. You could post flashcards as simple text that someone could drag and drop into the app to import. (Formats like Anki's .apkg…

I hadn’t heard of Fresh Cards before, but this looks lovely. I’ve used Anki over the years to great effect with second language learning and am fully bought into the paradigm, but I do find Anki quite clunky as a piece of software. Fresh Cards seems like it’s designed specifically in response to that. Does this use a similar repetition algorithm to Anki? Are there any obvious limitations versus Anki? What sort of UX…

Thanks for the response. I read up on how Anki's algorithm works and tried to match it closely, although Fresh Cards doesn't have the exact same algorithm. I actually ended up adding a way for you to provide your own scheduling algorithm via javascript if that's your sort of thing. I also built a web page just so you can see the effects of recalling correctly and incorrectly, as well as reviewing too early or too late: https://www.freshcardsapp.com/srs/simulator/

I also included my attempt at converting Anki's algorithm to javascript so you can compare and contrast. I'm open to any feedback on improving the algo that I do use!

In terms of limitations, probably the biggest one is that Anki lets you create cards using a template system, which Fresh Cards doesn't have yet. In Anki you can create one note and then from that have multiple flashcards auto-generated using several templates. You can also do something similar with cloze deletions. One of my goals is to get to 100% Anki compatibility over the next few months, so I'll be adding a template system and similar cloze-deletion system.

In terms of UX, the app actually has multiple review styles. When I first built the app, a lot of users actually wanted to re-review cards, but the app would strictly schedule them, so you didn't have a way to do that. (Anki lets you "review ahead", but for the most part tries to keep you on the schedule.)

To satisfy both users who wanted to review whatever they wanted, whenever they wanted and users who wanted to stick to a spaced repetition system, I designed the UX to act as a "playlist". You just hit play and it'll start reviewing the cards in the order that you see. (You can sort and filter based on spaced repetition algorithm, or use several other modes.)

The app is pretty much designed to just pick up and go without a lot of instruction, so it should be easy to get used to.

Re: Learn Exponentially

#108

Post by atahanacar »

redelbee wrote: post #18

I think a better title might be “Memorize Exponentially” because that seems to be the true gist of the article. There are undoubtedly many areas in which memorization is useful. I tend to use memorization as a second-order tool, in the sense that it is only useful to memorize once I’ve learned that memorization would be necessary. I memorize combinations to locks I unlock frequently. I memorize names of items I sell…

>There are undoubtedly many areas in which memorization is useful. I tend to use memorization as a second-order tool, in the sense that it is only useful to memorize once I’ve learned that memorization would be necessary.

For me (in medicine), it is the exact opposite. Memorization comes first, then you start to actually understand and learn things. Everything is so intermingled and there is no "learn this first to understand that". In addition, you can't start practicing things before memorizing them.

I feel like it's the same thing with the mnemonics. It's useful at first, but not very practical or efficient. However, as you use it more, you actually learn and stop using the mnemonic.

Re: Learn Exponentially

#109

Post by atahanacar »

Buttons840 wrote: post #76

There are a few e-ink note taking devices. They pride themselves on feeling like paper, but they're far more expensive than a notebook and offer trade-offs rather than clear advantages. For example, is using a text search on an e-ink device better than knowing "I wrote this down in the first 1/3rd of the notebook on the page with a coffee stain". Maybe? Maybe not. It's a trade-off. A physical notebook and an e-notebo…

Have you tried image occlusion for Anki? It does exactly what you mentioned: you occlude parts of an image (your notes in this case) and creates cards based on that. It's more tedious than an integrated solution would be, but I'm sure there are ways to make the card creation process more seamless. Maybe an Android e-ink where you can click "Share" on your notes, and choose the image occlusion "plugin app".

Re: Learn Exponentially

#110

Post by Denzel »

woojoo666 wrote: post #99

Title is misleading. This isn't learning exponentially. Anybody who has used a spaced repetition app (like Anki) knows that you generally learn around the same number of words every day. You're not going to be learning 10 words in a day and then one week later, learn 1000 words in a day. This article talks about how the reps needed to learn one piece of information, reduces exponentially over time. You might need 1 r…

The rate of repetitions scales sublinearly with respect to time.

Consider a uniform repetition approach that repeats each item once per day. The rate of our uniform repetition system scales linearly with respect to time. If one item takes 1 second to review, then I will lose 1 second per day. Therefore, N items causes me to lose N seconds per day. Even in the case where I don’t add items to my list, I lose N seconds per day.

In contrast, as you stated, the amount of time we lose per item per day reduces exponentially with spaced repetition, however. If we don’t add any items to our list, the review time asymptotically approaches 0 seconds/day.

Unrelated to all of this, I agree the title may be a touch embellished. :)

Post reply on HN