Scheming An Embedded SRS

I am planning a very light custom flash card / spaced repetition system (SRS) embedded in my website.

Why?

• writing prompts is good for learning
• I’m already forgetting what I learned ~a month ago
• SRS dramatically reduces reviews needed

Why not use an existing SRS? All systems I found are at least one of:

1. Separateness — cards live outside the source material
2. Review-only — cards are not meant to be read on their own
3. Permanency — cards are assumed to be permanent recurring tasks

Factors 1 and 2 discourage in-context volitional learning, which I desire when motivation is high. Factor 3 is about the realities of SRS: you will fall off, and that’s OK, because real life is completely disconnected from whatever you’re learning with an SRS that you fall off of.

More details below.

The Problem With Spaced Repetition Systems#

Here is my thesis: Rather than treating falling off SRS as a problem of motivation to be addressed (I submit: impossible), what if instead we assume that one will fall off an SRS, and make it friendly to be abandoned?

I have years of experience using both Anki and WaniKani, and falling off of both multiple times as their systems crush me under the weight of reviews.

An SRS assumes that you wish to learn something forever, and so it will schedule you recurring reviews (tasks) indefinitely. If you ever fall behind, you are basically hosed, barring herculean effort. Getting back on board means confronting hundreds or thousands of reviews.

What does one do? They simply abandon the system. And when they do this, their life is completely fine.

This illustrates the disconnect between the SRS and reality. The reality is, if you fall off an SRS, you didn’t need to learn all of the stuff you put in there. If you really did (e.g., you’re a med student, and you’ve Anki-decked all your exam material) you will drag yourself back on because it’s vital.

For the rest of us, even using an SRS has already selected you as a somewhat obsessive, detailed type of person, who will add reams of volitional material. Learning a language. Memorizing geography. Maybe writing four hundred prompts about how to make chicken stock. While in the limit (as t → ∞) you would be rewarded with a huge vocabulary or an encyclopedic internal atlas, real life involves one t at a time. And I submit to you that an amazing number of experiences in life are worth discarding your SRS review habit and instead focusing on living.

And so, falling off an SRS is our brain’s way of telling us: hey, I think right now it’s OK if you didn’t keep memorizing all that stuff. SRS is, in a way, artificially hacking our brain to expose it to stimulus that it doesn’t get in the real world, from which it would organically learn. I think people have different amounts they’re willing to tolerate this artificial stimulation. It’s probably a bell curve (because why wouldn’t it be). Some folks are way off to the right, and can use an SRS until eternity. Folks way off on the left would never even think of using an SRS for a minute. For the rest of us, there’s some capacity we have to first enjoy, then tolerate, our brains being stimulated by this artificial source of pointed knowledge, the SRS, until our brain subconsciously says, screw it, I’m not doing this crap any more.

SRS has a unique fight with real life because it’s a piece of technology that you are supposed to really commit to using, day after day, never stopping or resting. Other technology doesn’t punish you if you take a break.⁰¹

Rather than treating falling off SRS as a problem of motivation to be addressed (I submit: impossible), what if instead we assume that one will fall off an SRS, and make it friendly to be abandoned?

Furthermore, what if we harness the initial period of learning, where someone is so motivated that they’d like to setup an SRS, and allow for deeper, in-context, chewier learning—engaging with material volitionally and luxuriously, rather than sporadically, in random bits of stimulation, aiming to clear a review queue?

These factors motivate flash cards which are:

1. Embedded in the article material. Hat tip Orbit. But unlike Orbit, they are:
2. Viewable when no review is needed
3. Abandoned-native — no pressure to ever go through an enormous queue

I don’t mind the emails (point 3) so much, but kind of the whole philosophy is wrapped up in the combinations of points 1 and 2.

This philosophy is: Let me read and engage with my cards more than I strictly need to during my higher motivated period of peak interest. Good, juicy problems are meant to be chewed on, interesting concepts refreshed, in-context connections made. For learning interesting things isn’t just producing a point on demand in exponentially-increasing time jumps, it’s sitting with a whiteboard or an annotated paper or a book, pondering, making notes. Let me make digital flash cards like physical ones: keep them in reach, able to be played with, shown to friends. “Let me review that just one more time.” Let me feel like the cards are there and mine.

Anyway that’s my manifesto, I guess. Time to write some JavaScript.