Duolingo doesn't work because that's the point.

Same problem as dating apps: if you could actually learn a language with Duolingo, then you would stop using Duolingo. No good for business.

The hard part is how to trick people into believing that it works or even "it's better than nothing". Hence gamification.

Yes. Also, they don't need much vocabulary, no grammar concerns, no reading/writing.

We much overestimate how well kids learn, and how "easy" is for them. Many kids have language difficulties, and they usually know, and they don't feel too great about it.

FSI estimates are taken from diplomatics who study full time in preparation for their job. Not a good approximation for most people.

This is a far better one: https://www.cambridge.org/elt/blog/wp-content/uploads/2018/1...

I am working in this space as well. Habit formation / self-learning is a big easy because overwhelming majority of people prefers more structured / guided content. This makes SRS very niche.

The part of seeing "You still need to decide what’s worth remembering" as friction I strongly disagree with.

That is a very important part of the learning process and IMO should not be automated. It is difficult because learning is difficult, but if you kill that you also kill the spirit of self-learning!

And not mentioned in the article, but I think the most important factor is that everyone in the SRS niche knows Anki, and despite Anki faults, everyone has put the effort to learn how to use it, got used to it, found a good-enough workflow, and have zero incentive to move to a more expensive alternative even if it is better.

Yes, thou the main problem stated in the article ("memorizing the rectangles") is very real. Now, I don't think you need much magic / LLM stuff to overcome that problem:

1. Instead of the usual "1 new word, 1 card" have 2 - 3 new words on each card, and have each new word in 3 - 4 different cards (ideally with different inflection, meaning, nuance, etc) 2. Review cards fast and very few times each. Like 5 - 8 times / card (max) instead of usual 15 - 20. 3. Don't punish yourself, keep moving on even if you just half-remember. First familiarize and then internalize language patterns instead of just memorize words.

Review intentionally, totally concentrated on the task. 10 mins / day well done >>> 30 mins / day mindlessly.

Outcome: more fun, more effective learning, no memorizing rectangles.

> What is "good for one's brain" (apart from proper nutrition and absence of concussions) is a strong education and healthy lifelong social interactions

Sure, but I think this is more about the fact that what you don't use, you lose. Learning languages is hard, so even learning old Greek keeps you brain sharp as long as you enjoy it to some extent.

Gym for the brain is good, what you do with it can be better, but gym is still good.

Yes. AI transcription is great, AI translation is OK (depending on language pair), but TTS is still pretty awful for most languages.

> nor is there ever going to be a 100% freedom-from-error

That is not a problem. Language is messy, you don't need 100% accuracy to learn. The problem is that LLM errors are fundamentally different from human errors, and you won't even know how to recognize them.

Your interlocutors can work around human errors, because they tend to follow the same patterns in same language. But they will freak out with LLM errors.

> Natural language acquisition is almost entirely listening and talking

Listening and reading. Talking goes last. See Steve Kaufmann, for example

Spaced repetition = information retrieval. You need the other part, encoding information. Best "trick" to encode information effectively is to "make it yours" in the best way you can. Develop, explore, follow rabbit holes, ask yourself questions about what you are learning, engage emotionally with the content, write down what you learn in your own words, etc.