To adress the various weird issues:
We have different childhood memories. I spent a large part of my childhood reading books for fun. As a result, I had huge vocabulary and knew and understood tons of stuff not that usual for children, and I'd say it was a good base for any further learning. While other kids struggled learning the ortograph and grammar, and other subjects considered general knowledge, I was ok. I'd say that tops the authors memory "my grandpa was actively teaching me tons of stuff that slipped my mind".
The main problem is finding time for this. So, if your problem is lack of time, then perhaps looking for such "optimalisation" by drilling instead of reading is a good thing. But I think I'm a good living proof (and by far not the other one), that the bookworm's path definitely brings many benefits.
To that Russian learner's example: B1 after two years isn't bad. She managed to learn despite the method. And it is possible, that she actually reached her own goals, as she might have needed primarily to read. And if not, this was her fault, that she hadn't waved goodbye that teacher. I know lots of such examples, and many do things that would probably fall into the "active" category, as it is probably imagined in this article, yet they are totally useless. Guess what, doing just one activity doesn't work well.
.....
If at any point while reading this description, you told yourself, “Hey, this is pretty much how my lessons look like!” then run. Run the hell away from your teacher or language school. A visit to a local strip-club seems to be a better investment than this. At least you will know what you pay for.
Yeah, I'd agree. Most teachers are much less worth the money than most strippers. The stripper won't ever tell their public that not liking their performance is the customer's fault. Or say "this is the modern way to do striptease, dancing and taking off clothes is too old school and useless, just get excited without these elements". I could find many other such differences.
Self-study rules!
.....
The learning activity pyramid is a nice theory. But there are some catches, when it is applied to language learning.
For example: too much "practice doing" without "lecture" and "demonstration" leads to what I like to think of as "fluent neanderthals". And "Teach others", that is definitely a great activity, but should learners primarily focus on teaching others? Thinking of it, I actually know people getting paid for this. For teaching languages while being worse than some of their students.
.............
One minute of talking is worth 5-7 minutes of reading/listening (read more about the benefits of talking to yourself).
I'm all for discussing benefits of self talk. But it is not one or the other. If the one minute of talking is supported (and preceded) by 5-7 minutes of listening, I'd say we're getting somewhere. Really, there are tons of people with learned mistakes because of having just spoken and spoken, without actually listening much.
Unintentional learning takes place when you acquire vocabulary accidentally. It is a by-product of repeating a given piece of information a certain number of times.
It’s worth mentioning that is it also one of the default and (most useless) strategies of almost every language learner.
The body of research shows that you need to repeat a piece of information (unintentionally) between 20 and 50 times to put it into your long-term memory. 20 to 50 times! (one of many sources)
It takes way too much time. And time is the luxury a few of us can afford. Of course, One might argue that 20-50 repetitions are not that many. After all, if you read extensively and listen, you should get to this number of repetitions after some time.
Well, getting from B2 to C2 in three years (after two years of almost complete break) by almost exclusively reading and listening, that doesn't feel so bad But sure, when you show me lots of people who managed to do the same thing just by SRS and expensive classes (=people who tried and managed to stick to it without burning out), I'll even call my way inferior. But so far, it looks like every successful learner I meet has learnt a lot unintentionally. Of course combination of intentional and unintentional learning is the best, sure.
But condemning unintentional learning just because extremely few people are able to do the same thing faster through other means, that just doesn't make sense.
.....
Unless you learn three thousand words, reading is a very slow and inefficient activity.
And until you reach this number, your odds of learning words contextually are slight. Sure, you can infer the meaning, but there is a good chance that your guess will be incorrect.
And what about rare words which you might find useful?
What If I need to know the word “thimble” because that was my dog’s name, and I feel the need to share it with English speakers? How many thousands pages must I read to stumble across this word, say, ten times? Hell, I don’t remember when was the last time I heard this word in my native tongue!
What about other words like tangs, udder, piston, and so on? I need such words frequently during interpreting or teaching. Relying only on passive learning activities would make me an inefficient teacher/coach/interpreter.
So there you have it. L2 Learners are simply at a disadvantage as for the number of repetitions of words. If you want to optimize your language learning, limiting passive learning activities is one of the first things you should do.
I agree with a part of this statement. And if the author clearly targeted just beginners with the article, I'd agree a bit more. Those people trying to just listen to songs, some youtube, and avoid any coursebook, those are unlikely to get anywhere. But once you get to the 3000 (I'd even say 2000), get reading. And listening.
Btw 3000 is still not much, those words up to 5000 are still not rare. And the truth is, that if you don't encounter the word "thimble" in a few thousand pages, you actually don't need it that much.
If you're a translator, than by all means, learn every word with an exact translation right away, don't waste time on 20000 pages of books (that's the number I like to recommend intermediate learners as a good goal). But not everybody learns a language to be a translator. For most people, learning "thimble" a few months later, or not at all, changes little. And anybody in need of the word can learn it through other means.
I would love to agree with the last statement in this quote so much. It is such a truth that the rarer words that you still might want to learn are harder to learn through exposure, due to being presented less often. But it is also true, that those words (I'd say those between 5000 and 20000 on a frequency list) are much less represented even in resources for learners. Just saying.
Limiting "passive learning activities" is actually the last thing you should do. If a word appears twice per thousand pages, then reading less won't make you encounter it more. You should read more, if you've got the time. And let's not forget that real reading and listening teaches you much more than just vocabulary.
Instead of limiting the "passive" ways, increase the "active" ones. I mean it. It will help a lot to increase the exposure to the rarer words. But don't get rid of the "passive" stuff. I know lots of learners, who have drilled tons of words, by SRS or otherwise. They'll regurgitate them very efficiently and precisely. But they simply don't speak well. In this context, reading and listening actually does the "demonstration" and "practice doing" role, from the already mentioned pyramide. The same is true about another popular "active" activity, using the teachers. They learn well the content of the classes, but they are usually simply lost outside of it, because the classes are simply much less affordable and also much less realistic than a good tv series.