I disagree, if you are just doing you won’t be learning new things. I work at a place where there a lot of of long-timers are there and at that time there were only people that can from different fields. Those people work since 10, 20 years without even mastering what they do.

If they have to do something new they either give up or just try to come up with something on their own. Most of the time people have done things like that before and you can just read about them or look at their code.

You're right! I'd say doing is important for actually understanding and retaining information while reading is important for understanding the triumphs and failures of others and "stealing" ideas.

Interspersing the two is perhaps the best way of learning. For example you don't want to learn to swim by watching 12, hour-long videos on swimming techniques, costumes and its history. You do wanna watch a 15 minute instructions video on how to get your feet wet and then go ahead and actually get your feet wet and then come back and watch the next 15 minutes of video on how to step into the pool.

Strongly agree. I would always fall asleep in class in college.

Once I got into the workforce I would avoid this learning gap by creating side projects or taking extra work which was aimed at learning specific new things.

If I sat through talks or presentations I would just information dump into a text file which I would data mine later. The taking of the notes helped some with retention, but organizing it according to my own thought process helped make it more navigable later when I was stuck on something.

Works pretty well for me in my career. YMMV if you try the same thing.

I'm inclined to agree. There's a profound difference between how I remember something I've heard and how I remember something I've done - even if it's something I've heard through spaced repetition.

It really depends on who you are.

I learned Indonesian more or less from reading an excellent grammar book about twenty times over years.

And then I went there, and to my shock, I was able to speak it.

The high point was when I had an argument with people in the airport about the amount of airport tax and got them to concede they'd done the exchange rate right, all in Indonesian.

I would not recommend that way of learning language to anyone, but I just love grammar books, and I had a lot of spare time those years.

But I do agree with your point. Heck, I probably would have learned faster with a workbook and exercises.

I agree, but I can't exactly build a reactor, a SARS variant, or the holocene.

The first two are probably possible if legally dubious.

Yeah it's not the same trial-and-error as, say, making a basic Rails web app.

Semantic memory isn't procedural memory or episodic memory.

Just listening to something is completely different than being able to do something with what you just listened to.

Learning involves more than just recalling things from memory.

If you mean time spent, then yes, about 95% of time spent learning is doing. But you're talking about the cost, not the benefit. Learning by doing is very inefficient. You gain much deeper knowledge, but it can only possibly be in a very narrow area. If you spend only 5% of your time reading/listening/watching stuff, this more than pays for itself. You don't need to have experience programming a network stack before knowledge of it becomes useful. For almost everyone out there, knowledge is enough.

I’m not a fan of this dichotomy - you assume the two are mutually exclusive. There's analysis paralysis (no “doing”) on one extreme and blind ambition on the other (no book knowledge). I doubt many people advocate either.

Why can’t you build things while setting aside time to learn from other peoples mistakes?

I'd tend to agree with you and have perhaps leaned in a bit too hard on the doing for emphasis.

Not quite sure. Many things clicked while I was either idle or doing something completely unrelated.

But then applying that insight and playing with it in various dimensions is what built an intuition.

And that is the trouble with the headline. If watching at 3x speed gets you to actually trying something faster, that is clearly the way to go.

Totally trial and fail