People have differences in how they learn. Some can learn effectively through reading, others learn effectively through doing.
By doing but not reading, you have the potential to gain a more fundamental ground-up knowledge base about why certain approaches work better.
By reading but not doing, you have the potential to gain direct insight into others pre-established knowledge which means you can jump ahead without having a more fundamental knowledge _first_.
I prefer learning by doing, but I think both have valid merits and one is not better than the other. They are just different approaches to the same ends.
Hacking is about solving a series of specific problems. Reading will give you direct insight into another's pre-established knowledge, but it's probably not relevant, and it may not be correct.
Code kata in my opinion is the superior way to learn. Code through practice and repetition, just like every other skill.
It might be terrible advice for someone who is ignorant on a lot of worldly topics -- but if we take the demography of HN, literate, well-educated individuals who already have probably read many dozens to hundreds of books, it's not terrible advice. A lot of the books on that list are 'self-help' books, you read them, they make you feel good, you achieve basically nothing. It's just feel-good filler.
I'm _assuming_ that's what the parent comment meant by 'read less, do more'. It's not 'read nothing, do more'.
This should be read as a kind of "contrarian advice" i.e. advice designed not to be taken too literally but to counteract common wisdom that is often accepted uncritically. Common wisdom is of course "you should read more".
Personally I like reading very much and some time ago I started reading a lot of non-fiction of the kind listed in the article. It is a nice feeling when you like to do something and common wisdom encourages you to do more of it. But then I noticed that I don't get much out of these books other than a modicum of entertainment and some superficial erudition. So I stopped wasting my time. Now if I just want to read, I read fiction - it is more entertaining and honest.
In my opinion proper reading of a non-fiction book is a slow process when you try to really understand the ideas, criticize them and apply them in real life and it really blurs the line between reading and doing. A good rule of thumb is that some artifact (at the very least it may be some book notes) should appear as a result of this process. But it is very easy to fall back to "just reading" - reading a science textbook but not doing the problems, reading a software engineering book without writing any code, reading a business book without operating a business. If you do that, all you are doing is reading a fairly boring fiction book.
Grothendieck (the mathematician) was at some point giving lessons at a place which had no library. Someone pointed out that there were not many books to read there, so he replied: "Here, we don't read books, we write them."
I've not heard that quote before, but I'm going to guess it was more an apology for the poor state of the library than a serious suggestion about how to achieve what Grothendieck achieved.
I never thought it could be an apology (it's kind of a material detail, they could still find books elsewhere), but rather a mix of sincerity (at least regarding his creativity), stimulation and joking.
it's terrible advise for those who spend all their time doing and never reading. People who just stick to either reading OR doing will soon hit a point where progress is limited. Depth is impossible without a constant combination of both.
By doing but not reading, you have the potential to gain a more fundamental ground-up knowledge base about why certain approaches work better.
By reading but not doing, you have the potential to gain direct insight into others pre-established knowledge which means you can jump ahead without having a more fundamental knowledge _first_.
I prefer learning by doing, but I think both have valid merits and one is not better than the other. They are just different approaches to the same ends.