Scott discusses the “T” model of learning (where the stem on the T represents learning something in-depth) and self-education:
“In my experience, self-education tends to be very good at high-level ideas. [...] But I’ve noticed that the typical approach to self-education tends to be lousy at the deep, detailed knowledge of a field. [...] Let’s say, for example, you want the major focus of your learning efforts to be computer programming. [...] If [self-learning] is your approach, then the necessity to deeply educate yourself is even greater. Quick overviews of topics without understanding mathematics, operating systems or computer architecture won’t make you an expert.”
Unfortunately, this isn’t a good example. Many, many programmers don’t need much of an understanding of operating systems or computer architecture. I rarely use math beyond about grade 10. 90%+ of what is taught in the course of a typical computer science major (B.Sc.) will not be used after those courses are done. This is a huge waste of time and effort if the point is to make programs.
The best way to learn programming is to start with a simple, useful project. Learn what you need to know for that project. Then move to the next one. And so on. Awareness of coding possibilities for ways to solve a problem is important, but that doesn’t require in-depth learning like Scott is talking about. Rather, it involves cursory learning of large areas, exactly what Scott says is a strength of self-education.
Consider his discussion of the importance of curriculum:
“The reason acquiring deeper knowledge is difficult, is that the further you stare down the microscope, the less relevant it appears to the big picture. This is often why so many students lose motivation at school. Just how is understanding integrals, polymorphism or the ATP-cycle important for my life?”
But that’s just the thing. ATP cycles, polymorphism, and integrals are useless to the larger majority of people. This is one of the major weaknesses of universities - they pile information into people, where 90%+ of the time it will be useless (outside of future courses). They pile it on because, if you want to become a Professor and teach those concepts, you need to know them. This is largely circular. For practical use of knowledge, the university approach is wrongheaded.
Universities are largely a Faustian bargain: give us 4 of the best years of your life and x thousand dollars, and we will give you a degree which confers upon you status (i.e., will help you to get a job).
If you want to create or learn useful things, universities are, by and large, not the place to do that. For computer programming, for example, a diploma from a technical institute is usually going to entail more useful learning than a B.Sc. from a prestigious university. It is only because universities tend to attract smart people that, in this case, go on to be good at coding, that people get the impression that the universities are creating those abilities in people - the reality is almost the opposite.
If you try to ’self-educate’ by importing the university model, then that is the problem, not that self-education requires more discipline and so is more difficult. Once you step outside of something like university education, you can jettison the university learning model and actually focus on learning what makes sense to learn for whatever purpose you have.
Scott comes close to saying as much later on:
“In formal schooling, actually applying the ideas is a far goal. When you first learn statistics, most professors don’t expect you to start doing your own sampling or analysis. The actual use of the knowledge is put in a backseat to passing tests.
But if you’re going to sustain the motivation to complete a deep self-education curriculum, application must be put first. Otherwise, it is too easy to lose sight of the big picture and stop learning.
Effort needs to be made not just to learn the ideas, but to start applying them immediately. When I was previously teaching myself computer programming, I would always have a project I wanted to use the new-found skills on.”
I think Scott is mixing one thing up here. A rational goal isn’t some abstract “education.” What is education for? It is for using the ideas to create things that are useful, fun, or whatever.