Nicholas Nassim Taleb describes his education in Antifragile. It sounds much like CS Lewis's, as described in Surprised By Joy. He read as he wished from a young age, according to no program other than what he set for himself. Adults did not supervise or direct his free reading, so he would sometimes be exposed to things they might have found inappropriate had they known. (Neither Taleb or Lewis notes whether this is inappropriateness is violence, sexual, or abstract concepts beyond their likely understanding.) He attended school and did as well as he needed to at each point but found it unexciting and not very valuable. His father was a more devoted school student and won national awards in Lebanon, but did not insist that his son take the same route.
I have noted a few times that most geniuses seem to have been autodidacts. Taleb is clearly smarter than I am and his devotion to reading exceeded mine. I had also a devotion to socialising - perhaps that's it. Yet I understand what he is talking about. I used to say "never let school get in the way of your education" when I was in highschool and college. That was certainly founded in arrogance and a sort of laziness about what others thought I should be doing, but there was truth in it as well. I knew an enormous amount about a few subjects I had never taken a class in. In fact, there were no classes in Arthurian studies, baseball history and statistics, or history of the English language beyond an introductory. I added things over the years, taking up a topic with fanaticism for a few months or a few years, revisiting it years later. In between, I would pick up other stray bits without even trying. Once you have a structure, a schema, about a subject, anything that drifts past gets put in storage pretty easily. I call it "cuphooks." Once you have cupboards, and have put cuphooks in them, you can add in information at will. This is why I believe that teaching higher-level thinking skills to students too early is just bosh. Until they have a supply of cupboards and cuphooks, the information just lies around on the counter or on the table, useless.
I have therefore always loved books which supply a schema, even if it has to be modified later. Joel Garreau's Nine Nations of North America and David Hackett Fischer's Albion's Seed supplied cupboards and cuphooks. I have redesigned those cupboards since then, but they were extremely useful.
My children also have topics in which they are extremely knowledgeable. These are significantly independent of anything they learned at school.
I noticed early on that a temporary devotion to a subject in the humanities could make you an expert very quickly. I would talk about an historical or literary topic with a person who taught it at the college level and sense immediately that they did not know what I did. Not close, in some cases. There's a lot of literature, a lot of philosophy, a lot of history out there and no one knows more than a sampling. If you pick an obscure corner, you can get right up to the top quickly. The general tools that the professionals bring to tie things in together and give overview don't impress me. In theory, such things should be useful. However, they seem to be largely fashions, and these days, an imposed set of sociopolitical ideas.
I have also described in more than one place that experts and talented amateurs make different sorts of errors.
Enter Taleb, widely read in literature, philosophy, and mathematics. He suggests we avoid talking to middle people - which includes most college professors and popular writers - as we can learn little from them. He recommends attending only to the most informed, least fashionable, and most original thinkers on any subject. Beyond that, you are better off spending your time with auto mechanics, waiters, or others who do not rely on book knowledge to get by. I see the sense of this immediately, especially in the context of how he defines antifragile knowledge. I do have an immediate difficulty with the concept, however, because I am one of those middle people, and so are most of my friends. NNT notes that the middle people, middlebrow, can be very entertaining people to be with. They are conversant at a minor level with a hundred subjects. They know who Plato was and a few facts about him, so you can reference Plato in conversation with them in order to make a point quickly.
E D Hirsch made a name for himself in the 1980's by advocating a "cultural literacy" as a shared common knowledge to enable communication. I have thought until quite recently that this is an excellent way to go. I now see it is restricted to the middlebrows. It has strength because the current alternative is an assemblage of correct attitudes, even more a product of politics and fashion. Yet after reading Taleb I see weaknesses. It has no depth, so that no subject need be understood well. More importantly, it provides considerable inertia for the conventional wisdom. It is much harder to find a spot on which to stand to say "Socrates is wrong, and so is Plato," as Taleb does. As an autodidact, he references other ancient philosophers who he believes come closer to important truths - and I have never heard of most of them. They aren't part of the cultural literacy lists, neither the old canons, nor the new politically correct ones.
He disparages the type of close-ended knowledge that tests measure, with single correct answers according to forms. I have disparaged them myself (which is one reason I preferred Ron Hoeflin's Mega Test to other IQ tests), but not too loudly or too often, because they are the sort of thing I'm good at and I have an interest in keeping my own roads to status clear. Yet I emphatically see his point. His examples are pretty persuasive, and pretty damning, as well.
More to say, but I think I'll leave this here for the moment.