How to Learn (Almost) Anything | Hacker News Reader

How to Learn (Almost) Anything

litemind.com

92 pointscwan24 comments

Pro-tip for those still in university for a technical field:

Instead of going to lectures, completing assignments, and doing tutorials you would be better off learning engineering the way I learned. My way only works in those cases where your prof will actually assign a relevant book.

Step 1: Start at the very beginning of the book with some blank printer paper. Start reading, but be sure to use the paper to block off information further down the page. I actually used 2 sheets of paper, one for the opposite page and one for the section below what I was reading.

Step 2: When you encounter an example problem try to solve it. Make a good, honest, serious 10 minute try at figuring out what the formulas should be. This is especially important if you haven't even been given the formulas first. I "invented" the equations for electromechanical forces by thinking about what they were asking, visualizing it, and then applying basic calculus to the problem set. I also invented fluid flow through pipe networks, structural deflection, life time value assessments, and many other things. All of these concepts are stuck like granite in my mind for life.

Step 3: At exam time take all those assignments you haven't done and do them all at once in the 48 hours before the exam. If you were smart you would have asked your prof if he was ok with you handing in the completed assignments on your own schedule. The reason you leave these off to just before the exam is two fold. Firstly, you actually do the assignments without leaning on your friends to help you with the devilish details. Secondly, your mind doesn't go into auto-recall mode and remember THAT EXACT problem. It is actually trying to solve them for the first time.

That's all it takes. My university experience was a walk in the park once I figured this out, I actually held a full time job as a construction coordinator while fully enrolled at Waterloo.

On exams I could always get the bonus questions because I didn't just understand the formulas - I "invented" half of them. It's a much deeper understanding that started in first year and carried me through until the end of my fourth year. There are going to be exceptions (for me it was applied partial differential equations (except Laplace, that I understood fairly well)) but in general it is far easier to read through a 250 to 400 page book and cram for 48 hours than it is to trudge along going to class and getting stuck on stupid details in assignments.

This is a good tip. I do the same thing when reading math books. Trying to pick up advanced math on your own can be really challenging because there are almost no worked examples of, well, anything past the undergraduate level. The author supplies a series of exercises at the end of each chapter, no solutions, and you're expected to provide the rest. Fortunately every theorem that is actually presented can be viewed as an exercise that was too challenging to state without proof. Now every time I hit a theorem, lemma, etc. I do exactly what you describe. (The paper is really necessary, I'm incapable of not reading ahead and peeking at the solution, at which point the "problem" is ruined.) Often my solution is way off the book's, but this is the best/only way to "learn by doing" when it comes to higher math.

This is quite an interesting method, but be aware: it doesn't work for all people. Everyone has to figure out how they learn best, preferably by the start of high school.

In practice, maybe by the end of college. How can we help people figure figure how they learn, while they can benefit from it the most?

I truly learned how to learn by the end of college, and I would have found it invaluable to have learned it sooner. There is no doubt that I would have accomplished 100x more and learned 100x more during those years.

In my very first lecture in university the professor told the class that this is impossible. The only way to learn something is by doing. But not to despair, as even if you think you are no good at times, you will be 100x more productive by the end.

Very interesting, thanks!

How did you happen on this way of learning? It seems like one would need quite a bit of self-control and faith in this method to use it. Did you evolve it, did someone teach you, did you read it somewhere?

I was graduating the year Ontario was getting rid of OAC (grade 13) so there were twice as many graduates coming out of the high schools, even though the near by universities were only accepting 30% more applicants. Unfortunately (or fortunately, in retrospect) for me I was in the Grade 12 stream.

I took a detailed look at all the courses being offered and decided that I just wouldn't look at enough material in the grade 12 version of statistics, so I petitioned the school to allow me to take the grade 13 one. Of course there was a conflict with some other grade 12 course, so I was forced to learn it all from correspondence books. The books moved very quickly, and I found myself lost half the time until I finally told myself to stop skimming through the readings. I ended up applying to university with a 95 or something in the class.

In university I had forgotten that lesson until I was in the same mental space: lost and not knowing why I kept getting stuck. Then it dawned on me that I needed to actually read what in front of me and my method stemmed out of there. I think it helps that I'm one of those crazy visual thinkers, but most people can do it.

I prefer to sit in class, listen very carefully, do the assignment, but nothing else. I schedule my exams to immediately follow the classes. Works for me.

I guess there are different ways to learn.

You get to schedule your exams? Lucky you :)

I am a Mathematician. And all our exams are oral. You just have to find a professor with some time on her hands, and then you can take the exam.

that was good... thanks

Off-topic but related: If you are studying engineering or medicine or any field where there will be a significant penalty in money or harm to others of mistakes, DON'T settle for an A. Learn (practice) to get the right answer when working problems. In the real world there are no answers in the back of the book, and even when someone else is backstopping your work they can miss mistakes, too. Take the time when doing a problem to make sure you understand it and if it's complicated don't try to keep track of everything in your head.

My first employer out of college (The Instruction Set - RIP) exploited this by selling training courses that the technical staff would teach on rota:

- Per this article - staff quickly became experts in appropriate fields. There is nothing like explaining pointers,inheritance,device driver interrupts etc. to a confused (but motivated audience) to internalize it yourself.

- It was a great source of development work - one of the principles of the company was not hoarding knowledge - very often, teaching companies how to solve their problems would result in them saying something like "Wow, it really is hard, that's not our core business, can you help us" - and they would now have the knowledge to negotiate a sensible spec.

- It gave a sort of mini sabbatical to the tech. staff - every couple of months, you would be out teaching for a week or two, and dev. work would be scheduled around this.

The basic premise of the article is fairly obvious: to learn something teach it or do it. The best way by far to learn something is to teach it. When you have to explain to an audience the way something works you have a pretty good mental understanding yourself.

Interestingly, according to some studies, people remember 70% of the things that they say in speeches they give (at conferences, classes, conventions, etc.)

I agree with this; though, I do wonder how my Physics teacher knows the latitudes of various major cities...

Coastal cities? S/he may be a boater!

if you know the coastal cities' latitudes, then some basic geography will help one match up inside cities, too.

Maybe he used to work for a missile program? :)

My favorite example is this: Lets say you want to learn about a new Javascript library.

1. Spend a little bit of time reading about it, so you know about 20% of what you need to know, then convince a group to let you give a talk about using it.

2. Realize, "Oh crap, I know virtually nothing about this, I need to use it for something!" Build something with it and take notes as you go.

3. Your notes can be easily morphed into a presentation, which you give to the group on the assigned date.

The deadline forces you to take action, and the doing before the teaching makes you learn a lot of the material before doing the teaching, and then preparing for the final "teaching" part cements the final bits of knowledge you might have missed.

If you can't find a group to do a presentation, start a local one! Its really easy with the free tools available online now.

This works well with programming. I took two years of Java in high school (an intro course and an AP course) and now, as a HS senior, I probably couldn't declare an array for my life. However, I've self-taught myself python by programming a few GUI apps, and, even though having never taken a class on it, I'm infinitely more competent.

He touches lightly on it at the end, but this cone does not work for all people.

For example for me, HEAR and READ would be reversed.

And actually the rest are not accurate - for me - either.

The point isn't the details (I initially posted them, but edited it out) - the point is to figure out this cone for yourself, and work with it to best effect.

Found this on HN sometime back. Similar good read http://www.wired.com/medtech/health/magazine/16-05/ff_woznia...

I remember this discussion on HN. And I liked the idea. In prep school, I used to go through my whole Math binder for each bi-monthly test even though the tests were only on the last few chapters. It definitely helped me a lot.

After reading about the link you re-posted, I looked at some of the software that were replicating SuperMemo's method. (I can't remember the names right now) The issue I had was to find the material for what I was studying. Of course, the best way is probably to create your own material but that requires an extra time investment.

Reply on news.ycombinator.com