Ask HN: How Do You Learn? I'm really interested in knowing what system you employ to learn stuff. Personally: I'd make a list of all the available resources to learn about that topic -> narrow it down to those that (1) don't assume any prior knowledge where it makes sense (for example, you can't practically learn about ML if you don't know how to program) and (2) provide practice -> Go through each item in the (filtered) list and see which one clicks -> stick to it and practice -> experiment |
1. Lay of the land (read, watch videos, what's there?)
2. Map the territory (Anki, notes, Feynman method)
3. Drill down (where are my gaps in understanding?)
4. Emulate (Watch tutorials then recreate them from memory)
5. Synthesize (build something cool)
I'd then go back and find any gaps or things that I was unsure of or didn't make any sense and then study those with extra caution. It was a very mentally-intensive exercise but it had proved the most successful study technique thus far in my academics.
Every element of a system leaks all kinds of related information, when you think of why it must be that way, consider what sort of elements must exist to support it, what similarities they must have, etc. I look at something, just a small piece at a time, then explicitly think through everything I can tell and what I can guess from that small piece I've considered in isolation, look at a bit more, occasionally throw out a bad hypothesis as a clearer picture emerges, look at the deficiencies in what I've done and figure out what should exist to prevent or obviate it, and often find that yes, such a thing exists.
It's hard to explain, kind of a style I've developed on the job, where you really just sometimes need to know things yesterday. When I consult other resources, it clicks better because it takes the haziness out of my sometimes half developed ideas, but the work I put in beforehand makes me remember it close to as well as if I figured it all out myself.
In other words, find some practical application of the knowledge that you want to develop and keep going until you've delivered the outcome you set out to.
I go to the table of contents and start with the most interesting part. You don't always have to begin on the first page or the easiest one. If I feel bored or forced, I move to the next.
Keep reading different interesting bits. The concepts will naturally link to one another. The more concepts you have linked to a new concept, the easier the new concept is to understand.
Some here say learn by doing, and that's basically the same thing, except that sometimes the easiest to learn bit isn't practical. Knowledge is usually delivered in abstract chunks, with no relation to something else, and having something you plan to do will relate it and make it easier to learn.
And if you prefer a visual medium, please check out the Coursera course "Learning how to learn" [1] by the same person.
I am by no means a expert but thanks to just looking up stuff on YouTube I got my head wrapped around some technologies like docker, aws, etc that just didn't make sense when I read the docs.
1) Something that requires expertise: learn by doing, with help from a respected author. Build on that by picking harder stuff to do. Doing makes memory stick.
2) Things that are mighty interesting: dig in (much easier with the web), make notes until 'enough-for-now' ... then add-to-notes as new resources pop up. Oh, and systematize notes so adding - and finding - is easy.
2. Combine that with "by the book" learning. Actually sitting down and reading a book, a man page, etc.
3. Getting a feel for the community and how they do things. In C you may revolt if you look at some of the code. But if you immerse yourself with peers you can learn the reasons (or lack thereof) behind things. Why data is stored in a a bunch of fixed size arrays instead of objects or whatever. Arcane but important tidbits you may not get from most books.
All 3 in no particular order/weight.
Then, I tend to follow Feynman's advice: keep trying to explain it to myself until I can make it clear. I typically do this through writing and examples (code, analogy, whatever). I can tell this is working when it's painful, and (for me) it always reveals the parts I don't understand.
Now good ole R2 takes care of the beep beep boop while I learn via ecstasy mostly ...
NYC is great for ecstasy learning
Your learning method seems dry
Just click here-- I searched for you: "how to learn"
--> https://hn.algolia.com/?q=how+to+learn
You owe me $500 ;) ill PM you my venmo
I also recommend variations on that query, such as:
- "how to teach yourself"
- "how do you teach yourself"
- "systems for learning"
- "learning systems"
- "how to self-study"
etc.
that'll be $2000 ;)