Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: How Do You Learn?
35 points by KarimDaghari 11 months ago | hide | past | favorite | 16 comments
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

My tech learning process:

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)

What I found really helpful for university lectures was to write down as much as humanly possible (in the form of free hand notes on the computer) after the lecture, without taking notes during it. Usually I'd get stuck after a few sentences but then remember a whole bunch of concepts attached to that sentence and start writing for a while.

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.

My learning style is pretty close to a side channel attack. I just start doing the thing accepting that I must be doing very terribly to start with but looking to improve as I go. It sounds a bit like learning by doing, and it kind of is, but often, my focus isn't on doing at all, and I find that I've learned more than I've done, learning by aggressive deduction and induction.

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.

by doing.

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 break down all concepts into a glossary terms. This is quite important because the terms used somewhere can mean very different things elsewhere.

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.

If you're interested in how our brains work during learning, check out "A Mind for Numbers" by Dr. Barbara Oakley [0]. It may be geared towards STEM college students, but I found it a great resource for fine tuning my own process.

And if you prefer a visual medium, please check out the Coursera course "Learning how to learn" [1] by the same person.

[0]: https://barbaraoakley.com/books/a-mind-for-numbers/

[1]: https://www.coursera.org/learn/learning-how-to-learn

Can you elaborate on this? Reviews are mixed. A lot of people calling this course and the accompanying book fluff.

Watching YouTube videos. There are pretty good videos now on a lot of specific software topics (for example I find lots of helpful videos on web, mobile dev, infrastructure stuff). Then just cloning some example repos and breaking stuff until it makes sense.

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.

Two styles of learning.

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.

1. Immersion. Jumping straight in to implement a feature, or achieve a goal. Even if understanding is currently incomplete. Lookup docs and online info as needed to overcome roadblocks.

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.

When it's an entirely new topic, I always build a big picture first: glossary, Key Concepts and How They're Related, What Is Thing For. But that's just how my mind works--have some framework to hang things on.

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.

R2D2 desktop chat bot replaced my daily paper fact of the day rip calendar and it was Beep boop straight too the moon after that ...

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

practice a lot, redo exercises, always start with blank editor

Check it out OP -- I used that little search box at the bottom of the news.ycombinator.com website--

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"


that'll be $2000 ;)

This reminds me of all those Stack Overflow "closed as duplicate", where the link doesn't even answer the question.

FYI, OP did actually pay me for my consulting services. 100% serious, not being sarcastic. Thanks OP for your generosity!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact