I don't know if I should comment or not, but I just did it: I have an extremely uncommon (almost weird) way of motivating myself to learn something... I just have the motivation to prove to myself that I can be disciplined... I learned all country flags and 80 digits of Pi just because I wanted to show myself, that I can learn it. I am surprised, that this learned information stays in my brain (I didn't forget them yet). Other reasons for motivation are: - wanting to understand everything (curiosity) - when you begin to learn something, you get sucked into it and then want to learn everything (like a addiction to collecting information; avoiding unfinished things) - proving to yourself, that you can come out of your comfort zone and still find something interesting - I don't recommend it doing it because of that, but: To brag to others - finding reason in why it's important to learn this stuff (as said in the video - like applications irl) That's the way I got addicted to learning new things... At the moment I learn every capital city of every country, just because I can... But I feel always interested when I'm learning (like learning for an exam or sth.),. although I shouldn't have a reason to (except the ones I mentioned) And I mean that seriously: Also big topics work with these strategies of motivation (at least for me)
just bumped into this video w/o metalearning beforehand. Then I lost focus during the ad break and didnt exercise because I was watching youtube. I schould probably drill on my time managenent a little bit to return to the big picture... like your videos alot! Thx Artem
@@MineGames66 no, because later you may end up with the perception of there is no truth but the consequences of the past which is your graduation or whatever.
"Learn how knowledge is structured in field x". That is probably the most important part. I accomplish this by just following respected scholars of a particular field on Twitter. Sounds stupid, but their random streams of thought will in fact reveal the "landscape" of the field. What is the broad research goal? What is the bleeding edge? You don't want to accidentally fail to make sense of something only to discover it is an unresolved problem. You need a rough outline of the field before you can dive into any particular topic in particular. Now, I didn't always approach it this way, I used to chaotically stumble through the internet, sort of like sampling from a function over and over again until you actually have a good picture of its surface... That's painful, but also works. Maybe if you have no idea what people mean by "landscape of knowledge" (which sounds ridiculous in a vacuum), try the brute force approach once for a field that is reasonably out of your confort zone such that you initially consider yourself completely lost. The more detached from your existing knowledge tree it is, the better. The more experience you get, the better you get at asking the right questions, or to use our analogy of sampling a function again, correctly guess the places that sampling the function will most likely reveal a lot of its structure. It doesn't make sense to sample f(0.1) f(0.11) and f(0.111) meaning closely related topics, if you want to get an idea of the function at large.
Actually, even more important might be the "personal projects" part. You roughly know what you want to accomplish, but you don't know how to get there. Note that you accomplishing said goal is the only signal you have to validate if you are on the right track. There is no instant feedback after you perform action 1 to tell you if it goes in the right direction. You will stumble in the dark for quite a while and then the result is pretty binary at first. Success or failure. Only later will you be able to interpret the type of failure into a sort of direction of what went wrong and thus what to do differently.
The main struggle in being a "learner" - be it that you need to learn specifics on a topic to get on with your main project or on the other hand start from scratch, do the deep dive and follow the rabbithole - is self-motivation; i am not talking about having the will to perform in the first place, to that there is a simple solution - if you want to get somewhere, move. But the most demanding subtask is the coordinated switching between the domains of learning, trying, practicing, testing, verifying. I hope more focus will be put on interactive assistants that are coupled with sensoring that is already available in say your smartphone, your smartwatch, your pc or mac. We are at a point where it is possible to get behind the holy grail of fully understanding cognition and how it is processed, if only the entry-barrier hardware-wise and therefore financially wasn't that steep to actually create/tutor models/networks - discovering a world in which knowledge is easily graspable in its core not via industry-proven academic preformed abstractions that only work as long as you are in that specific and very narrow window of operation. A major step would be a unified general math, which does focus on interchangeability instead of specific applicability - have the engineers use what works, but do not block understanding under the doctrine of industry- or academia-first - only where it merges, overlaps, intersects we can find what is missing. I want the birds-eye-view, not these tiny peekholes into reality and how it "most likely seem to" work. Well, in the end i am just human and therefore lazy, lol - i just want to be flooded by understanding and then watch myself coming up with the most abstruse ideas that actually work. One may dream, aight?
I think watching videos to learn something is good because it allows you to understand concepts, it will also help you know how to ask a question. While I agree that you should eventually write programs yourself, I don't think it's important until you have a firm understanding of the fundamentals. I don't program along side videos I watch because it will break my concentration of trying to understand whats going on.
This is an awesome video! Thank you for sharing your thoughts about this topic! I would like to ask you: Do you have any material that I could read about this info? (Articles, theses, books)
Well, apart from the brilliant mentioned book "Ultralearning", there is also a great online course "Learning how to learn" by Barbara Oakely (www.coursera.org/learn/learning-how-to-learn ) and a book by Peter Hollins "The Science of Self-Learning", which you might find useful ;)
@@ArtemKirsanov Thanks for sharing.I hope you can PLEASE respond to my other message about how to read scientific papers more easily and more enjoyably and faster when you can. Thanks and hope to hear from you.
Tell you a funny thing, I've tried to do all these things throughout the years in one form or another but never tried to systematize it. Now I've seen your videos, which made me more conscious of it. (Especially utilizing the Zettelkasten method.) Thanks for the info, really helps.
Knowledge map of the domain at study helps.. overlapping software libraries (solvers) on subcategories is a good tool.. the wiring (interlinks/directionality) within the map would reflect the number of ways or the different routes to get to the point.👩🏻💻🙇🏻♂️
Что ты думаешь про использование Anki карточек? Могут ли они помочь воспроизвести информацию, если например, нужно рассказать целую лекцию на какую нибудь тематику, или если это продавец, то рассказать о продаваемых продуктах и их характеристиках?
The first step was great 👍 if i would add one as a self learner it would be research what you want to learn maybe you will fined more or maybe y9u don't actually like it
maybe this is just a me thing but i don't metalearning the way i should approach a topic because there is a chance that i am not able to understand it and get tunnel vision to a single perspective and can't learn at all i instead only learn what i actually trying to learn, for example i want to make a software i need to know what a software is and what is the process of making one research about that gonna lead me to thing that i don't know, learn from there and slowly build up the knowledge to make a software but if i decided to research for how a person make a software result gonna be a lot more situation dependent you might need to learn a system programming language for a computationally heavy program but if you just want to make a game then python or c# is better you can also package a website into an application and turn the whole thing into website development the person show me how to do it know that fact and their approach and actions are influenced by their goal but i don't know that or maybe i just like learning the fundamentals before the applications idk
Dude you’re almost as awesome as Grant Sanderson. Or maybe you’re more awesome because I think maybe You are Younger and English may Not be your first language. Grant does have that wonderful sonorous voice, though. In any case keep up the good work. I probably will be checking out short form. Thanks