Most active commenters

    ←back to thread

    Learning to Learn

    (kevin.the.li)
    320 points jklm | 11 comments | | HN request time: 0.001s | source | bottom
    Show context
    setgree ◴[] No.41914724[source]
    Something from Andrej Karpathy on learning that stuck with me [0]:

    > Learning is not supposed to be fun. It doesn't have to be actively not fun either, but the primary feeling should be that of effort. It should look a lot less like that "10 minute full body" workout from your local digital media creator and a lot more like a serious session at the gym. You want the mental equivalent of sweating. It's not that the quickie doesn't do anything, it's just that it is wildly suboptimal if you actually care to learn.

    [0] https://x.com/karpathy/status/1756380066580455557?lang=en

    replies(4): >>41914797 #>>41915614 #>>41915657 #>>41925299 #
    1. zer0tonin ◴[] No.41915657[source]
    This is complete bs.

    I've learned 2 languages to fluency by mostly watching movies. I've learned the linux cli by setting up a minecraft server for my friends in high school. I've learned programming by making IRC (and later Discord) bots for communities I was part of.

    All of this was fun, and it worked better than staring at a textbook and hoping that my "effort" pays off.

    replies(6): >>41915888 #>>41916097 #>>41916444 #>>41918016 #>>41918152 #>>41920519 #
    2. bityard ◴[] No.41915888[source]
    I think there is some conflation in this thread between "learning" and "practice" which are fairly different things.

    As an ADHD person, nothing shovels the dopamine into my neural receptors quite like going from zero to "knows enough to be dangerous" in a new hobby or field of knowledge. That's the fun part. But climbing the experience curve much further than that requires some amount of _deliberate_ study and beyond that deliberate _practice_ and experience in order to become something like an expert.

    Chasing questions down rabbit-holes is fast and entertaining but only takes you so far. Deliberate practice (studying) is mostly less fun, even when that thing is your life-long passion and/or career. But necessary if you want to be highly skilled in that area.

    replies(2): >>41917866 #>>41918883 #
    3. sksxihve ◴[] No.41916097[source]
    Agreed, consistency is more important than doing a marathon session. Anyone who has learned a musical instrument can tell you this, far better to practice 10 minutes a day everyday than 1 hour once a week.
    4. danielmarkbruce ◴[] No.41916444[source]
    Yeah, but Andrej is talking about learning much harder things in much more depth. He's a world class research scientist and engineer.

    Practically everyone on planet earth learns a language as a child. Learning how to use some commands in linux and and programming a bot are literally child's play. I learned to play soccer the way you speak of - i'm... ok at it. Messi did a different thing.

    5. epiccoleman ◴[] No.41917866[source]
    > nothing shovels the dopamine into my neural receptors quite like going from zero to "knows enough to be dangerous" in a new hobby or field of knowledge.

    Man, this describes me to a T. I love that feeling of the "first 75%" (or whatever percentage it is). Then I tend to lose interest in the long tail.

    replies(1): >>41950104 #
    6. auselen ◴[] No.41918016[source]
    Trying to provide a perspective..

    Let’s say because of your genetics, you can enjoy playing basketball, and you do that, and you have fun doing that, you get better at it… but that’s that, it won’t be that easy for you with other subjects or even get to next level at basketball. Then you need to submit to discipline… or are you willing to wait for things to become fun?

    > pays off

    That’s why it is important to chase things you are curious about, then you don’t need to wait for some return…

    7. Arisaka1 ◴[] No.41918152[source]
    >I've learned the linux cli by setting up a minecraft server for my friends in high school.

    What you've learned was "enough to set up a minecraft server", but definitely not as much as "learned the linux cli", and that's without even touching the obvious question "what does someone mean when they claim they learned the linux cli"?

    I fell for this trap too: I watched a tutorial by Nick Chapsas, then made my own ASP.NET Core Web API for a personal project, and thought "Wow I know ASP.NET MVC time to get a junior developer job".

    After a few resumes (there's .NET demand in my area) I landed an interview with a startup. The interviewer (who happened to be the cofounder and has 15 years of .NET experience, some working directly in Microsoft) started hitting me with questions like "what kind of objects can get constructed in a using block?", "what's the difference between readonly and const", "how can we identify that a payload comes from a mobile client if the endpoints are shared?", etc.

    That's when I realized that I knew enough to make me go "woohoo I have a .NET Core web API" but not enough to get a junior job. In fact, he was honest enough to tell me "you're barely entry level, and this wasn't even the technical interview it's just the screener".

    Off-topic but related with the event: I obviously didn't get the job, but he left me with an advice I'm actioning: "Stop jumping around and stick to one language, I don't know if that's gonna be C#, TypeScript, Go, Python, whatever. Deep dive something well. You're only hurting yourself in the long run".

    replies(1): >>41922814 #
    8. smeej ◴[] No.41918883[source]
    I realized this about myself years ago, but I don't (think I) have ADHD.

    I have always been able to learn faster than most people. No clear reason why. It appears I was just born this way.

    But if the Pareto Principle holds, and I'm learning twice as fast as average (estimating for the sake of easy math), that means when I'm a beginner learning the 20% of the skill that gives me 80% of the results, I'm learning like 32 times faster than somebody learning at an average rate who's in the 80% of work that produces 20% of the results, even though they're better at what they're doing. I look like an absolute rockstar out of the gate.

    Problem is that my ego has been tied up in that since I was little. Early life is all about learning high-leverage things as quickly as possible, and since that's also when you're forming your sense of who you are, it's a sticky trap. I have really struggled to build the patience for the rest of the grind, where even if I'm still learning twice as fast as average at the harder level, any average newbie is learning/improving twice as fast as me.

    The end result is that I'm moderately proficient in dozens and dozens of things, but I'm not an expert at anything except obtaining moderate proficiency.

    9. coolThingsFirst ◴[] No.41920519[source]
    That wont work with real science. Math requires the process which he describes. As do algorithms.

    In my experience it’s same with gym, lifting weights slightly too difficult.

    The least productive method that Ive used with very little results is jumping to problems almost impossible for me. It did absolutely nothing.

    10. zer0tonin ◴[] No.41922814[source]
    Watching tutorials is exactly what I would consider "not fun", and I'm not surprised it wasn't successful. On my side I had no problem getting into a software engineering career, so my point stands.
    11. tartoran ◴[] No.41950104{3}[source]
    >Then I tend to lose interest in the long tail.

    What was your initial goal? Or maybe you didn't have a conscious one to begin with and had an attraction or something like a curiosity for the subject. Once you covered enough ground you satisfied your curiosity and your interest faded. I think this is a very natural outcome and depending how you approach your learning subject you can achieve different outcomes. Try learning in a class setting where there is a set curriculum and where you could approach your learning in regular and consistent chunks. It may be boring at first but it could get much deeper into the topic.