This guide is in response to a request in the open thread.  I would like to improve it; If you have some improvement to contribute I would be delighted to hear it!  I hope it helps.  It was meant to be a written down form of; "wait-stop-think" before approaching a new area.

This list is mean't to be suggestive and not limiting.

I realise there are many object-level opportunities for better strategies but I didn't want to cover them in this meta-strategy.

It would be very easy to strawman this list. i.e. 1 could be a waste of time that people of half a brain don't need to cover.  However if your steelman each point it will hopefully make entire sense.  (I would love this document to be stronger, if there is an obvious strawman I probably missed it; feel free to make a suggestion for it to obviously read in the steel-form of the point.

 

Happy readings!


0. make sure you have a growth mindset. Nearly anything can be learnt or improved on. Aside from a few physical limits – i.e. being the best marathon runner is very difficult; but being a better marathon runner than you were yesterday is possible. (unknown time duration, changing one's mind)

 

  1. Make sure your chosen X is aligned with your actual goals (are you doing it because you want to?). When you want to learn a thing; is X that thing? (Example: if you want to exercise; maybe skiing isn't the best way to do it. Or maybe it is because you live in a snow country) (5-10 minutes)
  2. Check that you want to learn X and that will be progress towards a goal (or is a terminal goal – i.e. learning to draw faces can be your terminal, or can help you to paint a person's portrait). (5 minutes, assuming you know your goals)
  3. Make a list of what you think that X is. Break it down. Followed by what you know about X, and if possible what you think you are missing about X. (5-30 minutes, no more than an hour)
  4. Do some research to confirm that your rough definition of X is actually correct. Confirm that what you know already is true, if not – replace that existing knowledge with true things about X. Do not jump into everything yet. (1-2 hours, no more than 5 hours)
  5. Figure out what experts in the area know (by topic area name), try to find what strategies experts in the area use to go about improving themselves. (expert people are usually a pretty good way to find things out) (1-2 hours, no more than about 5 hours)
  6. Find out what common mistakes are when learning X, and see if you can avoid them. (learn by other people's mistakes where possible as it can save time) (1-2 hours, no more than 5 hours)
  7. Check if someone is teaching about X. Chances are that someone is, and someone has listed what relevant things they teach about X. We live in the information age, its probably all out there. If it's not, reconsider if you are learning the right thing. (if no learning is out there it might be hard to master without trial and error the hard way) (10-20mins, no more than 2 hours)
  8. Figure out the best resources on X. If this is taking too long; spend 10 minutes and then pick the best one so far. These can be books; people; wikipedia; Reddit or StackExchange; Metafilter; other website repositories; if X is actually safe – consider making a small investment and learn via trial and error. (i.e. frying an egg – the common mistakes probably won't kill you, you could invest in 50 eggs and try several methods to do it at little cost) (10mins, no more than 30mins)
  9. Confirm that these are still the original X, and not X2, or X3. (if you find you were actually looking for X2 or X3, go back over the early steps for Xn again. (5mins)
  10. Consider writing to 5 experts and asking them for advice in X or in finding out about X. (5*20mins)
  11. Get access to the best resources possible. Estimate how much resource they will take to go over (time, money) and confirm you are okay with those investments. (postage of a book; a few weeks, 1-2 hours to order the thing maximum)
  12. Delve in; make notes as you go. If things change along the way, re-evaluate. (unknown, depends on the size of the area you are looking for.  consider estimating word-speed, total content volume, amount of time it will take to cover the territory)
  13. Write out the best things you needed to learn and publish them for others. (remembering you had foundations to go on – publish these as well) (10-20 hours, depending on the size of the field, possibly a summary of how to go about finding object-level information best)
  14. try to find experiments you can conduct on yourself to confirm you are on the right track towards X. Or ways to measure yourself (measurement or testing is one of the most effective ways to learn) (1hour per experiment, 10-20 experiments)
  15. Try to teach X to other people. You can be empowering their lives, and teaching is a great way to learn, also making friends in the area of X is very helpful to keep you on task and enjoying X. (a lifetime, or also try 5-10 hours first, then 50 hours, then see if you like teaching)

Update: includes suggestion to search reddit, StackExchange; other web sources for the best resource.

Update: time estimate guide.

 

New Comment
39 comments, sorted by Click to highlight new comments since:

I think it would be good to put the points into subsections and then describe why each subsection is valuable. I have put some draft subsections below:

Have growth mindset

  1. Nearly anything can be learnt or improved on. Aside from a few physical limits – i.e. being the best marathon runner is very difficult; but being a better marathon runner than you were yesterday is possible.

Clarify what X is

  1. Make a list of what you think that X is. Break it down. Followed by what you know about X, and if possible what you think you are missing about X.
  2. Do some research to confirm that your rough definition of X is actually correct. Confirm that what you know already is true, if not – replace that existing knowledge with true things about X. Do not jump into everything yet.
  3. As you learn more about X, consider coming back to this point to confirm that these is still the original X, and not X2, or X3 etc.. (if you find you were actually looking for X2 or X3, go back over the early steps for Xn again.

Determine if it’s worth it to lean about X

  1. Make sure your chosen X is aligned with your actual goals (are you doing it because you want to?). When you want to learn a thing; is X that thing? (Example: if you want to exercise; maybe skiing isn't the best way to do it. Or maybe it is because you live in a snow country)
  2. Check that you want to learn X and that will be progress towards a goal (or is a terminal goal – i.e. learning to draw faces can be your terminal, or can help you to paint a person's portrait).
  3. Get access to the best resources possible. Estimate how much resource they will take to go over (time, money) and confirm you are okay with those investments.

Determine best practices/common mistakes

  1. Figure out what experts in the area know (by topic area name), try to find what strategies experts in the area use to go about improving themselves. (expert people are usually a pretty good way to find things out)
  2. Find out what common mistakes are when learning X, and see if you can avoid them. (learn by other people's mistakes where possible as it can save time)
  3. Check if someone is teaching about X. Chances are that someone is, and someone has listed what relevant things they teach about X. We live in the information age, its probably all out there. If it's not, reconsider if you are learning the right thing. (if no learning is out there it might be hard to master without trial and error the hard way)
  4. Figure out the best resources on X. If this is taking too long; spend 10 minutes and then pick the best one so far. These can be books; people; wikipedia; website repositories; if X is actually safe – consider making a small investment and learn via trial and error. (i.e. frying an egg – the common mistakes probably won't kill you, you could invest in 50 eggs and try several methods to do it at little cost)
  5. Consider writing to 5 experts and asking them for advice in X or in finding out about X.

While learning X, externalise what you have learnt.

  1. Delve in; make notes as you go. If things change along the way, re-evaluate.
  2. Write out the best things you needed to learn and publish them for others. (remembering you had foundations to go on – publish these as well)
  3. Try to teach X to other people. You can be empowering their lives, and teaching is a great way to learn, also making friends in the area of X is very helpful to keep you on task and enjoying X.

Try to apply what you learnt and find ways to improve what you have learnt

  1. try to find experiments you can conduct on yourself to confirm you are on the right track towards X. Or ways to measure yourself (measurement or testing is one of the most effective ways to learn)

Nice organization. Just wanted to say that you neglected to copy the ends of some of these lines.

I have fixed it. Thanks for letting know.

[-]Elo20

I was aiming for a linear process. Hopefully someone can comment on whether linear process works better for them or not. I will look at what might be the best way to divide things; I quite like some of your categories.

[-]Elo20

http://lesswrong.com/r/discussion/lw/mk8/stupid_questions_august_2015/cmei link to original thread question. Although its not all that important.

[-]Elo00

I note this article appeared in my feed; http://lifehacker.com/the-only-technique-to-learn-something-new-1709323497?utm_campaign=socialflow_lifehacker_facebook&utm_source=lifehacker_facebook&utm_medium=socialflow and has the sub-headings:

  • love it
  • read it
  • try it. but not too hard
  • get a teacher (plus the 10x rule - which is that coaching-time potentially multiplies your learning)
  • study the history, study the present
  • do easy projects first
  • study what you did
  • you are the average of five people around you
  • do it a lot
  • find your evil plan

Which are some similar points to the ones I covered.

I think your discussion post is better aligned than this article. I tried reading it...

I think I'm going to have to reject the article since it puts too much pressure on the talent and love aspects. I have never really loved anything that much, and if I was too worried about that, I'd end up in the "finding your passion" rut. Judging by some articles I've seen spring up recently, I don't think I'm alone. I appear to have talent in some areas, but I can't tell if I have talent in any areas that I do not already know I have talent in, nor can I measure the extent of my talent. I don't think this is a productive way to go about things since it's so ambiguous and provides too many avenues to drop the very idea of acquiring new skills.

Some people are just not that crazy about things and never got any obvious signs about a skill, but they still need reasonable advice on how to learn new things, because they can still learn some degree of mastery in those skills and those skills are still useful. I can make a significant difference by simply being a better programmer as opposed to a great programmer. Perhaps the issue is that we focus too much on doing things statistically unusually well as opposed to doing things because they're worth doing. If the thing you are trying to do is only worthwhile when you are the best at it, perhaps that's a signal that it's a relatively useless activity or there's already an over saturation of supply in it.

This is a very greatness focused article and I think that's actually a rather toxic mentality to approach things with. We should focus on standards and mastery (and personal benefit) instead of greatness. It also ignores generalists. I guess those are both rants for another time.

[-][anonymous]00

You've done it again Elo. Great guidelines.

[-]Elo-10

Edit2: thank you for your kind support! I found it good feedback :)

It has given me cause to have some concerns arising from you finding it useful.

The reason is - From interacting with you over the lesswrong site; I don't evaluate your opinion to be mainstream in society; or even mainstream within Lesswrong (to make this clear; thats a good thing for you). The concern is that you may be effectively a, "canary in the coal mine" kind of indication that I have not done what I set out to do in writing this.

I have two theories as to why you found it useful:

  1. you happened to need it and I provided it at an opportune time.

  2. This guide is not actually helpful to the boring ordinary humans; or the more-frequent-in-number and less-often-expressive lesswrong user.

I am glad you found it useful; but also I want to not deluding myself that this was helpful to many people if it was not. I expect it will lower the chance of me making more things like this if it is not helpful to the broader slice of people.

Can you perhaps enlighten me on your thoughts on 1 & 2?

Edit: for making my statement more clear.

[-][anonymous]10

That was really mean, dude. Think how you would feel if you gave someone a compliment then they said this about you.

[-]Elo10

:/ Can you understand my intention in the post? And can you help me re-word it to be less mean?

Edited; is that better?

I disagree with MattG. I don't think that it sounds mean as it currently stands (I did not see the original). However, NONE OF THIS MATTERS. Check with Clarity. What MattG and Tem42 think only matter when you are asking them a question.

[-][anonymous]10

I've felt amused, baffled, humbled and proud thinking about your reply but mostly grateful about the comment.

Elo I really appreciate it because I would like to set a precedence of open feedback and I love hearing what people think of myself since my self-awareness and insight isn't the best.

[-]Elo00

Thanks for the feedback; I feel like this whole event was handled entirely reasonably by all parties (of course nothing bad happened so it might feel less reasonable if something bad did come of it)

The original post is heavily modified from the first iteration. look at every explicit clarification of things and simplify it to one word. i.e. It has given me cause to have some concerns arising from you finding it useful." used to read, "it concerns me that you find it useful". which when read wrong is really not a nice line.

[-][anonymous]10

There's no way to word it in that way that doesn't sound mean. What I would have done was just say "Thanks! What did you find useful about it?" Which would have gotten the same information without passing judgement about someone's normalcy.

[-]Elo10

Oh no! I don't want to sound mean at all! I suspect that clarity understands their non-mainstream position. And I don't mean it as a bad thing at all; just as a point of declaration. (I tend to find value in non-mainstream myself)

I suppose that its entirely unclear that I value non-mainstream over mainstream unless you know the internal workings of my head. And therefore that non-mainstream is a compliment not an insult - is entirely unclear.

I am going to have to hope that Clarity reads all the posts and makes a reasonably charitable judgement about my comments...

[-][anonymous]10

Yes I feel like a bigger deal is being made out of this than it deserves. If anything, I would hope that more people took your approach. We have interacted on LessWrong enough to have made a sophisticated judgements about me that has made for a more useful interpretation of my response than otherwise, and you've expressed that which will help others who don't share that private knowledge.

I think it will be a real shame if you avoid making a similar comment in the future which benefits you and me both, but the community, altruistically, effectively and empathetic-ally, thinking what's best for two consenting adults, confuses one or both parties or complicates their utility calculations and/or theory of mind.

When reading you post, I read no mean intent from it and had no reason to fill in any blanks with it. Yep I'm reading all these posts and you're getting a well-deserved upvote for your reply.

Edit: I'll add that Elo, MattG and Lumifer, all participants in this chain of comments are clearly in the top 10 and highly like top 5 or higher or people who've made my stay on LessWrong..well..bearable to the rest of you. My post quality is gradually getting better and better based on feedback of any kind.

Edit 2: I'll add that I also found the original reply a bit funny and I'm having a bit of a giggle thinking that my first edit might distress Elo again. Though, I think his theory on that matter is fallacious and I'm looking forward to seeing if and how his ideas change.

[-]Elo00

I have a few more posts in the series. Several are partially completed, most of what is slowing me down is the time they take to create, I have estimated 22 hours of time to create the things I want to create; give or take; of similar content that I am in the process of preparing. (9 posts ~2-3 hours each)

I also like to write it. wait a week; re-read it; and clean it up to make more sense.

[-][anonymous]10

Maybe this is just the typical mind fallacy, but if I said I like something and someone else says "Oh no!" instead of "Thank you", I wouldn't feel very good.

[-][anonymous]00

I didn't read any negative connotations into it. I take no offense from it at all. I did find the side-stepping a bit awkward and that a bit offensive, because it implies Elo didn't want to incur the wraith of the community or my ire from something more direct. That would imply either/and I don't come across as receptive to negative evaluations (when in fact I have a slight preference for them to positive ones since I know which areas to improve then). Though, Elo is right in his caution since he has incurred right.

You know what they said, any information is good to a Bayesian, or something like that.

[-][anonymous]10

Perfect! Typical mind fallacy it is then.

[-][anonymous]00

I feel really good and cared for reading this. I also feel slightly ashamed and undeserving like I'm taking your attention away from potentially more important things. Thank you.

[-][anonymous]00

thanks for the feedback

re theory 1. doesn't conflict with any of my intuitions

re theory 2. I don't find that my preference for something as expressed on LessWrong predictably implies a contrarian opinion?

It concerns me greatly that you found it useful.

LOL

[-]Elo10

I hope my reasons are reasonable about that...

Well, you are on LW. It's... unusual to find here reactions which can uncharitably be summed up as "OMG, the freaks like it, what's WRONG with it??!?"

[-]Elo10

It sounds that bad? Dangit!

I am aiming to generate rationality content that is applicable to mainstream; clearly I am not doing that if my hypothesis was correct.

Does that statement make sense?

It sounds that bad? Dangit!

I have a nasty mind and a propensity for snark, so I can't guarantee that my reaction matches the mainstream :-P

I am aiming to generate rationality content that is applicable to mainstream

If you actually are aiming at a target, you need to define it a bit better. "Mainstream" is very very fuzzy and a lot of that mainstream is pretty dumb.

[-]Elo10

how about replace "mainstream" with "a large number of people"

That's just making things fuzzier.

There is usual advice that one should write for a particular audience. Who is your intended audience?

[-]Elo10

My purpose was, "generate rationality content". And create a guide where I could not previously find a guide describing how to learn a new skill X.

Who might find this guide useful? (a similar question to who is my intended audience)

  • people trying to learn a new skill X
  • rationalists trying to understand the meta-process of learning

I am not sure how to better test for those two groups of people.

people trying to learn a new skill X

So, let's try a couple of different examples. Let's take (1) the skill of sailing a small boat; and (2) the skill of starting a social interaction with strangers. Do you think your guide would be useful for these two specific examples?

rationalists trying to understand the meta-process of learning

Trying to understand how learning actually happens, or how, from your point of view, it's supposed to happen?

[-]Elo10
  1. people trying to learn a new skill X
  2. rationalists trying to understand the meta-process of learning

2 first - I meant - learning how to learn. or trying to get better at techniques of learning.

1a. sailing: I just ran through the first few points in my head (took about 5-10 minutes); and yes it seems to work for sailing.

1b. social interactions with strangers: yes it works. It occurs to me that there is no indication of how long each step should take. I am going to add that in as a guide.

[-][anonymous]00

Say you're aiming to catch a bear. So you put out a a pile of fish. A crow eats one. Does that mean that bears don't like fish? No, and while you're thinking about that you might any bears going for the fish.

[-]Elo00

Cryptic as that is; I understand; trouble is that the bears didn't eat many fish. So maybe these bears don't like the fish that I put out...

I find 3, 4 and 9 difficult to implement in teaching oneself a language.

[-]Elo30

example of 3 for learning a language - Chinese (I don't know much about Chinese) to a fluency level of being able to function while on holidays in china for a month:

Make a list of what you think that X is. Break it down. Followed by what you know about X, and if possible what you think you are missing about X.

  • Characters are words
  • there are a million or so characters
  • Characters are sometimes made up of other characters
  • the sounds for characters are unrelated to the noises for speaking those characters where noises do not build from other noises I don't know
  • directions
  • toliets
  • medical
  • food related words (I am going to stop because I feel like this is a good start of demonstrating this concept)

for 4 in this specific example - confirming how many characters there are, and the general map of characters to each other, and verbal language to each other. (I assume the difficulty in 4 was in not having 3 completed)

For 9 specifically - Don't buy a book about Chinese language history if you are looking to be fluent in Chinese for a month.

Does this help? Or are your difficulties elsewhere?

Yes, this helps.

[-]Elo00

I was thinking that now I need a step of "define X" earlier on. The problem with "learn a language" is its very broad; like people saying they want to learn programming. Or exercise more. Or lose weight. Its a not an SMART goal in not being specific.

Thoughts?