Ask HN: Do you maintain a log of what you learn? How often do you go over them?
Most of my leanings fades away as time elapses, except for those which I get a chance to correlate/apply often at life. How do you manage your leanings ? (copy-pasted with modifications from my comment on a similar thread posted earlier) I used to have an elaborate system, but I converged on a simple solution: I stash everything in a single Google Docs document. I made the conscious decision to optimize for ease of use, so that the friction/effort to write something down is minimized. At the same time, I also made a decision to not to adopt any organization system -- anything that increases the friction of use is eschewed. Search was all I needed. I've been using this system for the past 5 years or so, and it's been very productive. To extract ideas from it, I routinely re-read stuff (it's in log format, so it's very easy to read) and use the Fieldstone approach (Weinberg)[1] to coalesce similar and interesting thoughts and rewrite into larger thoughts. I've gotten a lot of actionable ideas this way (that I actually go on to execute on). So it's a system optimized for postprocessing rather than pre-processing. (The Fieldstone approach is a method from writing, rather than knowledge management.) I find that pre-processing systems are unsustainable over the long term unless you're exceptionally disciplined (also, it's hard to know how to structure knowledge until you've processed it -- most interesting ideas are a garbled mess when first encountered). I try to build systems that don't rely on sustained human discipline or the necessity of shoehorning into known organization units. [1] https://www.skmurphy.com/blog/2016/05/04/weinberg-on-writing... "I try to build systems that don't rely on sustained human discipline or the necessity of shoehorning into known organization units."
+1 I completely agree on this. The doc should have grown over time. Do you just append to the doc as you learn something new or do you organise learnings within the doc. I am asking this because if there is no structure there wouldn't there be too much of context switches when you go over them ? Yes I keep appending and rewriting. The doc keeps growing, but smaller thoughts that form part of a bigger idea get coalesced and rewritten. This is a constant process. Rewriting is a process of refining. Most ideas get rewritten and iterated on several times before anything coalesces. Also, it's common to have duplicate ideas at different times so coalescing them simultaneously consolidates them. For instance, when re-reading my doc, when I see multiple ideas related to say topic A, I put all topic A items together (just copy-and-paste into adjacent paragraphs -- nothing more), try to see the bigger picture and then try to rephrase them more succinctly. I constant rewrite and synthesize the disparate ideas into something more cohesive, kind of like snowballs clumping with some getting bigger over time. Some ideas do die in the process (you can copy-and-paste these into another Docs document), while others grow bigger and become more precise in their expression. The one idea behind the Fieldstone approach, often not considered in other methods, is engagement of the emotions. Of 100 ideas, I may only feel passionate about 5 at any given time because they resonate in my situation. For instance, some ideas/arguments/metanarratives which were meh to me pre-COVID now feel urgent, and with my current perspective, I've an emotional investment in pursuing and expanding them. You can have 100 ideas but you really on need to work on the few that you really care about, ie. those whose time has come. Only pursuing ideas that you're emotional about helps you focus and not be overwhelmed by the myriad ideas that are in the document. Inspiration is lumpy, but with a store of constantly refining/evolving potential ideas in the greenhouse -- which you cultivate over time -- you are better positioned to take advantage of those on the verge of blooming. It reminds me of what Richard Feynman said about convincing others that you're a genius: "always have 6 important problems percolating in your head; when you come across a new idea, test it against the 6. Once in a while you will hit jackpot and find a good match between problem and solution. Everybody will then think you're a genius for seeing the connection so quickly, when in reality you may have been thinking about the problem for years." (paraphrased) How long is your document currently, out of interest? 75 pages, 31k words. I think your approach sounds interesting. When you say you write things down in log format, do you use a field as a tag? Any chance you can paste an example entry? I don't usually tag my entries though I could. Here are two interesting ideas (by other people) that I came across, which I jotted down in order to process later. As I come across other related ideas, I cluster and rewrite when I feel like it, if they pique my interest. -- Credit cards are a form of microfinance for the privileged — they’re essentially 30 day interest free loans. -- Tobi Luetke: use systems thinking as a shared language. Drawing diagrams elucidates an idea and exposes assumptions. -- I find it terribly distracting to have all the things that I have tried to learn in my "active" memory. So I cycle between phases where I am engaged with a subject for a while, and then move on to the next item in my (circular) list until I cycle back to the original subject again. This is a serious impediment to developing any useful expertise in any area, but I have sort of made peace with this tendency. But not all is lost. It has been my experience that the next time I visit a previously learned topic, it's significantly easy for me to load everything to my "active" memory, and build-up from there. Apparently, you never truly forget the things that have once learned well. See here, for e.g.: https://www.sciencedaily.com/releases/2008/11/081117110834.h... > Scientists at the Max Planck Institute of Neurobiology have been able to show that new cell contacts established during a learning process stay put, even when they are no longer required. The reactivation of this temporarily inactivated "stock of contacts" enables a faster learning of things forgotten. I am no expert on memory, but from what I have read on the topic, our long-term memory is virtually unlimited, the main constraint being inability to recall rather than the memory content being erased. If you have once learned a topic well but haven't visited in a long time, you may not be able to recall it very easily, but you will be able to recognize or relearn it better than someone who doesn't have any exposure to the topic. "I am no expert on memory, but from what I have read on the topic, our long-term memory is virtually unlimited, the main constraint being inability to recall rather than the memory content being erased. If you have once learned a topic well but haven't visited in a long time, you may not be able to recall it very easily, but you will be able to recognize or relearn it better than someone who doesn't have any exposure to the topic. " I agree with your this part and personally my memory behaves the same way, but I am not sure if it should be like this, because then for example if I am having a conversation with someone about a topic I studied long back but did not recall recently then I am unable to contribute anything to the conversation although back in my mind I know that I studied or read about this but because of this poor memory thing I have to remain silent. I am more interested on how effectively we can apply the learnings to new problems. Sometimes it is not straight forward to connect the problem and the learnings. When application is not so straightforward, recalling the memory is out of question. Keeping a daily journal with relevant work and discoveries is great. You can use a simple text editor and keep appending the current date to the top of the file and writing a few a lines. You can do one per-project, or a general one for your daily life and activities. I reel back through the tape every few months when I'm curious about the timeline of development/research/discovery. Whittling it down, I think of it as Data, Insights, and Execution. First you need lots of data, then you get insights, and finally you can execute when you're close to maximizing on insights. Writing helps keep track of the first two, and if you want to teach then execution = communication, or how you package and present something. Does the logs stick to your memory when you go through a bunch of them together at once every few months ? Will it not be an overload at one go ? What do you mean by "learning?" because I mean something specific: the ability to adapt to solving new problems. It's not something that is so easily reified and commodified. If "learning" is remembering a password or a combination of keystrokes, I would argue that that is closer to "rote memorization" and does not really reflect the character or qualities of "learning" that this method is more suitable for. You should only look through your long-term notes to help solidify what you already know. If it's still blurry, then I would argue the learning phase is not complete. Short-term notes like a notebook and scratchpad, or a dry-erase board and camera on your phone, are better for actual acquiring of knowledge. You are spot on. Learning to me in this context is exactly what you have described "the ability to adapt to solving new problems". For e.g. within a social context a learning can be "when there is a mistake evaluate the intent rather the result". Another one related to work can be "Approach a product sale with an intent to solve the customer's pain and not to close the deal". There are too many of these learnings and many of them are deep. Thats why, I was wondering will one be able to get to the depth when many such learnings are reviewed at one go once every few months. It doesn't make sense to keep going over an ever-growing pile of notes, like a team of developers who's added so many features they don't have time to do anything but maintain them. You want to organize your notes so that you can go over them just when you need them -- when you're finally seeing that weird concurrency problem, that's when you skim over those notes you took on CountDownLatch or dining philosophers or whatever. So organize them and tag them so when the time comes, you can search for "composite key best practices" or "composing decorators" or whatever. Usually you get a fair amount of review just from the stuff you read while searching for other stuff you know is in there somewhere. Follow the same question in Ask HN asked a few hours before this one: https://news.ycombinator.com/item?id=23138530 I found a hackernews link like a month ago about making a daily TIL (today I learned) on github. I liked the idea enough so I made my own version: https://github.com/vincentntang/TIL It's where I store my code snippets, and a way to rubber ducky to myself. Everything I write on there is directly something I had to look up and didn't know off hand. Something to note is I do add things to my TIL during work hours and spend 10-15 minute tops writing each `.md` file I haven't really blogged much lately but this makes it easier to transition after. I wrote a React useState and useEffect guide this morning from my TIL notes https://www.vincentntang.com/react-use-effect-explained/ I highly recommend making a daily TIL since I can rich search all my notes for copy paste code I can use for projects My current philosophy is that programming is mostly about not making mistakes, so I'm most interested in cutting my error rate than in learning the latest tech, whatever that might be. To that end, I keep a mistake log, into which I insert an entry after doing something that either caused down-time or that caused my team to lose over an hour of productive time. It now has over 500 entries in story format, along with the relevant code. Some example titles (along with their tags in square brackets): - [gotchas] references to relative file paths are brittle when there are multiple entry points to a program - [gotchas] lazy evaluation cannot be cached in most systems - [unix] adding `sudo` affects $PATH and can cause commands not to be found - [gotchas] never confuse STDOUT with return value - [collaboration] never remove a flow in code without consulting all possible stake-holders - [algorithms] when calling a sorting function, always ask yourself if alphabetical or numerical I plan on posting a video with more details on my YouTube channel soon. I have a main text file I refer to and make edits in throughout the day ... as a developer I keep any programming specific nuggets here ... each topic has its area as in ubuntu tricks ___________ keeping one file makes it easier to search or browse ... separately I have other text files for broad categories like nyc, shop, bb ( catch all ) from a terminal ( Ubuntu ) I enter this to search svl 'some string here' where svl is an alias to a bash script containing stringPattern=$1 cat ~/MEGA/notes.txt | grep --color=auto -B 10 -A 15 -i "$stringPattern" I use mega nz for cloud storage so all this is mirrored on any box I use ... been using similar my entire career so this file today has over 90,000 lines of text notes ... I find such a process essential to be productive across a vast array of problem domains I keep a log of things like "here's the parameters for tar to extract a file" (yeah, I know, there's man pages, but not all systems have them). I review that item when I have to extract files from a tar archive. Until then, I don't care if I forget it - that's what the file is for. This approach works well for knowledge pockets where the discovery of application of knowledge is naive. When you want to extract a tar file and don't remember the command your brain know you have saved the command somewhere in a log and you can find it. But think of knowledge areas where you have to connect a few different learnings together to form an intuition on the application of knowledge to a problem. For stuff like that, I think that I keep the intuition without having to write it down and review it later. But I could be mistaken. If I don't use the intuition often enough to keep it available in my mental toolbox, then I don't use it often enough for forgetting it to be a major loss. I try to learn new things with the goal of: (1) attaining an intuitive sense for them (2) memorizing as little jargon as possible (only the amount necessary to look up things later). (3) making my life richer (4) making more accurate predictions about the world I encounter How? It varies. Lots of lectures and listening with no written notes. I had a professor who spoke so quickly I couldn't take notes as he spoke so I learned to listen as learning. Materially I make entries in a Google Doc which morphs often. The headings change and move around, entries are deleted and modified. - Precise dates and proper nouns do not interest me. IMHO that's all dry memorization, bits that don't connect to other bits, that's stamp collecting and tantamount to hoarding (which I think is gross, a literal failure of hygene). Started documenting the things I learn in a section of my website called "snippets". It's purpose is to be a registry of copy-paste resources, like a personal stack overflow, and I'm trying to keep it as small and short as possible. Also I don't write things there right away. First they go in a "backlog" in my notes, and once I require the knowledge but forget the details I transition it into a post on the website. This forces me to still remember stuff and helps keep only the relevant pieces around. Still very new to this approach, only time will tell if it's indeed worth the effort. The 2 stage approach sounds interesting. Do you cycle out notes from the second stage (snippets website). Do you follow different review frequencies for those 2 stages. Barely done this for a month now so the process is still young, but my plan is to keep around only the relevant snippets and update them in case anything changes. As to how often I review them, or the backlog, basically whenever I have time or I feel like doing it (trying not to put more pressure on me schedule than necessary). The only things that I do "urgently" would be to add elements to the backlog, and to fix errors in the published snippets. For everything else, I put a reminder in my calendar for the weekend, or after work in case I feel energetic enough. I simply just write them down as a form of blogging content. I personally think if you are on a personal journey to learn, everything you learn coupled with your own perspective could potentially be the eureka moment for another. Keep a journal! Here's my 'journal' to share my learning as a Software Engineer as well. Hope you enjoy it! https://blog.phuaxueyong.com/post/2020-02-29-articles-in-sec... I put new facts into Anki ( https://apps.ankiweb.net/ ) and then they get reviewed daily, with a steady, slow addition of new facts. It's a very effective way to keep everything in long-term memory. (It's also very difficult. You're trying to recall items that are just at the edge of your memory.) Interesting. Do you have fixed schedule every day to go through the facts in Anki or is it more on a ad hoc basis ? Yes, I use Anki every day. When I was working from an office, I used the commute time (I take the train) to practice Anki. Now, it's the evenings. I don't go over my notes. When I need some piece of info, either I know I noted this and go look for it or I just search with a few good keywords on my notes. Also, I just keep notes in one giant text file. I'd love some sort of fuzzy search though. I write things down in cherrytree nodes. https://www.giuspen.com/cherrytree/ With the goal of not having to go outside my normal flow of learning, I use: - Readwise.io: to capture and resurface highlights I make on my Kindle
- Pocket: to capture web reading and highlighting (readwise works here too) Simple, effective :) I keep a blog with short entries on how to do / repeat things.
Also writing it down helps me focus and get my head straight about what I'm trying to learn. What do you mean by learnings? Actual skills or facts like TIL? I meant the later