Zettelkasten Forum

How I refined my Youtube knowledge acquisition

Hi guys! I'm a typical lifelong learner. I'm constantly studying smth ranging from emotinal intelligence to NodeJs. In the last few years my content consuming process mostly moved to Youtube. And recently I've decided to make my workflow as smooth as possible. I'm so excited to share and get your feedback.

My typical Youtube learning workflow

My typical workflow while studying particular video usually includes 2 stages:

Fast aquaintance + timestamping

Initially I watch video at an accelerated rate (usually 1,5x). During this process I'm saving timestamps for each video moment, where I found a piece of valuable information. Usually I do it from my phone on the go or in a public transport. My "time-notches" go straight to Telegram ("Saved messages"). To save them I use "out of the box" Telegram feature. I just reply to Youtube video link message with timecode and Telegram automatically turns reply into the timepointed link.

Taking notes by timestamps

Getting home, I'm going over all "notches" and review the video in the corresponding timepoints to take notes into Obsidian.

This approach is OK in general. However, I found that for long videos it would be very handy to save a small piece of subtitles along with a timestamp. There are 3 benefits:

a. After watching the video and saving timestamps you can take notes mostly without any need to launch the video. That extremely speeds up the process.
b. You can prioritize your note taking process, what is extremelly valuable when your have to go through a long video and have limited amount of time. Having acquired the "whole picture" after watching the complete video, you can even exclude the particular timestamps from revision.
c. You can quickly go through your "info-notches" to get the whole picture even without taking notes of the video.

Having this in mind, I decided to create Telegram bot, which is able to pull out subtitles for the particular video timepoint and collect all these runtime "marks" to MD files (to be easily transferred to Obsidian with a click of a button).

The new scenario

Now I'm using my bot for saving Youtube video timestamps with subtitles around. The scenario is as following:

1. Send Youtube video link to the bot:

2. Then just send messages with timestamps formatted as following: "mm" (minutes), "mm:ss" (minutes seconds) or "hh:mm:ss" (hours minutes seconds). Bot turns them into links with timepoints and save with subtitles. Right after the timestamp (in the same message) you can also add any text. Bot will recognize it as a comment related to the particular timecode:

You can also send comments and pictures which are not related to the concrete timepoint.

3. When you are done watching one or several videos, just click "Collect all" button. Bot will pack your notches and comments to MD files, which you can download right away into your vault.

To end up with)

Just several points to end up with:
1. I'm planning to make bot as good as possible to satisfy Youtube learner needs. If you have any ideas on new features to add to the bot, just let me know. Also please notify me about bugs (if you meet them).
2. If bot is not responding to your input, just restart it with /start command
3. If you want to track bot updates and vote for feature priority, please follow this TG channel.


  • @apris, Welcome to the forums.

    Your YouTube learning workflow sounds like The Barbell Method of Reading technologically applied to YouTube.

    From The Barbell Method of Reading

    1. Read the book. Read swiftly but don’t skip any parts unless they make you vomit or put you to sleep. Mark all the passages that stand out and contain useful, interesting or inspiring information.
    2. Read the book a second time. But now you read the marked parts only. This time you make notes, connect them to past notes (Zettelkasten Method!) and think about what you’ve read. Make mindmaps, drawings, bullet points – everything that helps you to think more clearly.

    Using a 'Subtitle Dragon' to capture timestamps and Google's closed captions around highlights in a video is analogous to using a highlighter to highlight text in a book. I think this is ingenious.

    When I view a video for onboarding into the ZK, I use a much slower method. I envision the video as a lecture. I apply the lecture note-taking strategies and keep my eyes out, looking for new live-lecture and time-shifted lecture strategies.

    Effective Notetaking by Fiona McPherson provides me with some scientifically backed tips for live-streamed lectures. I apply the strategies the same way rather the stream is virtual or 'in real life".

    • the use of abbreviations and shorthand
    • Visual Note-Taking
    • pre-familiarity with the material
    • 90% of the success of knowledge transfer is on the lecture's shoulders - literally.

    From Will's ZK Note - Taking notes during a lecture 20220517

    Note-taking while watching a video or listening to a podcast is similar to a live lecture, except in those areas you have questions, you can stop/pause/rewind, which is impossible in a lecture. Although I can ask for clarification in real-time in a live lecture where I can't in a recorded one.

    Will Simpson
    My zettelkasten is for my ideas, not the ideas of others. I will try to remember this. I must keep doing my best even though I'm a failure. My peak cognition is behind me. One day soon, I will read my last book, write my last note, eat my last meal, and kiss my sweetie for the last time.

  • @Will thank you for the feedback! I will go deeper into this methodologies

  • Clever trick for the capturing. :)

    Just keep in mind that the capturing is the least time- and energyintensive part of the whole process. So, don't invest too much time in optimising and focus more on the part of actually developing the knowledge from the collected material.

    The ratio for a complex and useful book is ~ 16:1 (processing:first reading) (reconstructed from my personal work). If you are not practiced the ratio could be many fold higher. Estimating from my coaching experience it could be 4 to 6 times higher.

    So, if we go by the low end of my estimation focusing on learning how to capture is focusing on 1/64 of the whole process if you are not practiced.

    Meaning: Don't run into the suboptimisation trap.

    I am a Zettler

Sign In or Register to comment.