Active learning for lectures and podcasts

Active learning for lectures and podcasts

November 5, 2020

Something I’ve been thinking a little about is how to learn more effectively. My main philosophy around this is DRY (Don’t Repeat Yourself) and KISS (Keep It Simple Stupid).

My friend David introduced me to LiquidText, an extremely useful iPad app for indexing and closely analyzing documents. I shill it so highly because it makes it easier to A) start analyzing text in a structured manner, and B) lowers the activation energy needed to go back and look at notes. The big bang for buck I got out of 2018 was cloud iPad Apple Pencil notes and it’s apps. Before that, I took notes on paper, and the paper that always seemed to fall off the face of the earth. Why even bother to take notes if I would just lose them, and if I didn’t, they weren’t easy to organize, and even if they were, I still only had a physical copy.

For that reason, I also highly shill Kindles, because taking notes and highlighting is cloud based and easy to do, look up, or export. I’ve once mentioned that I think writing is the greatest invention ever - indexing and search are also in that league.

In the same spirit, I notice myself watching lecture videos, listening to podcasts, and even movies and taking notes on the side and wanting to index it in my own way. For example, at the 5:30 minute mark of Lecture 8, Dr. Knuth talks about XYZ.More often then not, any video worth watching isn’t only worth watching once. Anything worth studying really needs multiple passes, and the faster and easier you make those passes the better.

Video is unique because unlike documents, I cannot always get the video. For Youtube videos, it’s not so bad to download the videos, but storing a lot of of video is admittedly a hassle. Nobody downloads podcasts too, we stream it. These mediums are also extremely new (from a more balanced time scope), especially for learning, and I think optimizations can be made. I believe Apple’s iPad pencil technology is extremely well done and has created extremely high value if you can use it right. Chew on it a bit more: the FIRST cloud based, writing composable technology, that works smoothly. Hot take: most of us are desensitized to technology, especially design, integration, and smoothness advances. I’m excited to see more advancements on the video and audio front.

But that doesn’t mean there isn’t improvement over this process.

I intend on starting something simple for myself: a note-taker add-in for VLC that lets me add moments or bookmarks. This already exists, works nicely, but I’d like to add more features to it, and get it working better. Especially since this add-on is also standing on the shoulders of a previous popular add-on.

Potential Features

  • export notes
  • ideally link would work in web
  • sort by timestamp
  • maybe be better as a chrome extension for indexing video and time marks (this seems quite feasible and valuable)
  • should work on cloud (phones, iPads, other). Instead of sharing a tweet, sharing a clip, or a series of clips
  • File only needs to have its unique format
  • Can create a collage of videos
  • should work for podcasts too
  • mostly means we have to download the files and have them handy
  • could integrate nicely with dropbox still

Ideas for better video learning

  • indexing for subtitles, topic modelling, self index creation
    • searching moments this way to jump to that place
  • video subtitle hashing without ML and then doing topic modelling

References

[1] KISS and DRY