Learning effectively with podcasts
Integrating ListenNotes, Otter.ai, Roam, and walks around lakes
|Apr 6, 2020||21|
The amount of high-quality content available as podcasts is astounding, but coming up with an effective approach to listening, processing and actually learning from them is not trivial. In this edition, I share my approach in detail - how to actually learn and retain information, not just distract yourself from doing the dishes. I also talk about how my employer, the Minerva Project, is dealing with the Corona crisis, and provide a variety of links to community content about Roam and note taking.
Newsletter in Corona-times
It’s been more than a month since the last newsletter, and the world has changed. I’m lucky to have a secure job – in fact, my job has become even busier due to the crisis (some details below), but two parents working full-time, and taking care of an active six year old whose school it shutdown has been exhausting.
The topics of this newsletter are of course more relevant than ever, and I have a long list of topics I’d like to write about, but I can’t promise how regular the issues will be.
I would love to hear from you. Did you know that you can simply hit reply to this email, and I’ll get your response? Do you appreciate long pieces connecting different ideas, or shorter segments? Specific topics you’d like me to cover? Reach out!
Learning effectively with podcasts
– “I am now walking round after round around this beautiful, calm lake, with these incredible stars above me. And I'm feeling very, very good. My body's not tired at all. I am not disturbed by anyone. There's no noise or cars. I have time. I am learning incredibly interesting things. I'm not stressed out about my memory because I'm able to capture my thoughts and go back to them later. It reminds me a bit of monks or, you know, all kinds of people walking and learning.
I'm curious how this differs from reading and taking notes. Certainly being in front of my computer, I would have a much harder problem concentrating. moving my body is a kind of great background thing. I'm also curious if there's any difference in the processing, listening versus writing and reading and speaking and so on.”
I took these notes using Otter.ai, while walking around Lac Sauvabelin, an artificial tiny lake in a forest north of Lausanne, at night, listening to John Vervaeke’s “Awakening From the Meaning Crisis”, a lecture series of 50 episodes.
I’ve always enjoyed going for long walks. While we lived in Toronto, I would go out almost every night, traipsing through the ravines, or finding different quiet residential streets, while listening to audiobooks or podcasts. In Lausanne, we live in a tiny apartment, and I spend most of the day sitting in front of a computer, so going for a walk in the evening both gives me some privacy and a chance to move. I’ve listened to tech podcasts, French and German podcasts to improve my language, a lot of American politics etc. I’ve enjoyed it, but seldom felt the need to take notes or used it to further my learning in a systematic fashion.
In 2020, that has changed, due to three key enabling technologies, and a new set of interests. The catalyst was Roam, not only for being an excellent tool, and for reawakening my interest in serious note taking and knowledge management that I thought a lot about during my PhD period (video), but also because the Roam-adjacent community introduced me to a great swathe of new thinkers and new ideas which I have been diving enthusiastically into during the last few months.
During another walk, in Norway in March, I recorded the following:
Side thought: I guess my current interest in all of these things is a bit related to Maslow's hierarchy. That because I have a safe job that not only guarantees economic income, but also societal status, mental challenge, sense of purpose and interesting coworkers, I can see in a much longer perspective, I can start a project now, that might take me 10 years to finish and I can focus more on self-actualization.
Reading about the Listening Society I feel a little bit like when I was exploring the ideas of veganism and animal rights, because I had already been talking to Espen, and preparing myself to the possibilities, I was much more open minded to ideas about the fact that how we treat animals is cruel, whereas had I read the books without this preparation, I might have completely rejected their conclusions because they were too incongruent with my own identity and worldview. Similarly, because of my opening to ideas around meditation, trauma, therapy, mental growth, and so on, I am much more able to appreciate the ideas in the book The listening society.
So part of my thinking is that having been in academia for the past many years, the past three years as a post-doc, much of my reading concentrated on my discipline, and I spent a lot of time subconsciously worrying about the future, not knowing where I would end up (having to move for an academic appointment, etc). Since fall, I’ve been working full-time at the Minerva Project, and I think this has been tremendously important in giving me a stable platform where I can begin asking myself: “What else do I want to do with my life. What do I want to learn?”, and also the strength to be able to say “This might take me 10 years to learn, I better get going” rather than “That’s too long”.
I realize all this is fairly meta, and perhaps you were expecting a very technical description of how I process podcasts, so let’s get to it.
Finding good podcast episodes
Up until very recently, the way I listened to podcasts was using an app on my phone (in my case Overcast), subscribing to a number of podcasts, and seeing which new episodes appeared. However, once in a while I’d come across a Tweet or a website where someone mentioned an episode of a podcast that I was not subscribed to, and I thought “argh, I don’t really want to subscribe to this podcast, just to get a single episode”. It was when looking for a way to solve this, that I came across ListenNotes.
First a word about podcasts. A podcast is really a very simple (but brilliant) technology. At its core, each episode is just an MP3 file stored somewhere publicly accessible on the net, and the actual podcast is a simple text file, using the same RSS format which enabled Google Reader to follow blogs, describing in machine-readable format each episode, the title, description and link to the URL. An episode might be described as follows:
This looks messy, but is semantically encoded information that makes it trivial for your podcast application to display a list of titles, know which MP3 to fetch when you click on a certain title, etc. But how do you find these RSS feeds in the first place? Back when feed readers to read blogs using RSS feeds were popular, you used to go to a blog’s homepage, and your browser would tell you that it had an RSS feed. Clicking an icon would add this feed to your feed reader, or you could paste the homepage URL into your feed reader, and it would look through the homepage, and extract the feed URL (using metadata). However, now that we’re all on mobile devices, this doesn’t work very well.
I believe it was Apple that really propelled podcasts forwards with the inclusion of a native podcast player on the iPhone. And to make it easy for people to find interesting podcasts, they integrated podcast search and discovery into iTunes - they asked people to submit their feeds, and make sure it had enough metadata (nice channel image, description and tags) to enable Apple to suggest relevant podcasts, and ensure the listing in app store looks good. And since this was a huge market, most podcasters were happy to conform.
The crucial point is that Apple decided to open access to this database to any app or website. This means that I could build a new podcast app (like Overcast), and through a simple API request, I could provide my users with a search through the entire well-curated catalogue that Apple had already put together. And this of course made it even more attractive for podcasts to submit their information, and make sure their feeds were well formatted.
So for example, I can run a simple query on my command line:
curl "https://itunes.apple.com/search?term=future%20thinkers&entity=podcast" | jq .
(jq is a useful tool to format JSON data)
This performs a search against the Apple podcast database with the query “future thinkers”, and the result comes back immediately:
This is all the metadata required to display this channel beautifully in my podcast app or website.
So far so good, but it gets better - ListenNotes is a website which uses this data to build up a database of every podcast and every episode out there. This lets you search not only based on podcast title, but also on episode description (which for most podcasts at least contains the names of guests, books discussed, etc). Further, given that a podcast is simply a text file describing a set of MP3 files, and there is no requirement for all these MP3 files to be hosted at the same place, they give you the ability to create a custom “podcast playlist”, composed of the episodes you want to listen to, from any source.
You can either use their browser plugin, or go directly to their website and search for any term. If you find an episode you would like to listen to, you can add it with two clicks. And since this looks just like a normal podcast, it works with any podcast player natively. Let’s say I was interested in Andy Matuschak’s ideas, I might search for him, and find some really interesting episodes:
I add the ones I’m interested in to my ListenLater, and they’ll automatically download to my Overcast feed. Here’s an example of how it looks on my phone, you see episodes from several different podcasts listed together (you can find my public feed here, but you should really make your own instead):
When I’m getting into a new thinker, or preparing to read a complex book, I enjoy finding several interviews with the author from different perspectives – this gives me a high-level overview, and makes it much easier to effectively take notes once you dive into the book (or decide to skip it). I’ve also done the same for concepts I’m curious about, from meta-modernism to non-violent communication.
Side-note, aka lament, about academic metadata
I’ve been frustrated about the state of academic metadata for almost a decade. A great read, all the way from 2004, is a paper called “Why can't I manage academic papers like MP3s? The evolution and intent of metadata standard”. They talk about how back in the days when you ripped audio CDs, programs would actually upload fingerprints of the audio to a community server, and download community-supplied metadata. Later, MP3 files have embedded metadata (ID3 tags), meaning that you can move them seamlessly from iTunes to any other app, and it will “just work”…
16 years later, and Zotero has some support for automatically fetching metadata, PDFs have a metadata field that nobody are using, but we’re still mostly stuck in the stone-age… Why can’t I just drag and drop citations from a paper into my citation manager, and have the metadata imported, and the PDFs automatically downloaded?
Listening, and note taking
Now that I’ve filled my phone with incredibly interesting episodes, full of great thinkers and profound ideas, I need to actually listen, and process. Trying to keep all the ideas presented in a one hour podcast in your head until you come back is a recipe for frustration, but bringing a notebook with you while walking doesn’t make sense. Otter.ai to the rescue!
Otter is a tool for transcription, which pitches itself as a way of generating transcripts of meetings, and provides 10 hours of transcription for free every month. I had tried voice transcription a few times before, with mixed results, but I realize now why my experience with Otter is so different.
Previously, I might want to dictate a message while on the run, and use Siri or Google to take the input, before sending it. This kind of works for short messages, but the moment it makes a mistake, it can be quite annoying to try to fix, and it’s not appealing for longer thoughts. Otter does cloud processing, but it offers a live-preview while talking. However, you can turn off sending to the cloud over 3G, and in that case, it will only record. I highly prefer this, because I do not want to look at the words and get frustrated about the transcription.
It’s super-quick to record a note, simply switch to the app, hit the microphone, and start speaking. When you’re done, click stop. Once I’m at my computer, it uploads all of the snippets, processes them, and typically generates excellent transcriptions. And for the few times that it misses, it has your entire voice record, and you can click anywhere in the text to hear what you said at that moment, which makes it super-quick to clean up any mistakes. This is how it looks when I’m recording on the go:
And once I’m processing the note in the web interface
I have shared the above recording publicly, you can both see how I sound when I’m walking and talking, but also explore playing while Otter is highlighting the words, clicking on words to move around, etc.
This has really been a game changer. While listening, I might stop the podcast, and record a few key points, but I often add my own connections and reflections triggered by the podcast (as above). Standing in a beautiful forest, and just talking, you can be very reflective and calm. And while importing it into Roam takes a tiny bit of time, it’s vastly faster than trying to transcribe the entire text.
For some podcasts, I might just stop once in the middle, or even just summarize the key points at the end (this is similar to how I will sometimes quickly record a few key points after stepping out of a cafe where I’ve been meeting someone, or quickly capture some thoughts while playing with my son). For others, like Awakening from the knowledge crisis, I record lots of notes and ideas.
I mentioned above that I’m curious about the learning effect of this approach, compared to the traditional note taking while reading or sitting in a lecture. The Awakening from the Meaning Crisis series was originally a Youtube series, and I could have watched it on my laptop with my notes open. Apart from not wanting to spend that much time in front of the laptop, I believe that the walking and the fresh air has a huge impact on my concentration and interest. Not being able to take notes contemporaneously also forces me to think about what is important, and reframe it - which is obviously positive. And making it so effortless to capture, makes me much more likely to “think aloud” in ways that I find really useful when I look back at it.
Processing the notes
Having captured the notes, I later sit down with Otter and Roam generally a few days later. A recent trick is to create a new folder in Otter for episodes where you will be taking a lot of notes, to make it easier to process all the notes from an episode at once. I go into each note, select “Export as text”, and copy to Roam.
I read through the text quickly to make sure it makes sense, and if there are obvious mistakes, I fix them, or click on the text in Otter to listen to what I said, and correct.
If I only took a few notes from a podcast, I might just add where it came from, and a few tags. But for something like Awakening from the Meaning Crisis, I try to organize my notes much more carefully. However, this can be a very gradual process in the spirit of Progressive summarization – waiting a few days between you listened to the podcast, until you import from Otter actually gives you another chance to revisit your ideas, solidifying your memory, and generating new connections. The same is true for waiting some time before going back to your notes to clean-up, reorganize etc.
Looking at my public notes for Awakening from the Meaning Crisis, you’ll see that I’ve so far only deeply processed chapters 1 and 2.
Chapter 2 is a good example of where I’ve used a query to pull up my own reflections and questions to the top.
Once the content is part of Roam, you can of course apply all kinds of other practices to it, such as daily retros, and spaced repetition, which I’m planning to write about in upcoming issues.
In Summary (h/t Alban Brooke):
Find podcast episodes on Listen Notes
Add individual episodes to a custom podcast playlist
Listen to episode while walking
Periodically record voice memos into otter.ai to capture thoughts in real time
Clean up the otter.ai transcripts when you’re back at home
Import the transcripts into Roam
Apply progressive summarization and organization, in the same way you’d process written notes
Minerva and the Corona crisis
One side-effect of the current crisis has been the millions of students around the world (including my six year old son) who have suddenly been thrown into distance/online learning, with very little preparation, and highly variable results. At the Minerva Project, where I work, we have been designing pedagogical approaches, custom technologies and infrastructure for effective learning for more than six years, in use at the Minerva Schools, and a number of partners around the world (like HKUST, Laurel Springs High School, Recruit (JP), and Berkeley Law (video)).
Because our aim is not to replicate what happens in a traditional lecture, but to go far beyond, our approach requires trained instructors (we have multi-week training programs for partner instructors), custom heavily designed lesson plans, and a systematic approach to designing the entire curriculum. This means that people cannot simply switch to Minerva, like they might switch to Zoom – however, we are talking with a number of institutions who are interested in longer-term partnerships to improve resilience during the upcoming fall semester.
We’ve also opened a new program for international students who are accepted by top universities, but are unable to travel to begin class in September. Through Minerva Schools, we offer them acceptance as Minerva Visiting Scholars, letting them follow Minerva’s first year curriculum taught by Minerva instructors remotely for the first year, and then transfer these credits in full to continue studying at their target institution.
Our faculty and staff has also been sharing their many-year experience with teaching online through a variety of media, below is a list of some contributions. If you are interested in understanding more about Minerva, feel free to reach out (you can simply hit reply on this email). One of my favorite showcases of our platform is during the first few minutes of this video. I also highly recommend our book.
"Six Simple Ways to Keep Calm and Teach On" by Kara Gardner
James Genone on Bryan Alexander's The Future Trends Forum: "COVID-19, higher education, and technology"
"Eight Ways to Avoid Crickets”(Using active learning to improve online teaching) by Josh Fost
Education for Versatility and Virtue - Josh Fost
Tips for Teaching Math (and other STEM subjects) Online by Rena Levitt, Lucas Tambasco, and John Levitt
Running Interactive Philosophy Classes Online by Alex Hyun & Scott Wisor
"You had to cancel your lab course. What now?" by Abha Ahuja
Reimagining workplace learning during COVID-19 by Gloria Tam
Roam community updated
The unofficial Roam Toolkit (Chrome/Firefox plugin) has launched. One killer feature is ability to type things like ;in three days; and generate a proper date link, also spaced repetition. Check out the walkthrough video and find it on the Chrome webstore
Ryan Guill has a tool for flattening Roam output for use in other apps.
Nat Eliason’s Roam Course. The original, endorsed by Conor. Great generalist content, covering a huge range of use cases.
Cite to Write by Cortex Futura. Focused on academic use cases.
Shu Omi’s videos. Not formally a course, and free, but 16 short videos with massive bang for your bucks.
Amazing Roam prototype by Dhrumil Shah (pick of the week! 🔥)
Meeting Notes in Roam (organizing 1:1s) by Conor
What's So Great About Roam Research? by Nat Eliason
Weekly Review in Roam by Anonym.s
Roam: The Best Personal CRM Ever by Nat Eliason
Tiago Forte’s Interview with Sönke Ahrens on How to Take Smart Notes