I have periodically struggled with tab explosion and digital addiction. Over time, I have attempted a number of strategies to control this and use my time more productively. This has improved my quality of life and the quality of my work, though there remains substantial room for improvement.

Here are some of the things that have worked for me:

I used andreasgrill’s auto-selfcontrol (a Python script that automates SelfControl) to block a variety of URLs I deem to be of low value or to have high “black hole” potential. By automating and setting a schedule, I build times into my day that are open for me to check Twitter, Instagram, etc. while keeping the majority of time locked down. When I primarily worked on a personal laptop, I used RescueTime extensively as well to track what I was doing and analyze trends in productivity. Unfortunately, recent changes to macOS security have meant that automation of SelfControl is much more difficult than before and auto-selfcontrol no longer works.

I also applied the blocking approach to my mobile phone once Apple’s Screen Time became available. I am more stringent with my phone and have attempted to block all web browsers and news/social media apps, in the interest of restricting this content to browsing in front of a laptop rather than getting diverted by my phone throughout the day.

Twitter

Twitter is complicated for me, because it is a massive source of distraction and useless information but also a great venue for dissemination of useful knowledge. Without academic Twitter, I would probably encounter fewer research papers across a number of fields and have less insight into which ones were impactful. I would also have less context for what is going on in the world, and as someone who enjoys learning about politics and believes engagement with it is an important aspect of behaving ethically, I would be unhappy if I could not gain that context. Twitter also enables people from different disciplines to engage with each other about a common topic (though in too many cases this turns into a plethora of “Substack Matts” talking at each other).

For a long time, I had the perfect solution to this in the form of Nuzzel. I was able to route my Twitter consumption through the Nuzzel app, which would surface the most-shared URLs in my newsfeed to me and thus provide a much more digestible summary of what was going on but which I only needed to check periodically. This also allowed me to keep Twitter almost permanently blocked. Sadly, since Twitter bought and shut down Nuzzel, this option is no longer available to me and I have found myself using Twitter itself much more and/or missing out whenever I am detached from it.

Offload pages of interest into specialized apps and devices

At a high-level, much of what remains after restricting the lowest-value or highest-risk content breaks down into a few forms: books, papers, articles, videos.

I try to offload these from my browser as soon as possible into one of various specialized apps. If possible, I also try to move my consumption of these into an audio form (which allows me to superset reading with other activities) or onto a specialized device that is better-optimized for focused reading: my Onyx Boox Note 2, an e-ink Android tablet that has been extremely useful for my reading purposes. I like the tablet more than the alternatives because it runs a modded version of Android (so installing arbitrary apps is supported) and its native PDF reader is exceptionally good and highly flexible.

Books go to: various

Whenever I encounter a pointer to a book of interest, I save it in my Goodreads. I try have some high-level categories breaking things down by domain and by geography (if appropriate - as a large number of my books involve politics and history, tracking geography tends to help).

If I get a physical copy of a book and put it in my bookshelf, I use Libib to track its location. Each shelf of my personal bookshelf is its own entity in Libib, to make it easier to find what I’m looking for quickly.

For book drafts, digital books, and other book-like PDFs, I use git-annex to archive, version, and back up/replicate PDFs across multiple storage systems. A copy of my library is on Amazon S3 (plays nicer with git-annex than GCS) and another copy is on a 4TB external drive. This tool is amazing and I highly recommend it for managing one’s own multimedia. I own over 50GB combined of multimedia that I don’t need immediate access to; I can keep most of it on AWS and backed up to an external drive, rather than having to retain a copy on my local machine, thanks to git-annex. The one thing I would like is an indexing solution for git-annex archived PDFs; so far one doesn’t exist and searching in my digital library is challenging.

I have Scribd and Audible subscriptions and look for books on these services sometimes. Scribd is interesting because of its “Netflix for books” character, the existence of both traditional e-books and audiobooks in its catalog (I like its traditional reader flow but not its audiobook functionality), and the quality of its app’s reader flow on tablets. I found my subscription worth it. Audible allows me to download books for usage in venues other than a smartphone or laptop (in particular: I can download books to my Underwater Audio waterproof iPod Shuffle and listen to books while swimming), and has a larger selection of books due to some exploitative deals they cut with publishers.

I would love to use Libby more frequently but unfortunately the library often doesn’t have the books that I want, and the app does not work well on my Onyx.

Papers go to: Paperpile

For a long time I had no solution to the problem of storing my papers, especially since I began to boycott Mendeley due to its ownership by Elsevier. Eventually, I bit the bullet and began paying for a subscription to Paperpile. This is far superior to Mendeley, has an Android/iOS app that is pretty good, and has exceptional support through its Chrome extension and iOS app for adding new papers via URL. I highly recommend it.

When I discover a paper of interest, I immediately move it into Paperpile rather than letting it sit in a tab. I read my papers almost exclusively on my Onyx, where I can mark them up, add them to various categories, and mark them as completed or abandoned. Paperpile also allows me to bookmark a paper even if I don’t have a copy of the PDF and provides useful functionality for discovering such papers.

As with my Goodreads, I have grouped my papers into coarse categories and created different folders for each of these categories which papers can inhabit; they can belong to multiple categories at once (for example, something can sit in both “Machine learning” and “remote sensing”).

I’m a big fan of Connected Papers for paper discovery and would love to see integration of this tool with Paperpile.

Articles go to: Instapaper, Feedly, Reeder

Instapaper was the first “offloading” tool I used to get things out of my Chrome tabs. Instapaper has immense flexibility for extracting the article content from a URL and for saving both on laptop and on mobile. You can even archive text through email.

My one regret is that my Instapaper is not well-categorized at the moment. I pay for the service and have access to the API; I am hoping that I can use this to download the text of all my archived articles (approximately 5,000) and use topic modeling to create some of these categories.

Instapaper also has a decent text to speech converter and a speed reading functionality.

On my laptop, I’ve occasionally used Reeder as a client of my Instapaper, though I’ve found it to not play nicely all the time with the API. Reeder has a nice UI and has some additional reader support tools I’ve played around with like Bionic Reading.

I also use Feedly to subscribe to newsletters when possible; I can also add this subscription to Reeder above. In practice, I only occasionally read newsletters.

Other PDFs go to: a folder in Dropbox

There are a number of other PDFs or true “documents” I encounter that aren’t books or papers (for example, slide decks, informal lecture notes, syllabi, dissertations, CVs). These for now are being put in a “Miscellaneous Documents” folder in my Dropbox.

YouTube videos go to: Watch Later

I actually don’t watch a lot of YouTube videos; I tend to be more audio than video for historical reasons (see below). I’ve realized over time however that this has made me fall behind on more technical areas, especially math, physics, and more theoretical computer science, which require some visual involvement to be effectively communicated. As I’ve tried to learn more about ML on graphs, for example, I’ve discovered a number of YouTube videos of interest. I’ve begun archiving these in my “Watch later” queue, but will need a better approach soon.

Other URLs go to: an Evernote page

The number of URLs of interest that are not one of the above categories has numbered on the order of 50-500 over the past 6 years; this is manageable enough that I store these exceptions in my Chrome “read later” queue immediately and in an Evernote note over time.

A long time back, I used to use Kippt as a generic bookmarking service for non-textual content. Unfortunately the company went under before I had a chance to validate the service.

Preference for audio

I find that I learn better from audio than from video. Thus, I try hard to channel my content consumption into audio. I use three podcast apps for different purposes.

Overcast

I use Overcast for daily podcast consumption. You can create an ordered set of playlists, and the app will automatically choose the first non-empty playlist as your “queue”. (There is a playlist called “Queue” that always sits at the top of your playlists, but if it’s empty then the next non-empty playlist will be pulled up.) I use this to listen to the news and to any podcast that I want to heavily prioritize.

I used to use Overcast for all podcast consumption, but unfortunately it did not scale to my very large number of subscriptions (greater than 500). A potential solution is to reduce my total number of podcast subscriptions, as having so many is absurd.

Pocket Casts

I use Pocket Casts to subscribe to a large number of podcasts and get random access to their episodes. I use this to set a “discretionary” playlist that I dip into periodically throughout today, with no expectation that I will finish the entire playlist. Anything I didn’t listen to or only partially listened to will eventually be archived by Pocket Casts’s settings. Pocket Casts has only one queue, which one must explicitly add or remove episodes to, but many “folders” into which incoming episodes can be slotted.

Google Podcasts

While Google Podcasts leaves much to be desired, it is exceptional at podcast search. Whenever I am trying to discover a discussion of a new topic or interviews with a person of interest that none of my existing subscriptions have helped me with, I use this tool to search for relevant keywords and usually find something of interest.