09 November 2021

Protocol: “Spotify for readers: How tech is inventing better ways to read the internet”

But if you really want to understand where reading is going, you have to talk about Spotify. The big green music app is the analog nearly everyone in the space seems to turn to for guidance. Springwater mentioned Spotify early and often in our conversations. So did Jeroen Seghers, the founder of an app called Upnext that aims to offer this kind of experience for articles as well as podcasts and videos.

After all, what does Spotify do? It takes a corpus of stuff (music) and finds endless new ways to show it to users. Users can save the stuff they know they like (a library), explore things curated by other users (playlists) or turn to the app’s machine-learning tools for ultra-personalized recommendations (Discover Weekly and the like).

So now imagine a reading app. You can save all the articles, tweet threads, PDFs and Wikipedia pages you want into your library. You can follow other users to see what they’re saving, or check out what a curator thinks you might be into. The more you read, the more the app begins to understand that you like celebrity profiles, you’re learning a lot about NFTs right now, you worship at the altar of Paul Graham and you’ll read anything anyone writes about “The Bachelorette”. Now, every time you open the app, it’s like a magazine made just for you.

David Pierce

Some interesting concepts in this article, such as being able to read podcasts (their transcripts anyway) and highlight sections just as you would do with a written article. Other proposals sound rather awful though: a daily email reminder with random highlights from past articles feels completely useless to me, an annoyance I would immediately turn off. For context, some of these startups are invite-only betas, or available exclusively on iOS, so I would wait until they manage to become profitable and expand to other platforms before even considering them.

As for the analogy with Spotify, I feel that it is missing the mark quite a bit. Music listening and reading are two rather different activities, and people engage in them for distinct reasons and at different times. Music is intended more for unwinding and detaching from daily routines, or it may accompany chores in the background, whereas reading tends to be more intentional and focused – you can’t usually work or cook while reading an article, unless you do a poor job of both. In-depth articles and fiction also take much longer to read – dozens of minutes to hours and days – compared to listening to a song, which is over in three to five minutes. That’s why people will tend to be more careful and discriminate when selecting something to read than what music to play.

The algorithmic model Spotify uses for playlists and recommendations works because Spotify has access to virtually all digital music tracks and the cost of failure is rather low; if you don’t like a song, you can skip it or wait a couple of minutes for it to finish. The barrier to success for reading is much higher, and the pool of available content much smaller, as I doubt these new apps have deals with most online publications to skip their paywalls and provide unrestricted access. In conclusion, I don’t think comparing this to the streaming music model makes much sense.

Personally, I have been using my Kindle as a read-later device for years already, and I don’t see any of these startups improving on my existing workflow. It involves some manual work to send articles to the Kindle and then retrieve my highlights, but the reading experience makes up for it. There are several ways to deliver articles from the web to the Kindle, and, depending on the website where the article was published, I sometimes must try a couple of them before succeeding:

  • The first option, and the most straightforward, is the Send to Kindle Chrome extension. It strips away most extraneous content and sends the current article directly to Kindle, but unfortunately it fails under some situations. When the article contains many high-resolution pictures the delivery fails, possibly because of the large size of the resulting file. With some publishers, the extension also misses portions of the article, probably because of the complicated underlying HTML structure.
  • The second thing I try is to pass the article through an online reading service such as Instapaper or Outline. They have a better success rate generally, but sometimes can also skip sections, so I try to check at least the opening and closing lines to see if they match the original text.
  • Feedbin, my Google Reader replacement for all these years, added a ‘Pages’ section a while back, which functions as a read-later service through a bookmarklet. Once added to Feedbin, I can then use the Share menu to send the text to Kindle. While Feedbin is the best at capturing the article contents without gaps, I tend to use it as a backup option because the file downloaded on my Kindle disregards my local font settings for some weird reason, and I don’t care for the Arial-like font I am forced into.
  • As a last resort, you can also manually copy the text into a Word document and send it by email to a unique Kindle address, from where it gets automatically converted and delivered to the device.
A screenshot of the Kindle Library
Another tip from a reddit thread: you can take screenshots on the Kindle by tapping two opposite corners of the screen simultaneously

One downside of the Kindle solution is the lack of a good offline Windows app. The native app doesn’t show documents, only books purchased from the Kindle store, so I can’t read my articles while on a trip on my tablet, I have to bring my Kindle along. I have seen some people mention that the Kindle Cloud Reader could be installed as a PWA in Chrome or Edge and used to download articles offline, and I’m definitely going to try this the next time I have a longer trip planned.

Post a Comment