Werden wir Helden für einen Tag

Home | About | Archive

My Twitter addiction

Posted on Jan 30, 2022 by Chung-hong Chan

Episode #1: The demise of RSS

I think two of the most important historical moments that most people don’t consider important enough are the slow, painful and gruesome death of Bloglines (circa 2010 to 2015) and the rapid sudden death of Google Reader (July 1 2013). In this blog, I didn’t write about the demises of Bloglines and Google Reader. Back then, I could not understand the significance of the two events. At that point I had used Google Reader for a very long time and I can still remember how cumbersome it was when Google Reader is no longer available. I read blogs from there, I read tech news from there, I even browsed Flickr photos from there (Apple called it Photocast), I kept track of search results from there. I read tweets from there at some point too!

For a short time, I switched to another online RSS reader called Feedly. During the short window of Google’s announcement (March 2013) and the shutdown (July 2013), I managed to export all of my RSS feeds as an OPML file and then imported all of them to Feedly. But the habit didn’t stick. I perceived at the time that the (tech) world has decided to move on with RSS. The “next generation” of thing written on the wall is social media.

Yes, social media had been quite big already in 2013. Facebook acquired Instagram in 2012 and iPhone 5 was popular. Honestly, there should be no causal relationship between the demise of RSS and the boom of social media since 2013. But I still think the demise of Google Reader should be considered one of the turning points of tech history in the 2010s. Please remember: in 2013, Arab Spring was still fresh in the air. Obama started his 2nd term. Many scholars still attributes the winning of Obama to his social media campaign. 1

Now we know what happened after 2013. Twitter stopped providing RSS in 2013. Even browsers such as Firefox stopped supporting RSS (feed detection and feed reading) in 2018. The number of Facebook users has doubled since 2013. (Q4 2013: 1,228 millions; Q3 2021: 2910). Tik Tok was not a thing in 2013. Algorithmic sorting of timeline was only introduced in Instagram and Twitter in 2016 (yes, it was not a big part of the two platforms at the beginning). Instagram replaced Flickr. Facebook and Twitter replaced blogs and basically any form of writing, including proper journalism. The term “fake news” was not used in our language in 2013.

All human attention is now concentrated (or captivated) in those social media platforms. And it creates a lot of problems.

Episode #2: Twitter addiction


Source: TED

Media scholars often find something like this: with self-report data, they find no (or very tiny) media effects related to social media. With empirical digital trace data, they find effects. This conflict leads to constant argument among media scholars about whether things like Filter Bubbles or Echo Chambers are real. To me this kind of academic argument is quite meaningless, because we can instantly resolve argument by seeing things from the platforms’ perspective. Facebook used to publish a lot of research papers in those big journals. I think the most important two are Bonds et al. (2012) (manipulation of emotions) and Kramer et al. (2014) (boasting voter turnout through the “I Voted” button). Yes, the effect sizes are tiny but very significant, due to its extremely large sample size. But the records are pretty straight: even Facebook in its 2013 scale (half the size of what it is now) can have the ability to swing real life outcomes — political outcomes even — if they want to. Just remember, they always know better about their platforms than any of the independent scholars who participate in those ivory tower debates.

There are actually some scholars seeing things from the platforms’ perspective. In her TED talk, media sociologist (now more like an epidemiologist) Zeynep Tufekci said we’re building a dystopia just to make people click on ads. We might not have clicked on online ads on social media. But certainly, social media change us. Or at the very least — I admit — me. In my case, negatively. 2

After saying this, I think it’s time to talk about the real subject: I was addicted to Twitter. I spent pathologically a lot of time on social media: first Facebook and Instagram, and then Twitter. It took me a lot of courage to admit this thing publicly and I am ashamed of myself. Admitting this would be perceived as having a weak mind and it is socially desirable to understate the influence of social media on one’s life. That’s probably why the self-report studies usually can’t find a meaningful media effect. But please be realistic: facing the powerful social media platforms, we are vulnerable to being manipulated by them. They have collected a lot of data about us and probably they can generalize verifiable rules about us better than we do. They are not only good at software or having a lot of data, they also know how our “wetware” works (more on this later). Not only having the best computer scientists, they also have a large team of psychologists and computational social scientists, probably all the brightest minds walking on this earth. Their software makes the best decision based on the persistent data about us; the persuasion architectures are (almost) perfect when it comes to making us addicted. We are also like to downplay the effectiveness of algorithmic solutions. The most famous one being there is no AGI (Artificial General Intelligence). Well, making us addicted doesn’t need AGI. Making us addicted is a very well defined task which machine learning systems excel at; surely easier than a winning game of Go against the world best players. Any drug lord would say it is fairly easy. The command-and-conquer task is even easier when our “we-think-it’s-unique-and-wonderful” GI (General Intelligence) — our wetware — is imperfect. We made a lot of stupid mistakes, my decision to use social media too much being one.

In May 2018 I killed my Facebook account. Last year, I stopped updating my Instagram. I used my will power (my “wetware”) to stop myself from using these two. The key is I actually don’t care about what’s going on with the people and the media outlets or whatever on Facebook and Instagram. Why did I spend so much time on the two?

But there is still one single platform which I can’t quit, and it sort of filled the void created by leaving Facebook and Instagram. And that’s Twitter.

Unlike Facebook and Instagram, I thought I care about what the people say on Twitter. There are (so) many researchers on Twitter. I made an excuse that I need to keep track of what’s going on in my field through the window of Twitter. Also, I thought that I need to “promote” my own work through Twitter. With these two reasons, I legitimized my usage. But that’s just rationalization.

Just like any addiction, there are underlying emotional problems. Addiction is a symptom of my problems. Upon self-reflection, the addiction is a product of my frustration with my (insert any negative adjective here) academic career. My stagnant but “can-end-any-time-soon” academic career means I desperately need external validations. Twitter, unlike other parts of academic life, is instant gratification and thus it felt so good to be there. And that’s how Twitter defeated my wetware. Upon meta-reflection and I tended to forget this: my wife had told me this problem long before my self-reflection and had suggested me many solutions. But I dismissed her for so many times with the above rationalization that I need Twitter for my career.

The previous satire is a self-criticism of how much I despise my attention seeking behaviors on Twitter. Moreover, the addiction also means Twitter sucked up a large portion of my time. I tried to count the time I was on Twitter and the bare minimum was 1 hour of my wake time per day. It could easily be 2 hrs or even more. Academic people are famous for being “busy”. What if I can liberate the hours I’ve spent on Twitter to something more productive, let’s say my long stagnant book project or any other hanging project that is half-baked? At that point, it’s clear that my Twitter usage dragged my performance. It hurts me emotionally too.

I don’t want to be the one
The battles always choose
‘Cause inside I realize
That I’m the one confused

I don’t know what’s worth fighting for
Or why I have to scream
I don’t know why I instigate
And say what I don’t mean
I don’t know how I got this way
I know it’s not alright
So, I’m breaking the habit
I’m breaking the habit tonight

And then it’s time to break the habit.

Episode #3: uBlock & Leechblock

There are many ways to make the cut. One way is like how I did to my Facebook and Instagram. But for Twitter, I can’t just use my wetware because the pull is too strong and I’ve lost hope on my wetware. I need software too. Due to the pandemic, my iPhone is mostly used as a timer for my cooking. The major problem is my working computer and that’s the area I needed to work on. Actually, I’ve made several changes to my computer related to this. But I will focus on measures against Twitter in this post.

I’ve watched the presentation by Luke Westby at Rustconf on being “Extremely Offline” using Rust. I’m then interested in blocking the web. It was a gradual process. My first step was to use uBlock to block the data tracking and any part of any website that I found problematic. If you use Twitter (the ordinary Twitter interface), I highly recommend blocking the stupid but unswitch-off-able “Trending” sidebar using uBlock (as of writing the CSS element is: twitter.com##div.r-1udh08x.r-1ifxtd0.r-rs99b7.r-1phboty.r-1867qdf.r-k0dy70.r-1ysxnx4.css-1dbjc4n:nth-of-type(3)). During the pandemic here in Germany, that sidebar is constantly infested with disinformation and anti-vaxxer’s hoaxes. These information made me anxious and angry, it affected my emotion.

And then I stepped up the game and added another blocker: Leechblock. This is a real game changer and it saves me.


The "doomsday clock" - 2 minutes

Leechblock blocks time-wasting sites, as simple as that. I have experimented with many ways to block Twitter. My final verdict is to allow 2 minutes of usage every 3 hours. Therefore, the total number of minutes I can spend on Twitter is \(\frac{24}{3} \cdot 2 = 16\) minutes. I deliberately break it into super short period of 2 minutes so that I can’t engage with the site for a long time. I also can’t doomscroll and just briefly skim through a few tweets at most.

You might argue: it is extremely easy to circumvent this. Uninstalling the two blockers or even resetting those rules can easily revert this. That’s certainly true. I think I can apply Professor Margaret Roberts’ (the “Roberts” in “King-Pan-Roberts”) research here. Any (technical) information control system is porous and there must be a way to circumvent it. The (real) information control is actually through three Fs: Fear, Friction, and Flooding 3. My reflection above says how much I fear the damage of my Twitter addiction. Applying the blockers can create friction. And more importantly, I need some other things to “flood” (a.k.a. distract) myself from using Twitter. I uptake many projects, such as intensive writing, programming, and reading. Or listen to music or playing videogames with my wife.

With the sharp cut in my Twitter usage, my productivity soars. I have more time to work on things I always want to work on. But then come another problem: how can I keep track of research papers and news? I don’t want to legitimate my usage of Twitter with this again!

It’s time to revive my old friend.

Episode #4: Make RSS great again



Rather than curating the curators to curate information for me, I rather curate the information myself again like the early 10s.



I dusted off my old OPML file 4 and set up Elfeed on my own machine. I added some scientific journals’ RSS feeds there too. It is not easy to discover those feeds. In some cases, I need to pick a feed from the source code.

There is no more excuse to be on Twitter “to keep track of research”.

Postscript: the Q & A

Q: Why don’t you just simply delete your Twitter account like your Facebook account?

A: I might, but not at this point. Keeping the social media presence, unfortunately, is strategically important for my career.

Q: How are you going to promote your work on Twitter?

A: Unfortunately, many academic people are still on Twitter and thus I still need to tweet my published work. I will use that 16-minute budget.

Q: How about I’ve published something in a journal that doesn’t provide RSS feed? Does it mean it can’t reach you?

A: I am glad that you asked. Unfortunately, it can’t. In general, if you think your paper is interesting and you want me to read it, please contact me by e-mail.

Q: It is pretty easy for me to stop using Twitter. There is no need to take these drastic approaches.

A: It works for you but not for me. Wetware is not universal.


  1. So do 2016 Trump, but in a different way. 

  2. It is not my intention to talk about the academic debate related to the media effect of social media. Saying there are negligible effects from social media is self-delusional. A d = 0.01 effect size found by digital trace data, when it is happening in a 2.6 billion user base of the 7.9 billion total world population, is a big effect in terms of people affected. 

  3. Roberts, M. E. (2014). Fear, friction, and flooding: Methods of online information control. Harvard University. 

  4. While reinstating the OPML, I also checked how many of the feeds are still reachable. Actually, most of them are dead. Almost 90% of the Hong Kong blogosphere it was is not longer there. 


Powered by Jekyll and profdr theme