Reorganizing my Twitter

July 26, 2020 8:45am

I've reorganized my Twitter and the way I use it. It became really hard to keep track of multiple topics of interest that are so disjoint. I always had the feeling of missing out on the important tweets and depending too much on Twitter's recommendation algorithm and that lack of control was unnerving. But the problem is that switching Twitter timeline to show sorted by date meant that I had to go through a way, way too much information, even if I'm following (only) ~350 accounts.

I split much of people I follow (and more) into lists (all are public and will be updated over time): https://twitter.com/abstractalgo/lists and got back to https://tweetdeck.twitter.com/ (if you didn't already know about it, it will literally change your life). This process has been more painful than I imagined or desired. Twitter's lists management features are a disaster, a lot of clicking and precision selecting. I can't imagine doing this with like 1000 followed people or more. I even signed up for Twitter Developer to create some simple app to allow me to bulk-manage lists, because I found no good product for that.

Besides creating lists, I also added several search queries, such as #openscience OR #openaccess OR #OA and "reference manager" OR @mendeley_com OR @zotero OR "reference management". I will update and refine these over time as well.

One thing that I became aware of very quickly is existence of very active accounts and hashtags and the problem they introduce. For example #AcademicChatter and @AcademicChatter are very interesting things to follow and stay on top of, but the amount of content they generate is just too massive, they bury the timeline way too much and practically spam and obscure everything else in it.

So, current state is that I'm at ~12 TweetDeck columns. I have no fear of missing out, I can filter content much more quickly and precisely, and by traversing a single stream of tweets my brain experiences less "context switching", by consuming the tweets all on the same topic, and that has helped me a lot. I actually realized that I wasn't missing out that much (if I was missing anything at all), but I am happy that I was able to confirm it by having a total control. Another thing that is helping a lot is the "Clear" option on TweetDeck - it allows you to basically clear the timeline (each column separately) up until that point in time, so that you can progressively consume the timelines throughout the day and be sure that you haven't missed anything.

Now that problem of "consuming the Twitter data (by topics)" has been resolved, there's still a problem of generating the data (by topics). My followers range very different areas and interests, just as the people I'm following do, from computer graphics, front-end programming, interesting low level coders to VC and startup community, designers and artists, "second brain" community, people in open science and AGI researchers... Now, each group offers something that I need, or find helpful or important - a creativity, original opinion, beauty, high quality opinions,... you name it. While I managed to separate topics into certain groups, I wonder how people that follow me can make distinctions in the content I tweet. Yes, not every account is precisely specialized in tweeting about exactly one topic, but that is a problem to a certain degree.

I thought about creating multiple account, i.e. "Dragan Okanovic #AGI", "Dragan Okanovic #OpenScience",... but the amount of overhead is just too big (even with TweetDeck) and doubtfully useful. Because my primary job is not social media management, I don't think that it's worth it, so for now my followers would have to put up with me tweeting about science (computation and AI), meta-science (open science and related software), startups and programming, and occasional beautiful art. :)