It’s finally happened. As many have feared, the bird site has been purchased by a “free speech absolutist” billionaire. The ugly side of the internet has responded in all its typical trollish glee. Meanwhile, we’re all left to wonder just what the future will bring to the site. Maybe it descends quickly into an unmoderated troll paradise, leaving a smoking $44 billion-sized crater, or it might stay business-as-usual for a period as it struggles to pay for itself before dying. I don’t expect the instant vaporization of the site, but time will tell.
Update as I write: Things move fast. Looks like management’s clearly signaling they’re going to try to squeeze revenue out of their biggest, most active and valuable userbase. Anyone’s who’s seen a beloved video game go into a monetization death spiral will be getting some very familiar vibes.
Update 2: Now VINE is on the table? My little confidence is deteriorating rapidly now as spaghetti is flying willy-nilly at the walls.
Top of mind for many of us is what will happen to Data Twitter. That amusing, eclectic, hyper-varied mix of people that cuts across all countries and industries imaginable, is one of the jewels of the data community. I’ve personally found a job after a Q4 layoff with their help and this very newsletter would not have existed without the support of many folks on it. I’ve also given countless instances of interview, career, and resume advice to random strangers that pinged me through it. No one wants to see such a great and welcoming resource go away.
The people continues to be what continues to be awesome about the group, but the platform provided the glue and interaction media that allowed it to thrive. The recent acquisition has highlighted how the platform itself is a massive single point of failure, as people start wondering how we’ll maintain the connections with the platform potentially gone.
As people who have to do occasional engineering and are sometimes tasked with building out resilient production systems, it probably makes many of us itch that there’s no practical replacement that will fill the void for the foreseeable future. There is no magical “build a duplicate Twitter for redundancy” option. So people, myself included, are currently cobbling together backup plans in case Twitter dies unexpectedly by sharing where they can be found on other social networks. There’s even some chatter about heading off to Linkedin and bringing the lighthearted banter and shitposting culture there.
Community can survive in the oddest of places, like in the comments section of a blog, or a small forum on some website, or in a chat server. But the platform dictates many of the available modes of interaction, so finding a replacement that satisfies both familiar modes of interaction as well as a similar population will be impossible for a while. We’re going to be in a salvage-what-we-can mode for a while.
Side note: I’ve cobbled together a very simple, publicly-editable google sheet with the handful slack/discord servers that I know exist. It's not exhaustive, but use it as a way to connect in case of emergency, and please feel free to add entries. Anything that’s remotely relevant might be useful to others.
Digital life has single points of failure everywhere
I like digital life much more than most of actual real life. But the one place that real life completely trounces digital life is how real life relationships tend to be multi-faceted and therefore so much more resilient to disruption compared to digital.
Think about all the people that you are friends with online. I’ll bet that the vast majority of them, there’s multiple points of failure that would make it so that you have an extremely difficult time reconnecting with any of those people. If an email address is stolen, a social media account deleted, a social network gets acquired and rapidly rots, how many friends and conversation partners would you lose? Hundreds, possibly thousands?
The massive reach and scale that digital lives gives us the ability to track and follow thousands of people simultaneously without having to remember their individual phone numbers or home addresses like back in the olden days. But that scale has also forced us to offload the cognitive load of even merely remembering the relationships to The System. My pea brain knows the majority of people on Twitter simply by the color and tone of their profile pic and I get confused when those change. As I try to move to other platforms, I have to constantly reference my existing connections to even remember who to follow again.
Moreover, these connections aren’t redundant like they would be in a real-life relationship. I can still physically reconnect with people I lose touch with in real life if they happen to live nearby — by virtue of bumping into them on the street or in a market or something (let’s assume I go outside for argument’s sake). I’m also more likely to have multiple ways with contacting real-life friends because maybe we’d go to the same club or coffee shop, or share the same interests and thus go to the same events. All of these “strategies” are heavily curtailed in the digital social landscape we find ourselves in.
So we need some redundancy in our digital lives
Luckily, since Twitter is merely dying a slow(ish) death and not an instantaneous one, we all have time to put into place backup plans for when it does become unusable. We can download our account data, save our connections, and at least get a list of people we need to find out there in the big open internet.
So long as we can maintain a handful of resilient ties to chunks of the community in various places, the forces of triadic closure will lead to reconnecting with more people over time if the whole thing explodes and we’re scattered to the winds.
For example, if you’ve managed to find and connect to some data friends on Linkedin, or join one of the handful of large data discords/slacks, you’ll be able to rediscover some other people that were active in Data Twitter by virtue of friend-of-a-friend type suggestion features. Those connections will eventually connect to part of the hive mind that will inform you if a new gathering spot has hit critical mass. It's an order of magnitude more difficult if you don't have even that tiny starting point for snowballing.
But for such a snowball strategy to work, it requires a minimum of two things:
People need to actively connect with each other on multiple services before a massive failure event wipes out existing connections data
People need to be willing to share their identities across multiple services, especially if there's pseudonymity involved
Many of us are luckily aware enough of the potential for the eventual failure of Twitter to take the opportunity now to share ways to connect. It was the kick in the pants that many needed to actually overcome the inertia. For you, pay attention and take advantage of the opportunity to connect! You’ll never know when it’ll be the last time you see a given user on a given platform, and if you value the connection, take it! If somehow Twitter makes a miraculous recovery, you’ll have at least made stronger links with a bunch of people.
Things get murkier when there's pseudonymity involved across networks. For example, if all of us decided to move to Reddit, the norns there is that everyone uses a handle instead of their real names, whereas the conventions of data Twitter were the reverse and real names were significantly more common. It's a different sort of place with different expectations where you probably can't rebuild the network there.
Either way, some people are willing to share their identities across platforms, and again it’s an opportunity that is not to be squandered.
But what about newcomers who won’t have a network?
Since there won’t be a social network to replace Twitter for a while, probably the best place to direct people who are new to data science is one of the bigger data slack channels. Those already have hundreds of people available who can engage with newcomers and provide the welcome that those folks need.
We’re all going to be adrift in little silos for a while until stuff settles down. A lot of people who remember the ancient days of Web 1.0, with the fragmented forums based on certain web sites and usenet groups will actually be pretty familiar with the situation we find ourselves in. We’re not going to have communities that we didn’t know pop up in our timelines thanks to The Algorithm anymore.
Luckily! The Old Fashioned way will still exist. Back in the olden days of Web 1.0, people stumbled upon communities by seeing references to them in other places. I found out that there’s a site and forum where most amateur gem cutters hang out on by seeing a mention of it on some random post somewhere. So if you can, help with the SEO and link to the communities you care about. We’ll be in this weird web 1.5 state for a until new gathering places get discovered. If all we have are a bunch of blog posts that point to everyone’s favorite gathering spots, then so be it.
For now, keep connecting to your favorite data folk wherever they say they’re going to be. Even if it’s a service that might fall out of favor in the near future, it’s just one more tiny thread that will keep our valuable social fabric together in times of need.
If you need to find me on Mastodon, I’m @Randy_Au@recsys.social. It’s a small Mastodon server lightly themed towards data folk like us and run by Karl Higley. Your server choice doesn’t stop you from following people on other servers and see their posts. While quirky, Mastodon does fill the need to have a microblog service that is seen an interactable amongst a number of friends and strangers.
You can also always find me on this newsletter’s Discord server, which is much more laid back and chill about everything.
Other news
Remember that post I wrote about how healthcare data could use some of our data science love? Well a few people seemed to have formed a nonprofit called payless.health around working on that data and have put out some bounties. If you've got some interest in doing some healthcare data science for some cold, hard, cash, take a look at their dolthub space, which includes some details about what they’re looking for and trying to build.
Standing offer: If you created something and would like me to review or share it w/ the data community — my mailbox and Twitter DMs are open.
New thing: I’m also considering occasionally hosting guests posts written by other people. If you’re interested in writing something a data-related post to either show off work, share an experience, or need help coming up with a topic, please contact me.
About this newsletter
I’m Randy Au, Quantitative UX researcher, former data analyst, and general-purpose data and tech nerd. Counting Stuff is a weekly newsletter about the less-than-sexy aspects of data science, UX research and tech. With some excursions into other fun topics.
All photos/drawings used are taken/created by Randy unless otherwise credited.
randyau.com — Curated archive of evergreen posts.
Approaching Significance Discord —where data folk hang out and can talk a bit about data, and a bit about everything else. Randy moderates the discord.
Support the newsletter:
This newsletter is free and will continue to stay that way every Tuesday, share it with your friends without guilt! But if you like the content and want to send some love, here’s some options:
Share posts with other people
Consider a paid Substack subscription or a small one-time Ko-fi donation
Tweet me with comments and questions
Get merch! If shirts and stickers are more your style — There’s a survivorship bias shirt!