Notes (sigh) on a Tweet

How to use social media posts responsibly ... and how to do the opposite

One of the topics I addressed in my German-language Cancel Culture Transfer and largely left out of The Cancel Culture Panic is the question of how specifically legacy media do/should engage with social media posts. I do talk about it a little, of course, but in the German book I tried to walk readers through how articles about "cancel culture”, especially in Europe, move directly from a tweet some person sent to these really really big claims about “the Left”, “the Culture”, “the Discourse” blah blah bah. And how they do so, as far as I am concerned, way too directly. The idea that “social media users” are doing something as a collectivity (the vaunted “online mob”) is both super important to “cancel culture” and anti-wokeness discourses … AND immensely slippery.

Part of why I excised that portion in the English edition is that I found this tendency was more widespread in German media than elsewhere. I don’t really know why. There are some famous instances in the US — Brett Stephens evidently searching for his name on Twitter and going berserk when he found a tweet that made fun of him, which had received something like nine likes. You might say: well, Adrian, it would be pretty weird if this was something German speakers did more readily than others. Sure, but I’d say this is probably not about authorial intent, it’s about editorial standards. I’m sure desperate writers who want to prove that someone somewhere thinks a particular thing, but who have a firm deadline to hit, google or search social media platforms for the relevant keywords all the time all over the world. The point is that any good editor or fact checker would be like: hey, buddy, love the effort, but are we sure this counts as evidence?

Because that’s the problem with the language game “on social media”: “on social media” users celebrated, criticized, fulminated? Well, yeah, sure. Someone did. There are hundreds of millions of tweets put out every single day, and I’ve heard that it’s more than a billion posts on Instagram. You’re gonna find just about any two word combination in that immense discursive profusion — so when you’ve found it, you’d have to ask: how relevant is this, how representative, or representative of what. This is the questions that a good editor pushes a writer to ask. And one that German editors as a rule seem to push their writers on less than others.

Let me show you an example from today — so, yes, this is a post about a single tweet, a sub-stack of a tweet, if you will. But hear me out, this isn’t about this tweet. It’s about how that tweet is used as evidence. It’s about how legacy media reflect on social media and the editorial process that should come to bear when it does so.

Der Tagesspiegel is a center-left (I think? Basically?) daily based in Berlin, Germany. This is the article that caught my attention this morning.

"Supporting Hamas publicly used to be unthinkable. But now the haters of Israel have lost all shame and say openly want they want.” Here’s the article (which is an installment of Leber’s regular column):

The title of the piece is: “Where is the firewall against the Islamists?”

I should explain the framing: The “Brandmauer” (“firewall”) is a metaphor German political parties use when discussing the relationship to Nazis, Neo-Nazis and specifically the far right Alternative für Deutschland (AfD). It refers to the tendency to basically not give an inch to the far right — not to strategically cooperate on a few select issues, for instance. Which already makes this framing fairly remarkable: the article doesn’t really say “individuals are parotting Hamas talking points”; it’s saying “there’s an institutional tendency to not sufficiently reject Hamas talking points”. You may think that this is true, but I hope we can all agree that this is a pretty big claim.

Here’s the first piece of evidence we get for deep and abiding sympathies for Hamas among online leftists — it is in fact the opening of the piece:

“The latest perfidy is the comparison with Nelson Mandela. After all, he was called a terrorist for a long time, but later he was celebrated as a hero. That's why it's better not to call Hamas fighters terrorists.

One could take the trouble to explain that Mandela did not have people murdered, did not order massacres or suicide bombings, was neither a religious fanatic nor an anti-Semite nor a gay hater and did not have rivals thrown alive from high-rise buildings. You can also simply block people who make such absurd comparisons on social networks.”

And yet, I would note, Sebastian Leber did not do that. He did not block, he screenshotted. Which is fine, of course. But I’d point out that that’s a very different online behavior: he’s evoking this image of himself beset on all sides by crazed leftists posting their Mandela-comparisons and Che Guevara memes, so many that even with all his blocking he still has to see their stupid opinions. I understand that everyone social media feeds are different, but I can honestly say I’ve never seen any of these things. So something about the pose is a little strange: he never outright claims that there’s a ton of this discourse online, but his invocation of “block[ing] people who make such absurd comparisons on social networks” does suggest that there’s a fair amount of it. Enough to show up unbidden in his feed. It’s a kind of thing that comes into your feed and you’re like “ugh, not again.”

The first example we get of this proliferation of Mandela-Hamas comparisons is the following tweet by a Twitter user named “Eloy_007”.

Seems pretty straightforward. It’s — the caption notes — an “example for a nonsensical equation in online debates”. Those darn online debates! With their nonsensical equations!

You can probably see where this is going if you’ve spent more than a microsecond on social media, or have spent any time reporting on social media. Because you will have thought: “Oh, weird, why didn’t they leave on the timestamp and the number of likes and retweets?” It’s particularly unfortunate as it would give us a better sense of Leber’s thesis: how widespread is this nonsense? How did people in this user’s online environment react? Who is “Eloy_007”? Is he a public figure, does he have a lot of followers, is he an organizer, etc. etc.? And what did the replies look like to a tweet like this? Feels like that would give us a fuller picture of how widespread these “nonsensical equations” are in “online debates”! Curious online debaters would like to know.

“Eloy_007” appears to be a Denver-based Marine corps veteran whose daughter just graduated from UC Boulder (congrats!). He joined X in early 2024. His feed is indeed pretty rough. (In fact, if you’re wondering why I cut off the top of his profile — it’s the slogan “Stop the Genocide” alongside what I think is a picture of a dead child, which I don’t feel I should reproduce in dunking on a news article. Feel free to find his profile and check my work.) This user retweets a lot of information about the war in Gaza, and doesn’t seem particularly discriminate about its sources. I didn’t go down every rabbit hole, but he certainly might want to fact check a lot of the claims about Israel and the war he retweets or puts out there. But he also — and that seems to me significant — has as of 05/13 a grand total of 321 followers, including at least five that very much look like crypto-scammers. So it’s not what you’d call a high-impact account. And I think that before I demand that a Denver-area dad do better fact checking, a Berlin-based daily might fact check that sort of thing.

Still, perhaps “Eloy_007” went viral with this one tweet! His followers (insofar as they are not of the “my tits in bio”-variety) are all clearly posting about Gaza, so you could see how his reach might be bigger when networked with a group, however small, of likeminded followers. It took a lot of scrolling to find the relevant tweet that Der Tagesspiegel reproduced, and no points if you guessed the obvious next step in my argument.

You read that right: zero likes, zero retweets (as of 9:30 am PST Monday, May 13). And even I — who is essentially banned on Twitter/X for inactivity and having my Bluesky handle in my bio (I think? Or I’m just super unpopular?) have never seen an impression count that low. Thirty! Just thirty users even had to swipe past this post in their feed? Impression figures tend to be incredibly inflated (and probably bogus) — thirty frankly feels like its own kind of accomplishment. Is it an “example”? I guess in some very formal sense it is: it exists in the world and it does do the thing the author says some people do. Fine. But as an introduction to a phenomenon that is about collectives and groups — about the left, about protestors, about “online debates”, etc. — the fact that no one has seen this tweet, does seem to impact its ability to serve as an example.

And more to the point: take another look at the dude’s followers: his followers are mostly virulently anti-Israel voices who post almost exclusively about Gaza. And yet none of them retweeted this! If anything, that might be remarkable. I won’t say it is because X’s algorithms are so fucked these days that it’s quite possible they were simply never shown this content. Which is why it’s so damn hard to use posts on X as evidence of anything (unless they are from very popular users or associated with official groups)! Sebastian Leber was one of the 30 people who had this in their feed at all, and his editors didn’t bother to ask him whether that was really a strong piece of evidence to open an article about how lefties were a bunch of Hamas-lovers who can’t do historical comparisons right.

Now, if you’ve been on X a bit, you’ll have been screaming the next point at your screen for the past few paragraphs: This was never in Leber’s feed. He searched for “Mandela” and “Hamas” and screenshotted the result. How do I know this? Because “Mandela” and “Hamas” are boldfaced in the screenshot reproduced in Der Tagesspiegel. Those indicate search terms. (I guess this could be his editors searching for the tweet to screenshot it, but … in that case wouldn’t they have seen the number of likes and been like “uh oh”?) Just for future reference, guys: if you x out your search term and reload the page, the bold facing disappears.

I think this matters: Leber wasn’t horrified by what social media algorithms washed into his feed and wrote a piece in reaction to it. More likely he wanted a tweet that made a specific comparison, and searched for it. This is what came up. There are other hits you get when searching “Hamas” and “Mandela” on Twitter (many of them now the Tagesspiegel-article) — some of them jokes, some of them dumb, some of them trollish, some of them hot flaming garbage. The point is: using these things as evidence is pretty tricky. What’s said on social media is not irrelevant or not worth reporting on. But what you see there is also not as straightforward evidentiary as what — and just taking a random example here — you might read in a paper. A tweet may reflect a broadly held belief, and it may indeed be news. And there’s no hard and fast rule for saying when it starts being so. But I do think some basic reflection on where that line is, is key when reporting on “online debates”. And Mr. Leber, who according to his bio on Der Tagesspiegel, “spends too much time on the internet”, should probably spend more time figuring out where that line is.

In some way, reporting or reflecting on online discourse always risks having the same Achilles heel as everything digital: the danger of narcissism. There’s the risk, in other words, that what you meet in the infinite expanse of the Internet is simply yourself, again — your obsessions, biases, petty grievances. If 10,000 accounts repost an open letter, is it newsworthy? The deeply frustrating answer is: It may be, but it may not be. Still, it does seem important that we be able to demarcate a zone of irrelevance. A zone where someone saying something cannot count, one way or the other, as evidence for anything. And reporting on that zone, even though you know how recherché the niche you’re drawing from is, I fear, everything that reporters like to accuse online debates of: it’s tendentious, myopic and dangerously close to disinformation.

Update 05/14: This post got some traction on Bluesky among people who seemed to think that this wasn’t really a big deal. (I’d by the way agree that this isn’t a huge deal — it is a pretty perfect example of a problem we all struggle with when talking, and especially writing professionally, about the internet and social media. That’s why I wrote about it.) Still, I feel like I need to clarify something: my point isn’t that picking out a single tweet out of a large group of tweets is somehow illegitimate. You might still want to think about what it means to pick on a private individual in this way, especially if you have a huge platform — but that’s between you and your maker. My point is that the fact that the best example of a discourse you can find on Twitter has zero retweets and zero likes might indicate the discourse isn’t as widespread as you’re assuming or making it seem. That doesn’t mean you can’t talk about it, of course. But then you’re kind of talking about an online curio (“get a load of these guys who think xyz!”). When I was researching early uses of “cancel culture” on Twitter, all my evidence looked like that — someone confidently using the word and like two people reacting at all. And I pointed that out in the book.

Anyway, after people got mad at me I decided to grab all the tweets that seem to make the Mandela-comparison as of Monday afternoon PST. Because I am ninety thousand years old, I did it the old-fashioned way and just screenshotted everything. I found about 60 tweets. You can find them all gathered in a Google folder here. I’d invite you to click through them — many of them are pretty crass, but nothing that I think would get anyone in trouble in their jurisdictions or turn their stomachs. And I’d invite you to ask yourselves: does this have the gestalt of a significant trend? The likes and retweets are reliably zero or close to zero — the most popular of the tweets had like 20 retweets. Again, you might read this data differently than I did. But I think it’s good to document what we’re talking about here. And the question I’d leave you with: what are other discourses we could probably find this kind of documentation for (60 tweets, low-follower accounts, bizarrely low engagement) — and would we judge those discourses to be significant? Worthy of our attention?