Recently, like a lot of journalists, I joined Bluesky, a social media platform that is enjoying a burst of postelection growth and positive press attention. It’s been lauded as a “kinder, gentler”—and, perhaps most importantly, more left-wing—alternative to X, which is increasingly seen as infested with what a Bluesky user might call “MAGA chuds.”
While I thought some of the critiques of X were overstated, over the last six months or so I’ve increasingly soured on it. It felt like an ever more hostile, hateful place, the technology seemed more broken every day, and I am not a fan of owner Elon Musk’s recent conspiracy theorizing and all-in support for Donald Trump. It seemed like time to scope out a potential alternative.
This was a mistake.
On December 6, I made my first post on Bluesky—which was actually launched by Twitter in 2019, before becoming an independent company two years later. As I soon found out, it is an exceptionally angry place. And in part because of a widespread culture of impunity when it comes to violent threats among some of its users, it comes across as a potentially dangerous one—in a way X, or Twitter, never did for me in my decade-plus of actively using that platform. Bluesky has either made a conscious decision to take a laissez-faire attitude toward serious threats of violence, or its moderators are incapable of guarding against them, or both.
There’s at least some evidence for the latter theory. While many left-wing people announced they were leaving X after the election, one million users joined Bluesky that week. The results weren’t pretty. As The Verge reported on November 17, “the Bluesky Safety team posted Friday that it received 42,000 moderation reports in the preceding 24 hours.” That’s more than 10 percent of the number received in the entirety of 2023, which was 360,000.
But given what I’ve learned about Bluesky’s “moderation” over the last week, I feel compelled to inform the site’s users—and potential users—about its staggeringly negligent policies toward violent threats and doxxing.
The background here is that a subset of users on Bluesky disagree with my reporting on youth gender medicine—a subject I’ve been investigating for almost a decade, and have written about frequently, including in The Atlantic and The Economist. (I’m currently working on a book about it, commissioned by an imprint of Penguin Random House.) I’m not going to go deep here, but I’d argue that my reporting is in line with what is now the mainstream liberal position: See this Washington Post editorial highlighting “scientists’ failure to study these treatments slowly and systematically as they developed them.”
But perhaps because I wrote about this controversy earlier than most journalists, and have done so in major outlets, I’ve become a symbol of bigotry and hatred to a group of activists and online trolls as well as advocacy orgs like GLAAD that push misinformation about the purported safety and efficacy of these treatments, and attempt to punish journalists like Abigail Shrier for covering the controversy at all.
Bluesky appears to have attracted a particularly high number of these trolls, and even before I arrived on the platform, some of them were making sure I wouldn’t feel welcome there. Nora Reed, an online influencer and cultural critic, wrote in November that “I think we need a plan for if Jesse Singal shows up here in advance.”
“Honestly?” responded one user. “[G]un.” Another user then replied, lamenting the absence of technology that would allow them to shoot me via the internet:
Anyway, I joined nonetheless. I figured these messages were coming from a small group of disturbed people. There’s no lack of disturbed people on X—albeit ones who usually have very different politics—and I really did want an alternative to X.
When I arrived, I was was bombarded with messages from people telling me to kill myself, or expressing their opinion that I should be killed. When a Change.org petition signed by 25,000 people failed to get me booted off the platform—likely due to my having never come close to violating any rule—the anger only spread further.
Now, to a certain extent, this is the price of being a public figure on the internet, especially if you write about controversial subjects. If you take the occasional stray “kill yourself” or “i’ll kiill [sic] u” too seriously, you’ll go crazy, and you’ll give the trolls attention they don’t deserve. And generally speaking, mature social media sites do at least a decent job of booting the most pernicious users off the platform: You can simply report abusive accounts, block them, and move on.
Of course, it’s impossible to perform this task perfectly, to everyone’s satisfaction, once a site exceeds a certain size—especially given that it requires subjective judgments—but on Twitter, my experience has been that if someone threatens you with death, they’ll be banned.
Bluesky’s norms surrounding violent threats, meanwhile, seem to be far weaker than Twitter’s. A handful of those directed at me were lurid and specific enough to worry me a bit. For example, on December 10, a user named @billkezos.bsky.social (a bowdlerization of “Kill Bezos”), with about ten thousand followers, posted “Jesse Singal. 2 to the chest. 1 to the forehead a little less than [an] inch above the nasal bridge.” (One of his friends disagreed, arguing that I should be murdered by being beaten to death with a tire iron “methodically,” which does make for a memorable image, at least.)
I tagged Bluesky’s online safety handle at the time, writing “Either @safety.bsky.app takes this stuff seriously and bans these sorts of users, or the site will become unusable.” The site did not take action, as far as I could tell.
I have since wiped my account to start fresh, forgetting to take a screenshot of that post, but luckily @billkezos.bsky.social reposted it in mockery, so I still have access to it:
Then, on Saturday, I woke up to a handful of emails and DMs from individuals who had noticed that some of Bluesky’s users were circulating an address they believed was associated with me. The source appeared to be a post that Bill Kezos had made on Friday, which had gone viral by Bluesky standards, racking up hundreds of reposts. Users began calling for someone to go to what they thought was my address and hurt or kill me.
Now, attempted doxxing wasn’t entirely new to me. A couple years ago, in the course of my reporting on a deranged Northern Irish teenager who was something of an internet terrorist, said teenager became enraged and spammed me on every conceivable channel with what he believed to be my address. He had a long history of engaging in real-world acts of harassment, so I had to go to the Brooklyn address in question and explain to the very confused, middle-aged, Caribbean-born residents there what was going on.
But this time around, the situation seemed significantly worse, because it wasn’t one potentially sociopathic—but apparently nonviolent—teenager across an ocean who thought he had my address, but a sizable group of Bluesky users who had expressed very violent tendencies.
I should be clear that it’s rare for even severe instances of online threats and doxxing to escalate into real-world violence. Maybe I was just being paranoid! But this was the first time, in many years of being far too online, that I’d encountered this volume of threats, and this many people calling for real-life action against me.
So just after 8:00 a.m. Saturday, I emailed Bluesky’s moderators about the situation and got a quick response: “Thanks for reporting. We’ll look into it and take the necessary action.” I also alerted Jay Graber, the site’s CEO.
Then, for hours, I watched as the address spread more and more. I didn’t have any reason to fear for my own personal safety, but I was growing increasingly alarmed about the possibility of one of Bluesky’s many deeply dysregulated users throwing a brick through a window, or lingering creepily outside the address in question. On top of that, my brother, who looks like me, actually was staying in that neighborhood at the time. What if someone accosted him?
Bluesky’s team was absolutely nowhere to be found. I kept emailing Graber, updating her on the situation, and cc’ing Aaron Rodericks, the head of trust and safety. All I wanted was for someone to get back to me to explain whether they were actually doing anything to mitigate the situation. They never did.
At some point after posting my supposed address, Bill Kezos got suspended, at least, though a bunch of users then began circulating a screenshot of the post that included the address, demanding that the site “Free Bill Kezos.”
At 2:25 p.m., more than six hours after I’d first made contact with Bluesky, I emailed the moderation team as well as Graber, and asked, “Do you guys genuinely not have the technology to just temporarily ban the string [the address in question]?” Obviously this wasn’t an ideal solution, but it would temporarily stop the spread of the address. Bluesky still didn’t reply, but about an hour after that, the address didn’t come up when I searched for it. The hard-liners switched to posting the address in meme form, with some complaining that Bluesky was “protecting” me, but things generally calmed down.
Basically, it took something like seven hours for the site to implement this simple safety measure—long after the bulk of the damage was done. And as far as I can tell, their team hadn’t even thought to do so until I suggested it, though I’d be willing to accept that there was some technological obstacle that I don’t understand.
I didn’t want to write about this experience, and initially turned down The Free Press’s invitation to do so. I’d already engaged far too much with some broken and violent-seeming people on the internet, and had given them far too much consideration. And journalists who write about getting doxxed always get accused of “playing the victim,” or trying to build a profile, or whatever else. (I’m also giving some terrible people with no other platform the attention they crave.)
But this morning, I reached a breaking point. I was alerted that the Bill Kezos account was back.
Bluesky was aware of an account that has threatened to shoot me—that has specifically described where on my body he would shoot me—and posted what he thought was my address. Those actions weren’t enough to ban him, despite the fact that this wasn’t his first time being suspended, according to his own posts. The original attempted doxxing post was still up as well. It was still up as of this afternoon. Needless to say, so were a huge number of the other most violent posts about me.
In response to a detailed email of the claims that I make in this article, I finally heard back from Aaron Rodericks, the head of trust and safety. He said in a statement that “On Saturday, Bluesky’s moderation team. . . deployed a tool to remove these posts [with the address]. Threats of violence are in violation of our community guidelines, and our moderation team accordingly actioned those posts and accounts as well.”
Again: When he sent me this statement the post that started this whole thing was still up.
It was only after I followed up with Rodericks and pointed this out that it was taken down, but the account itself remains intact.
Bluesky is sending a clear signal that violent threats, attempted doxxing, and repeated suspensions are not sufficient to lose the ability to post there—this is the site they have chosen to present as an alternative to Twitter.
To be clear, I don’t expect Bluesky, or any social media site, to be able to remove each and every single threat during a moment of viral outrage. But what’s happened in the last week is like something out of a how-not-to-moderate textbook: Bluesky failed to ban this individual for an obvious, lurid death threat, and then that user did the most obvious thing imaginable: Realizing no one was going to step in, he pushed the line further. Had the users actively threatening me on Bluesky done so on Twitter, their accounts would have been banished days ago.
Bluesky happens to be left-wing, but I don’t think the lesson here is that left-wingers are particularly violent. Rather, the lesson is simply that humans are human, and online, their behavior is shaped by both the prevailing norms in their community, and whether rules constraining that behavior exist and are enforced. Barring a significant change in its policies, Bluesky shouldn’t be viewed as a remotely viable alternative to X. The latter has major problems of its own, but it has at least decided not to be an incubator for death threats.
