I made a burner account on Twitter recently, mostly out of morbid curiosity as I do every few months to poke at the festering, bloated corpse, and came away with a fairly settled opinion: the platform doesn't read like a social network anymore so much as a right-wing satire show that's forgotten it's supposed to be funny. The "comedy" on offer is just straightforward hatred and hypocrisy, performed at volume, with the studio audience cheering along. I'd call it dark comedy if anyone involved seemed to be in on the joke.
The Bit About Bluesky
Somewhere in the scroll I found a user mocking Bluesky for the amount of pornography on there, in the tone of someone who'd just discovered a smoking gun. Within about thirty seconds of that post, I was served a blatantly sexualised ad for Grok. Not implied. Not suggestive-if-you-squint. Straightforwardly pornographic, sitting in my feed, on the platform whose user had just finished laughing at the other site for exactly that.
It isn't just showing up in the feed either, as it turns out.
That's not a stray ad somebody slipped past moderation. That's one of Grok's actual App Store listing images — the promotional screenshots Apple's review team signed off on, sitting on the page you land on before you've even downloaded the thing. "Companions" is Grok's name for the feature — anime avatars with names like Ani, built specifically to flirt, and in Ani's case to escalate into explicit territory once a user's "affection level" climbs high enough. 9to5Mac covered the original rollout as a potential App Store rules violation, given Apple explicitly bans content depicting sexual organs or activity "intended to stimulate erotic rather than aesthetic or emotional feelings." Apple apparently agreed enough to act on it eventually: MacRumors reported that Apple privately threatened to pull Grok from the App Store entirely over sexualised imagery, rejecting xAI's first attempted fix as insufficient before eventually, reluctantly, letting a later version through. Whatever that review process actually screens for, it apparently isn't this — the listing itself is the evidence, not a leak from it.
I'd love to say this was an isolated glitch in ad targeting. It isn't. Reporting from the Center for Countering Digital Hate found that over an eleven-day window following Grok's "Edit Image" rollout in late December 2025, the tool generated millions of sexualised images, with researchers estimating tens of thousands depicting minors — a figure serious enough that NPR reported Indonesia and Malaysia temporarily blocked the chatbot over it, and the UK's Ofcom opened a formal investigation. X's official response, paraphrased generously, was that this was a user problem rather than a product problem.
It gets worse, not better, the further you follow it. By June 2026, Forbes reported that well over half of all traffic through xAI's Grok product was now driven by pornographic image generation, explicit video, and adult roleplay — not as an accidental byproduct, but, according to former employees cited in the same reporting, as a deliberate strategic pivot once rival AI companies started banning explicit generation outright. So the joke isn't just that the ad showed up. It's that the ad showing up is, by the company's own apparent internal logic, the point.
Meanwhile Bluesky's actual policy on this, for anyone who'd rather check than assume, is that adult content is opt-in and hidden by default, with explicit material requiring users to deliberately enable it and self-label accordingly, and CSAM removed without human review the moment it's flagged. I'm not pretending Bluesky is some pristine utopia free of bad actors — no open network is — but the structural difference between "you have to go looking for it and the rules say you must label it" and "it's serving as an unsolicited ad while half your traffic is explicit content by design" is not a small one. It's the entire argument, really.
The Same Five Minutes
The Bluesky-porn bit wasn't even the only recurring joke doing the rounds. In the same five-minute window I saw at least two posts insisting Bluesky has no actual users — which is news to my roughly 2,400 followers, and apparently also news to the platform's own registered user count, which sat at over 40 million by late 2025, with daily active users in the low millions on top of that. I'm not claiming Bluesky rivals X on raw scale — it doesn't, and nobody serious says otherwise — but "no actual users" and "tens of millions of registered accounts with a multi-million daily active base" are not the same claim, and one of them is checkable in about the time it took to type the joke. And then, in the same stretch of scroll, a post asking what slur people should start using for Bluesky users. Not a joke about the platform. A genuine request for one, replies and all. I'm not going to repeat what got suggested. I'll just note that this happened in the same five minutes as the "no users" jokes and the porn-shaming, on a platform that's allegedly there to laugh at everyone else's hypocrisy rather than supply its own.
The Wider Bit
None of this happens in a vacuum, which is the part that makes the "comedy" framing land for me rather than just feeling like an isolated bad ad. Academic research from Berkeley, published in PLOS One, tracked hate speech on X across the months following Musk's 2022 acquisition and found weekly rates running roughly 50% higher than before the takeover, with transphobic slurs specifically up around 260%. That's not a vibe I picked up from doomscrolling. That's a peer-reviewed measurement of the room I was standing in.
And then there's "Patriot Rising," a handle I scrolled past with a small collection of Christian emoji bolted to the end of it, posting the kind of thing that made me genuinely sad rather than annoyed — which is, I think, the more honest reaction. The hatred I expected. The costume it was wearing wasn't. There's a particular flavour of cognitive dissonance in watching someone perform piety with one hand and contempt with the other, and not appear to notice the join.
I'll say this plainly, because I think it's worth being upfront about where I'm standing when I write any of this: I'm Pagan, and this is exactly the kind of Christian that's put me off being preached to by Christians at all. Not Christianity as a faith — I've no particular stick up my arse about people who actually practise theirs quietly and decently. It's this specific strain, the emoji-and-outrage kind, treating the religion as a costume for cruelty rather than a reason to be less cruel, that's done more to sour me on being lectured at than any actual theological disagreement ever has.
Where That Leaves Me
I don't think the platform is secretly self-aware about any of this. I think it's accidentally become a parody of itself, and the people inside it are too committed to the bit to clock that the audience is laughing at them rather than with them. The hypocrisy isn't a side effect. It's load-bearing — the outrage about pornography elsewhere only works as a punchline if nobody checks what's running in the background on the platform telling the joke.
I deleted the burner account. The research, unfortunately, I'm keeping.