There was a brief, halcyon period in the 2010s when lots of people in white collar jobs were first encountering both high-speed internet and the smartphone boom and at the time, it genuinely felt possible to get a feeling for what was hot. Thanks to shared platforms and access to actual data about what people were watching and sharing, we were able to see what was popular and to discern the shape of the culture.
Now? Now it’s so hard to know what people genuinely like that Amol Rajan can convince himself that he can sustain an eponymous media brand. What happened?
The answer, in large part, is the slow death of social media. You might think that data-led platforms, designed to maximise engagement by working out what people like and most want to consume and then firing it at their eyeballs at potentially-ruinous volume, would give us a window into what exactly that stuff is. Instead, the past 15 years have seen a slow erosion of the data which used to give us an idea of what people were consuming online.
Time was, you could just look. Twitter was once a researcher’s playground, where anyone with a bit of coding nous could see which topics were trending and which links were being shared, and by how many. Since Elon Musk took over, that data has been locked behind a paywall priced at tens of thousands of dollars a month.
Meta switched off CrowdTangle, the tool that let journalists see in near-real-time what was spreading across Facebook and Instagram, in August 2024, replacing it with a “Content Library” that’s open only to approved academics and nonprofits; which is to say, not to the press.
The public-facing signals have quietly been dimmed too: YouTube hid dislike counts in 2021; and both Instagram and Facebook let users hide likes the same year. Meta noted with horror at the time that people used them “to get a sense for what’s trending or popular”, which was entirely the point.
X made likes private in 2024 and if you’ve tried recently to work out when a TikTok or Reel was actually posted, you’ll know it’s become oddly difficult. Even where the numbers are still visible, there’s no telling whether they mean anything.
Suggested Reading
Forget the Terminator – the age of AI weaponry is already here
Faking views on the internet isn’t a new thing. When I worked in digital marketing in around 2011, the going rate for buying YouTube views was £1 per 100. There were no questions asked as to where those “views” were coming from . But now, spoofing numbers has become big business.
Earlier this year, the band Geese were at the centre of controversy when it emerged that their management had, allegedly, been paying an online marketing firm to promote them. That firm used trend simulation, viral clip distribution, and social media/TikTok networks to boost algorithmic engagement for the band, making them look significantly more popular than they in fact were.
This extends to fan accounts too. Hyperonline acts such as Sombr have fan accounts dedicated to posting candid clips of concerts, and other “behind the scenes” stuff, demonstrating the organic popularity and obsessive fandom of the artist… except, it turns out, fans suspect these are run by his own team.
This extends to pretty much everything online now. Cast your mind back to early 2026, when the American “looksmaxxer” Clavicular was suddenly everywhere on people’s social feeds, seemingly out of nowhere. Clavicular is a man whose life is livestreamed on niche video platforms such as Kick.
His content was edited and pushed relentlessly across more mainstream sites such as X and Instagram by dedicated “clipping” services, charging tens of thousands of dollars per month for the service. This was enough to convince mainstream media of his popularity, leading to profiles in the New York Times and Guardian and catwalk appearances at Fashion Week.
Were significant numbers of real people actually watching the vapid antics of an online meatpuppet now charged with rape? No one knows! Equally, no one knows who was paying for the clips to be made and distributed.
The latest version of this has been the boom of “virtual influencers” on Instagram and TikTok. These include Jean Philanthrope, an AI-generated, freakish-looking avatar, just human enough in appearance to trigger an “is that a real person?” response in the viewer, racking up (it appears) tens of millions of views.
Except – are these things popular? While claims abound that AI influencers are getting millions of clicks and making thousands of dollars per video, try verifying any of these numbers and you quickly come a cropper. Meanwhile, a selection of grifters are trying to make money out of selling “how to” guides to credulous chancers who think that it’s really possible to make money this way. Perhaps that’s where the real money is.
While much has been made of the ruinous impact of AI on everything, it’s worth considering whether this won’t be one of the most significant consequences. Social media, and the wider web, has been flooded with fakes: fake views, fake comments, fake reviews, fake shares, fake clips, and now even fake influencers. Where has all the authenticity gone?
Suggested Reading
Twitter, Bluesky and the death of social media
Except, of course, “authentic” popularity has always been manufactured. Music charts were manipulated by payola (labels paying radio stations to play specific records) throughout the 20th Century. Authors and publishers have been caught bulk-buying hardback runs to juice the bestseller charts. Breakout stars of stage and screen were often beneficiaries of multi-million dollar marketing campaigns carefully constructed to hide their existence from an unwitting public longing for a good old “talent will out” story.
Maybe this new era isn’t quite so bad after all. The AI influencers seem to have no long-term traction, the views dropping off a cliff once the novelty of a character wears off. Geese continue to sell out shows in real life, regardless of the “reality” of their online hype. Perhaps all this is just a useful reminder that taste is something we have, not something we follow, and that basing what we watch, consume, listen to or follow on data leads to boring, herdlike behaviour and mediocre art.
Think of it like this – if we can’t tell what’s popular anymore, we are freed from the tyranny of “big taste” and can instead learn to ignore the numbers and just watch and listen to what we like rather than what we are told everyone else thinks is good. Which, fine, doesn’t bode well for Amol Rajan, but may be good for culture in the long run.
Most popular
I was at Eton with Charles Spencer – this is how it explains his battle with the royals
Only a lunatic would escalate the war in Iran – and that lunatic is Donald Trump
Forget the Terminator – the age of AI weaponry is already here
The Budget that smelled like an early election
