Why Online Trends Keep Looking Bigger Than They Are

Online trends often look much larger than they really are because social platforms reward visibility, not durable cultural adoption. In a new interview, NYU cyber-ethnographer Ruby Thelot argues that virality has become a misleading way to measure what is actually spreading across society.

Thelot, whose book In Defense of Being-On-Line is due this fall, says the internet is now so fragmented that many “viral” moments are really just localized bursts inside separate digital communities rather than true mass movements. That shift, he argues, helps explain why seemingly huge phenomena can fade quickly while quieter ideas endure.

Why virality is no longer a reliable test of what is real

Thelot’s core argument is that culture and attention have been distorted by the same logic that governs modern metrics everywhere: once a measure becomes the goal, it stops measuring what it was meant to measure. In his view, the chase for views, shares and mentions has turned virality itself into a product.

That matters because a fast-moving spike of attention can be manufactured without creating lasting social significance. A clip can travel widely, but that does not mean it will reshape behavior, language or identity in a meaningful way.

Thelot argues that the online world has become so segmented that many apparent trends are only truly “viral” within narrow communities, not across society at large.

He frames this as a Goodhart’s law problem: when platforms, creators and media outlets optimize for attention, the headline numbers become easier to inflate, while the deeper question — whether something has cultural staying power — gets pushed aside.

The result is an environment where internet users, marketers and journalists can all mistake a temporary burst for a broad shift. Thelot says that is why many high-profile online phenomena feel louder than they actually are.

How the internet became a cluster of digital islands

Thelot says the modern internet is best understood as a patchwork of “digital islands,” each with its own habits, feeds, jokes and reference points. What looks universal from the outside may actually be highly specific to a platform, a subculture or even a small network of users.

In that kind of environment, virality does not spread evenly from one shared public into another. Instead, content jumps between enclaves, where it may be remixed, misunderstood or amplified for reasons that have little to do with broad cultural appeal.

This helps explain why some online terms or subcultural markers suddenly appear everywhere, then disappear just as quickly. The surface signal can be intense even when the underlying adoption is thin.

Why “big trends” can be misleading

Thelot points to the risk of confusing attention cascades with genuine mass adoption. When an item is pushed repeatedly into feeds, it can seem ubiquitous even if most people are only seeing it because algorithms are optimizing their experience.

That distinction matters for readers trying to tell whether something is genuinely part of the culture or simply being force-fed by platform mechanics. According to Thelot, the answer is often less exciting than the hype suggests.

• Some phenomena are true cross-platform shifts.
• Many others are niche signals amplified by recommendation systems.
• Short-lived waves can still feel enormous because feeds make them visible again and again.
• Durability, not raw reach, is the better test of cultural impact.

What the dating-app debate gets wrong

Thelot also uses online dating discourse as an example of how culture writers and platforms can overread the mood of the internet. In recent years, the conversation around “dating burnout” has become a familiar story line, prompting app makers to rebrand and reposition themselves around AI and other features.

Bumble’s decision in May to drop its signature swipe feature is one recent example of how the industry has responded to that supposed fatigue. But Thelot says the narrative is more complicated than a simple collapse in romantic optimism.

He points to “heteropessimism,” a term used to describe the frustration some straight women feel about dating straight men, as evidence that the feeling is real for a subset of users but not necessarily a totalizing social condition.

To test that assumption, he and a friend examined roughly 1,000 Instagram and TikTok videos with more than 1 million views each, looking specifically for negative framing. The result, he says, was that about a quarter of the sampled content fit that category.

That does not mean disappointment in dating is imaginary. It does suggest that the loudest complaints may receive disproportionate attention relative to the broader conversation.

Thelot’s reading of the data suggests that the sense of universal romantic despair may be overstated by commentators who are interpreting the most visible posts as the most representative ones.

How much of the dating conversation is actually negative?

According to Thelot’s review of high-view videos on major platforms, the negative share hovered around 25 percent. He argues that is enough to keep the topic culturally sticky, but not enough to prove that the entire dating landscape is defined by burnout.

In other words, the conversation can be real without being total. Online audiences often mistake the scale of a debate for its universality.

What prediction markets reveal about participation culture

Another topic Thelot addresses is the popularity of prediction markets such as Kalshi and Polymarket, where users wager on everything from entertainment outcomes to environmental events. Their rise has been sold as a sign that younger users want skin in the game.

Thelot offers a more sociological reading. He compares today’s speculation culture with the meme-coin trading groups that flourished between 2020 and 2024, especially among young men organized in group chats and informal communities often referred to as “the trenches.”

His point is not simply that the behavior is gambling. It is that gambling, in these settings, is social. People do not only trade alone to maximize returns; they trade together, share information together and, crucially, lose together.

That shared risk creates a kind of belonging. The value of the wager is secondary to the feeling of proximity to whatever is being discussed, predicted or bet on.

Thelot says prediction markets are less interesting as financial tools than as a way for people to participate in the culture they are consuming.

Why people want to be close to the event

In Thelot’s view, the modern media user increasingly wants participation rather than passive observation. A wager, a bet or a speculative trade can function as a shortcut to closeness — a way to feel embedded in an event, celebrity moment or public storyline.

That logic helps explain why the financial upside is not always the main attraction. For many users, the appeal lies in being inside the action instead of merely watching from the sidelines.

• Prediction markets collapse entertainment, speculation and social identity into one behavior.
• Group chats and community chatter intensify the sense of shared involvement.
• Losses can matter as much socially as wins do financially.
• The attraction is often emotional and communal, not purely economic.

Why Thelot thinks AI hype is built on the wrong idea

Thelot is also skeptical of the idea that AI will transform the world because it becomes vastly more intelligent than humans. He does not believe intelligence is the main bottleneck holding society back.

Instead, he argues that many of the world’s hardest problems are about coordination: getting large groups of people to cooperate, align and act toward a common goal. In that sense, the issue is less about raw cognitive ability and more about social organization.

He also warns that once intelligence is treated as the core metric, it can begin to crowd out other human capacities. The more a society quantifies what it values, the more it risks elevating one dimension at the expense of others.

That is why he sees some superintelligence narratives as philosophically narrow and potentially dangerous. They project a specific elite ideal of intelligence onto machines and then treat that projection as destiny.

What gets lost when intelligence is everything?

According to Thelot, a fixation on measurable intelligence can leave emotional, artistic and bodily forms of intelligence undervalued. He draws on philosopher Gilbert Simondon to suggest that technological progress often leaves behind alternative ways of knowing and being.

That argument places AI inside a broader cultural pattern: when a society rewards one form of excellence, it can make other forms harder to see. The danger is not only machine overreach, but also human self-flattening.

In Thelot’s telling, the most important question is not whether AI will outthink people. It is whether people will let the technology narrow their conception of what intelligence even means.

How should readers tell a real trend from a manufactured one?

Thelot’s answer is to stop treating raw visibility as proof of significance. A real trend should do more than spike: it should leave traces in language, behavior, institutions or everyday life after the initial wave has passed.

That is a harder standard than counting views or mentions, but it is also more useful. The point is not to ignore what becomes popular online, but to ask whether the popularity is broad, durable and meaningful beyond the feed.

For journalists, marketers and ordinary users alike, that means a healthier skepticism toward the apparent consensus of the internet. What trends tell us most clearly, Thelot suggests, is not always what people care about — but what platforms can most efficiently surface, package and repeat.

What this means for culture, tech and the next hype cycle

Thelot’s analysis arrives at a moment when nearly every corner of the digital economy is trying to convert attention into identity, transactions or subscription revenue. Whether it is dating apps, betting platforms, AI tools or social networks, the business model increasingly depends on keeping users engaged enough to participate.

That makes the difference between attention and adoption more important than ever. A platform can dominate the conversation without winning real commitment from users, just as a product can appear culturally unavoidable while remaining shallowly adopted.

His warning is that internet users and commentators should be careful not to confuse the feeling of saturation with genuine transformation. In a fragmented media environment, the loudest trend may be the least representative one.

That idea is likely to shape how Thelot’s book is received when it arrives this fall. It also offers a useful correction for an era that often treats every spike as proof of a new era.

Key takeaways from Thelot’s critique of trends

• Virality is no longer a dependable measure of cultural importance.
• The internet is fragmented into separate communities that make trends seem larger than they are.
• Dating pessimism is real but likely less universal than viral discourse suggests.
• Prediction markets reflect a desire for participation and belonging, not just profit.
• AI hype may be overfocused on intelligence and underfocused on human coordination.

Timeline of the recent trend conversation

Ultimately, Thelot’s case is not that online culture is fake. It is that the internet has become a misleading stage on which scale, speed and repetition can overpower substance. If that is true, then the next time a term, app or subculture appears to take over the world, the first question should not be how viral it is — but whether it will still matter once the feed moves on.