The internet peaked with “the dress,” and then it unraveled

If you were on the internet on February 26, 2015, you saw The Dress. Prompted by a comment on Tumblr, BuzzFeed writer Cates Holderness posted a simple low-quality image of a striped dress, with the headline “What Colors Are This Dress?” The answers: blue and black or white and gold. The URL: “help-am-i-going-insane-its-definitely-blue.”
Do you really need me to tell you what happened next? In just a few days, the BuzzFeed post got 73 million page views, inspiring debate across the world. Seemingly every news outlet (including this one) weighed in on the phenomenon. How was it possible that this one image divided people so neatly into two camps? You either saw — with zero hint of variability — the dress as black and blue, or white and gold. There was no ambiguity. Only a baffling sense of indignation: How could anyone see it differently?
Looking back, the posting of “the dress” represented the high-water mark of “fun” on the mid-2010s internet. Back then, the whole media ecosystem was built around social sharing of viral stories. It seemed like a hopeful path for media. BuzzFeed and its competitors Vice and Vox Media (which owns this publication) were once worth billions of dollars.
The social-sharing ecosystem made for websites that would, for better or worse, simply ape each other’s most successful content, hoping to replicate a viral moment. It also fostered an internet monoculture. Which could be fun! Wherever you were on the internet, whatever news site you read, the dress would find YOU. It was a shared experience. As were so many other irreverent moments (indeed, the exact same day as the dress, you probably also saw news of two llamas escaping a retirement community in Arizona.)
Since 2015, the engines of that monoculture have sputtered. Today, BuzzFeed’s news division no longer exists; the company’s stock is trading at around 50 cents a share (it debuted at about $10). Vice has stopped publishing on its website and laid off hundreds of staffers. Vox Media is still standing (woo!), but its reported value is a fraction of what it used to be (sigh).
The dress brought us together. It was both a metaphor and a warning about how our shared sense of reality can so easily be torn apart.
Whether you saw gold and white or black and blue, the meme revealed a truth about human perception. Psychologists call it naive realism. It’s the feeling that our perception of the world reflects its physical truth. If we perceive a dress as looking blue, we assume the actual pigments inside the dress generating the color are blue. It’s hard to believe it could be any other color.
But it’s naive because this is not how our perceptual systems work. I’ve written about this a lot at Vox. The dress and other viral illusions like the similarly ambiguous “Yanny” vs. “Laurel” audio reveal the true nature of how our brains work. We’re guessing. As I reported in 2019:
Scientists still haven’t figured out precisely why some people see the dress in one shade and some see it in another. Their best guess so far is that different people’s brains are making different assumptions about the quality of the light falling on the dress. Is it in bright daylight? Or under an indoor light bulb? Your brain tries to compensate for the different types of lighting to make a guess about the dress’s true color.
Why would one brain assume daylight and another assume indoor bulbs? A weird clue has arisen in studies that try to correlate the color people assume the dress to be with other personal characteristics, like how much time they spend in daylight. One paper found a striking correlation: The time you naturally like to go to sleep and wake up — called a chronotype — could be correlated with dress perception. Night owls, or people who like to go to bed really late and wake up later in the morning, are more likely to see the dress as black and blue. Larks, a.k.a. early risers, are more likely to see it as white and gold.
In 2020, I talked to Pascal Wallisch, a neuroscientist at New York University who has researched this topic. He thinks the correlation is rooted in life experience:
Other scientists I talked to were less convinced this was the full answer (there are other potential personality traits and lifetime experiences that could factor in as well, they said). Even if there’s more to this story than chronotype, there’s an enduring lesson here. Our differing life experiences can set us up to make different assumptions about the world than others. Unfortunately, as a collective, we still don’t have a lot of self-awareness about this process.
“Your brain makes a lot of unconscious inferences, and it doesn’t tell you that it’s an inference,” Wallisch told me. “You see whatever you see. Your brain doesn’t tell you, ‘I took into account how much daylight I’ve seen in my life.’”
Moments like the dress are a useful check on our interpretations. We need intellectual humility to ask ourselves: Could my perceptions be wrong?
The dress was an omen because, in many ways, since 2015, the internet has become a worse and worse place to do this humble gut check (not that it was ever a great place for it). It’s become more siloed.
Its users are seemingly less generous to one another (not that they were ever super generous!). Shaming and mocking are dominant conversational forms (though, yes, irreverence and fun can still be had).
This all matters because our shared sense of reality has fractured in so many important ways. There were huge divides on how people perceived the pandemic, the vaccines that arose to help us through it, the results of the 2020 election. Not all of this is due to the internet, of course. A lot of factors influence motivated reasoning and motivated perceptions, the idea that we see what we want to see. There are leaders and influencers who stoke the flames of conspiracy and misinformation. But in a similar way to how our previous experiences can motivate us to see a dress in one shade or another, they can warp our perception of current events, too.
Though, I will admit: Maybe my perception of a more siloed internet is off! It’s hard to gauge. Algorithm-based feeds today are more bespoke than ever before. I can’t know for sure whether my version of the social internet is like anyone else’s. My TikTok feed features a lot of people retiling their bathrooms. That can’t possibly be the average user’s experience, right?
I have no idea if we’re all seeing the same things — and even less of an idea if we’re interpreting them the same way.
More chaos is coming, I fear. AI tools are making it easier and easier to manipulate images and videos. Every day, it gets easier to generate content that plays into people’s perceptual biases and confirms their prior beliefs — and easier to warp perceptions of the present and possibly even change memories of the past.
The dress represents, arguably, a simpler time on the internet, but also offers a mirror to some of our most frustrating psychological tendencies. What I wonder all the time is: What piece of content is out there, right now, generating different perceptual experiences in people, but we don’t even know we’re seeing it differently?