And how memetics will guide future tech products
“Who controls the memes controls the universe” tweeted Elon Musk, referencing a line from Dune. Somehow it feels both trollish and profound.
If you live on the internet, you know it when you see it. Memes are everywhere. More importantly, it feels like memes matter today. A well-crafted internet joke like Musk’s can yield more attention than a well-funded PR campaign.
Lately, I’ve been reading a fascinating book on the topic. First, I learned the academic definition of memes is far more expansive than just internet jokes. Richard Dawkins coined the term meme to refer to fashions, ceremonies, customs, and technologies that spread across human brains. The mechanism: mimesis, better known as imitation.
[Dawkins] discussed [meme] propagation by jumping from brain to brain, likened them to parasites infecting a host, treated them as physically realised living structures, and showed how mutually assisting memes will gang together in groups just as genes do. Most importantly, he treated the meme as a replicator in its own right. Everything you have learned by imitation from someone else is a meme. But we must be clear what is meant by the word ‘imitation’, because our whole understanding of memetics depends on it. Dawkins said that memes jump from ‘brain to brain via a process which, in the broad sense, can be called imitation’ (1976, p. 192)…
Everything that is passed from person to person in this way is a meme. This includes all the words in your vocabulary, the stories you know, the skills and habits you have picked up from others and the games you like to play. It includes the songs you sing and the rules you obey. So, for example, whenever you drive on the left (or the right!), eat curry with lager or pizza and coke, whistle the theme tune from ‘Neighbours’ or even shake hands, you are dealing in memes. Each of these memes has evolved in its own unique way with its own history, but each of them is using your behaviour to get itself copied.Susan Blackmore “The Meme Machine”
Think of memes as the smallest atomic unit of culture. Some memes are funny, some are relatable, and some are not very useful and don’t spread widely.
Memetic theory says these mind-viruses compete against each other for their slot in the next human brain. Some memes make it. Others don’t. (Even more interestingly, some memes make it without regard to their real-world usefulness—only their ability to replicate most effectively.) Memes, they argue, are part of a Darwinian system. This is all rooted in the idea of Universal Darwinism, which says evolution applies to any “replicator” with the following conditions:
- selection – the fittest survive
- variation – there are slight changes between copies
- heredity – the offspring inherits characteristics from the parent
Memes satisfy these conditions and replicate “cultural instructions” just like genes do.
Of course, our lives are increasingly digital. Arguably, more culture is mediated through media and tech platforms than in real life. Thinkers like Richard Dawkins and Daniel Dennett were talking about cultural evolution and the theory of mind abstractly more than, say, lip-syncing TikToks. Even so, types of social media posts—as digital shipping containers for memes—help them propagate.
To social media’s credit, they’ve taken the Darwin-governed world of memes and built fairly Darwinian systems around them. Now, a retweet button here or a like button there governs how memes spread (or struggle for life). Digital memes satisfy all the evolutionary pre-conditions:
- selection – some posts get comments & likes, going viral and becoming part of public awareness
- variation – every story gets told slightly differently
- heredity – Meme image templates, quote tweeting, TikTok duets all derive from existing content
You can expect a lineage to form from the most “successful” genres of social media posts. Viral templates survive for a reason.
Seeing social media through a memetic lens is, admittedly, a bit confusing. It’s hard to know where the shipping container begins and ends. In theory, they’re merely vehicles for culture to hitch a ride on. (As Blackmore wrote, memes are “substrate-neutral.”) But the mechanisms are worth studying. The difference between a Twitter meme (e.g. “Time for a thread”) and newspaper meme (e.g. “Area man…”) is the memetic evolution has been turned up a few notches. It spawns imitators way faster. Anyone can chime in with their own version.
If memes behave like genes, perhaps the various tech platforms resemble the different climate zones, where possessing a certain trait is highly adaptive. (Aside: Is Facebook akin to a tropical rainforest, enabling the most memeo-diversity? It might not be provable or even helpful to ponder.) The point is certain memes are naturally selected based on their product choices.
Finally, and most notably, memes travel across regions at a startling pace. Like an invasive species caught in a ballast tank, digital shipping containers spread memes farther and faster than ever. Every human brain is now a potential target.
The first great meme machine
In 2006, internet analyst Mary Meeker was asked how Google could justify its acquisition of YouTube for $1.65B. Meeker made the bull case and pointed out that YouTube felt like a natural evolution:
“What’s interesting to me about the evolution of media today is that these 3 minutes clips–and I’m generalizing–that are amateur in nature are often times very funny. Because people happen to be in the right place at the right time. When I first watched some funny videos on the internet, I was reminded of the early days of MTV when music was presented in a different way. And I was also reminded of the early days of Saturday Night Live…my hope and my bet is that [YouTube] was an event that will push the traditional content creators to focus more aggressively on monetizing their content on the internet.Mary Meeker (full clip)
The analysis still looks spot-on 14 years later.
What made YouTube special was its ability to aggregate short, punchy clips that are “amateur in nature.” The rest is history. With time, that ushered in a whole creator economy and encouraged media corporates to buy in. Now, banks value YouTube at $200B – $300B alone, making it one of the greatest M&A deals of all time.
YouTube’s scale and early start allowed it to enjoy a near-monopoly on virality, at least in video form. “Damn Daniel” and countless other memes from YouTube have ported over to everyday conversation. Previously, viral internet ideas spread as static images with editable text. Video added a new level of fidelity.
Now, YouTube’s scale allows the world’s ideas, jokes, and culture get replicated to the tune of 100 billion hours per day. Like any good social technology, YouTube unleashed an extremely powerful broadcast tool. In its wake, trillions of bits of cultural material spread from brain to brain. YouTube perhaps the first great meme machine on the internet.
In terms of meme replication power, YouTube enjoyed the last decade without a credible challenger. That is, until the rise of TikTok.
The second great (video) meme machine
TikTok has famously taken video virality to another level. There’s some great write-ups about the history of TikTok, and how it pivoted its app Musical.ly into what it is today. Personally, I think the most remarkable feature is TikTok’s explicit focus on memes.
TikTok’s main differentiator is audio-related. In every video, the audio itself can easily be “forked”—or copied—into a half-original new creation. All you have to do is film something compelling over it. This is not a huge secret. Every TikTok user knows there’s a song ID at the bottom of every video, and if pressed, will show you all the other videos using that song. From there, it will even encourage you to film your own imitation, lowering the friction to creation.
Audio forking spawns thousands of imitations
Currently, Fleetwood Mac’s song “Dreams” is back in the charts again thanks to Doggface208 posting a TikTok video lip-syncing on a skateboard while drinking Ocean Spray. Thousands of others have “forked” the Fleetwood Mac audio, and replicated DoggFace208’s video while skateboarding on their own. Other hit songs have minted entire video genres. [See video left]
The platform offers more than just dancing along to famous songs. The same applies to videos with original audio, which also can go equally viral as people iterate on the audio-meme. Often, this is done with the “duet” feature, where creators anticipate remixes on a script they create.
TikTok’s chief innovation is often said to be the creation tools, the mobile focus, the constrained time limit, or the superior algorithm. Those are all important. But I suggest the most important factor is this “A/V forking”–splitting audio and video–that encourages imitation.
By splitting content into the audio and visual, the hurdle for creating an entertaining video gets (1) easier for the creator and (2) exponentially more competitive in aggregate. Often, the original post of a viral TikTok is outdone by imitators that “perform” over the audio in an even more entertaining way. As a result, TikTok’s native content is far more explosive and gripping compared to YouTube.
The way TikTok focuses on memes is intentional and apparently the product of rigorous focus group testing. From The Information‘s paywalled piece:
TikTok executives were initially hesitant about the memes. But things changed in 2019, when Tiktok did an in-depth survey of its U.S. users, focusing on groups older than middle-school age. The study, called “tribe research,” screened and analyzed users’ profiles and usage of the app. It found that TikTok’s users felt disenfranchised on Snap and Instagram, according to a person familiar with the study. The people familiar with the study said they were surprised to find that the app was particularly popular among minorities, the LGBT community, and poorer whites. The common ground, according to the survey, appeared to be a fondness for memes.
The survey results prompted ByteDance to tailor the app toward memes. TikTok began directing user traffic to more meme-related content and started creating its own campaigns targeting memes. The company added the features that allow users to add texts directly to videos, enabling them to create memes more easily on TikTok. The popularity of memes in the U.S. surprised ByteDance employees. The company’s typical go-to market strategy overseas usually focused on recruiting top influencers from sites such as Youtube and Instagram, sponsoring big offline events and advertising campaigns.
This is obvious to anyone who’s used TikTok—it’s the main appeal—but it’s worth pointing out the innovation is audio- and meme-related. And it’s consciously built into the UI of the app.
Time for some meme theory
In TikTok’s case, the inherited meme is often the soundtrack—but includes other behaviors (e.g. skateboarding with Oceanspray while looking blissed out). TikTok memes that survive not only capture the zeitgeist, but represent all the things that are selected for as “good” on the platform.
Previously, most social platforms innovated on social status & metrics that show a “proof of work,” as Eugene Wei has noted. Broadly, these selection mechanisms translate to things humans care about in real-life (e.g. status). Likes, comments, and superior algorithms have also helped fine-tune the selection process of identifying what content is most “fit.”
TikTok seems to be playing a different game, evolutionarily. There’s less emphasis on likes and social capital (selection), and more on spawning new imitations (heredity). Technically speaking, by relying less on 100% original content, there’s increased heredity and decreased variation. The videos—and by extension the memes inside—are extra “fit.” This brings up the average TikTok video’s quality, which, in turn, makes the most “successful” videos exponentially better.
The lack of originality doesn’t come at a cost, either. As Jia Tolentino wrote in How TikTok Holds Our Attention, creators have embraced this new evolutionary process. There’s no shame in remixing:
“Adam Friedman has begun producing music directly for influencers, and engineering it for maximum TikTok success. “We start with the snippet, and if it does well on TikTok we’ll produce the full song,” he told me. I suggested that some people might think there was a kind of artistic integrity missing from this process. “The influencer is playing a central role in our culture, and it’s not new,” he said. “There’ve always been socialites, people of influence, the Paris World’s Fair. Whatever mecca that people go to for culture is where they go to for culture, and in this moment it’s TikTok.”
It’s hard not to compare this to hip-hop production, which was once criticized as stealing from classic songs and only adding a new vocal overlay. But it’s its own art form, and being able to reinterpret existing music has simply created a superior end product. (Not to mention, it’s grown into the number 1 genre in America.)
While status is still baked into TikTok’s system, that’s just table stakes now. Heredity—or remixing—seems to be a key new lever for social products. Going forward, perhaps the new playbook will center on designing systems for memetic evolution.
Where else could “forking” work?
It’s not always wise to pick the hottest tech narrative today and try to copy their secret sauce. In TikTok’s case, A/V forking has been baked into the platform for years. Its head start might actually be unstoppable, even with Facebook in the race. The content is the platform now, and the feedback loops are hard to replicate.
As far as I can tell, there aren’t tons of examples of social media remixing-as-a-feature, aside from the Twitter quote-tweet or the Tumblr retweet. The walls between posts are generally rigid. Originality and scarcity work.
YouTube is already trying to insert more TikTok-like short videos into its mobile app. But it doesn’t easily allow A/V forking.
Instagram might be the best candidate to make reposting easier (for one, it already allows reposting a tagged Story). Allowing this for static posts as well might be more impactful than Reels, its weak TikTok copycat. But it would also inflate the scarce currency of likes, risking their golden goose.
Another candidate might be Spotify, which could literally broker music remixing in-app. Spotify recently discontinued support Algoriddim’s DJay, which was a fantastic source for live DJing from the Spotify catalogue. Recently, Spotify started allowing podcasters to sample full songs with Anchor, enabling a type of remixing. Allowing professional-quality music to be created in an IP-compliant way could be really meaningful for the future of music, and may be a future goal.
Interactive content like Netflix’s Bandersnatch, could be forked and crowdsourced leading to better stories and experiences. Or perhaps more feasibly in the world of gaming, allowing user-created worlds be forked more easily (Fortnite does this to some degree with Creative Mode). Finally, I’ve seen some features on Reddit 3rd party app (Apollo) that makes forked image memes more visible. (You can click into the OP more easily.)
However, it might be up to startup challengers to truly innovate on content forking in the spirit of TikTok. Apps building in the airpods & audio space are currently hot, in part because they’re unique experiences in a world filled with images and videos.
Clubhouse and Roadtrip feel fresh because they allow synchronized remixing of live audio (aka discussions – what’s old is new). Riff works as a kind of shipping container for voice messages across different platforms. Shuffle allows snippets of podcasts get shared easily, and could perhaps add remixing where users re-interpret snippets with their own two cents. One interesting app called Voisey has built a TikTok-like video app for submitting hip-hop beats and allowing live-recording over them, creating a highly iterative hip-hop feed. (H/T to Vidy Thatte, who I know is building something cool there.) I’ve covered a few other audio startups in my newsletter here.
Beyond audio-first apps and social media, it’s less clear. Forking is a term borrowed from the world of open-source software and cryptocurrency and both remain great candidates. Other crazy areas could include journalism (forking stories based on agreed-upon facts) somehow on the blockchain. Things like legal-contracts could also be forked, but that’s getting beyond the original point.
In short, the next great meme machine could take forking to another level. Remixing will likely be part of the toolkit, as TikTok proved to have superior memes. Meme theorists like Dawkins might have more to offer than Skinner (and the legion of psychologists employed by Silicon Valley) when it comes to making compelling social products.
Futher reading/listening: Mary Meeker, Eugene Wei, Susan Blackmore, Jia Tolentino, Turner Novak, and Vidy Thatte have all inspired ideas here. Thanks to Brice Gigedot and Drew Austin for seeing early drafts.