American Idle

I promised one final piece on TikTok, focused primarily on the network effects of creativity. And this is that, in part. But it discusses a bunch of other topics, some only tangentially related to TikTok.

All the points I wanted to cover seem hyperlinked in a sprawling loose tangle. This could easily have been several standalone posts. I've been stuck on how to structure it.

Some people find my posts too long. I’m sympathetic to the modern plague of shortened attention spans, but I also don’t want lazy readers. At the same time, this piece felt like it was missing a through line that would help pull a reader through.

And then I had a minor epiphany, or perhaps it was a moment of delusion. Either way, it provided an organizing conceit: I decided to write this piece in the style of the TikTok FYP feed. That is, a series of short bits, laid out vertically in a long scrolling feed.

This piece is long, but if you get bored in any one section, you can just scroll on the next one; they're separated by horizontal rules for easy visual scanning. You can also read them out of order. There are lots of cross-references, though, so if you skip some of the segments, others may not make complete sense. However, it’s ultimately not a big deal.

If I had more time, I might have built this essay as a series of full-screen cards that you could swipe from one to the next. Or perhaps tap from one to the next, like Robin Sloan’s tap essay (I wish there a way to export this piece into a form like that, if someone built that already let me know). And if I were even more ambitious, I would've used some Anki-like spaced repetition algorithm to randomize the order in which the following text chunks are presented to you, shuffling it each time a reader jumped in.The most meta way for me to ship this essay would have been as a series of TikTok videos. It would have been the Snowfall of TikTok essays. That would have also taken a year of my life (which, being locked inside because of a pandemic might be the time to attempt something like that?). Also, I am camera shy.

But as it is, this is what you get.


By network effects of creativity, I mean that every additional user on TikTok makes every other user more creative.

This exists in a weak form on every social network and on the internet at large. The connected age means we are exposed to so much from so many more people than at any point in human history. That can't help but compound creativity.

Various memes and trends pass around on networks like Instagram and Twitter. But there, you still have to create your own version of a meme from scratch, even if, on Twitter, it's as simple as copying and pasting.

But TikTok has a strong form of this type of network effect. They explicitly lower the barrier to the literal remixing of everyone else's content. In their app, they have a wealth of features that make it dead simple to grab any element from another TikTok and incorporate it into a new TikTok.


The barrier to entry in editing video is really high as anyone who has used a non-linear editor like Premiere or compositing software like After Effects can attest. TikTok abstracted a lot of formerly complex video editing processes into effects and filters that even an amateur can use.

Instagram launched one-click photo filters (after Hipstamatic, of course, though Hipstamatic lacked the feed which is like the spine of modern social apps), and later Instagram added additional features for editing Stories, and even some separate apps like Boomerang that were later re-incorporated back into Instagram as features.

Snapchat has a gazillion video filters, too, though many are what I think of as simple facial cosmetic FX.

YouTube has launched almost no creator tools of note ever. WTF.

TikTok launches seemingly a new video effect or filter every week. I regularly log in and see creators using some filter I've never heard of, and some of them are just flat out bonkers. What creators can accomplish with some of these filters I can't even fathom how I'd replicate in something like the Adobe Creative Suite.

Kili So Silly (@kili.so.silly) has created a short video on TikTok with music original sound. | #stitch with @xxelacxx #TimeWarpScan #fyp #foryou

JeremyLynch (@jeremylynch) has created a short video on TikTok with music Despicable Me (From "Despicable Me"). | This freaks me out watching it back 😅 #timewarp #timewarpchallenge

TikTok’s Warp Scan filter is a bizarre concept for a filter in and of itself, but the myriad of ways TikTok users put it to use just shows what happens when you throw random tools to the masses and allow for emergent creativity. It only takes a handful of innovators to unleash a meme tsunami.


A longstanding economics debate is why we haven't seen the effects of the internet in our productivity figures. I won't rehash every side of every argument there.

But I know this: to take someone else's video and insert a reaction video of my own playing alongside it on the same screen is not easy in a traditional NLE. I'm not saying it's the moon landing, but it's not trivial.

On TikTok, you can just press the Duet button and start talking into your phone, and soon you have a side-by-side of the original video and your reaction video (you can choose from any number of their preset layouts for reactions). That's an explicit productivity boost; I can measure it in time saved for the same output.

You won't see that show up in GDP per capita figures, but it's real.


Remember 2 Girls, 1 Cup? If you've seen it, how could you not?

What interested me was less the video, which just horrified me, but the reaction videos of people watching it. Because 2 Girls, 1 Cup was a short video, I think it was a minute or two long, you could simply watch the face of someone watching the video and sync every reaction to every horrific beat of the video now forever haunting your memory, even though the original video wasn't visible on screen. The fun of the 2 Girls 1 Cup reaction video, but reaction videos in general, is that shared context.

Until TikTok came along, there wasn't an easy way to do reaction videos to other videos and have them make sense unless the original video had so much distribution that it was common knowledge. Or you could put the reaction video alongside or on top of or beneath the original video, but that required skill in using a non-linear video editor to lay those out and synchronize their timelines.

With TikTok's Duet feature, you can instantly record a side-by-side reaction video to anyone else's video. Duet is the quote tweet of TikTok. Or you don't have to do a reaction video at all. The Duet feature is designed simply to allow you to record a video that will play back alongside another video. It can be used for reaction videos, sure, but also to just provide a running commentary on other videos, and there are entire accounts built around both concepts.

But again, the Duet feature is built at such a low level that you can treat the feature as a primitive to replicate any number of other editing tools.


One such tool is to use the Duet feature as a dynamic matte. Since you know where your video will be placed in relation to the original poster's video, you can build a video mosaic.


Another is to use the Duet feature to, well, literally record duets.

But if you allow Duets to stack, well, eventually, one Wellerman can bring the whole chorus to your yard.

Someone truly ambitious could adjust the playback speed of various levels of Inception from the film Inception and stack them and synchronize them in TikTok using the Duet feature. If I had more time I'd do this myself, but the time has come for some time-rich kid out there to take this on.


Knowing that others can Duet your video means you can post any number of videos as prompts.

For example, you can read one side of the dialogue in a two-hander.

sara (@saraecheagaray) has created a short video on TikTok with music 人生のメリーゴーランド (Jazzical Lounge ver.) [『ハウルの動く城』より]. | #duet with @thechrisbarnett NOW I can say my accent sucks #fyp #foryou #acting


Knowing that TikTok has a Stitch feature, you can also post a question in a video and expect that some number of people will use Stitch an answer to your question and distribute that as a new video.

A popular prompt is "Tell me you're X without telling me you're X" or any number of its variants like "Show me you're X without showing me you're X."

Stitch wasn't necessarily designed to be used in this way, but as a primitive it's well-suited to any number of uses, including making TikTok a sort of video Quora.


Video prompts can come from not only other TikTok videos but commenters.

Some TikTok videos are made in response to requests posted in the comments. The comment is excerpted and published as an on screen text overlay at the beginning of the response video.

This is another of the nested feedback loops within the global feedback loop that is the FYP talent show. Once one example of this went viral, then the entire community adopted this as one of the norms of the community.


The Daily Show with Jon Stewart was a show almost entirely built around the reaction video. Stewart would play some clip of a politician being a hypocrite, or some Fox News anchor spouting their usual performative indignation, and then the camera would cut back to Stewart, his face frozen in some emoji mask of shock: eyes wide, mouth agape.

Social networks, and entertainment networks like TikTok, have completed the work of democratizing reactions. Yes, there's no reason you need to react to everything. But it's human nature. This is the social contract of the social media era. If you dare to shout your opinion or publish your work to the masses, the masses can choose to shout back.

Gossip litigates and fleshes out the boundaries of acceptable behavior within groups. Whereas gossip used to be contained, social networks now give it global distribution. This is one reason of many we've seen in-group and out-group boundaries drawn in bolder weight in this era. For every wide-eyed look of horror by Jon Stewart, you had the furrowed brow of disbelief that is Tucker Carlson's signature look, like someone in his elevator car passed gas.

Now extend that to clapbacks on the internet and you have a world in which back-channel gossip, a useful release valve and distribution channel for information about our peers, has become an open dialogue. The grapevine became the public feed, and every day, kangaroo court is in session.


TikTok's Duet feature belongs in the social media hall of fame of primitives alongside features like Follow and the Like button.

What feature better epitomizes the remix, react culture of the internet? Paul Ford once wrote that "Why Wasn't I Consulted?" is the fundamental question of the web. By then, social networks were well on their way to taking over from the web, and in the process, installing the plumbing by which the masses could finally directly opine to the masses, who could, in return, directly consult back. The "reply guy" is the consultant class of the internet, and mansplaining is its verb.

Yes, there are quote tweets and replies, but the TikTok Duet is the video analog, so simple and elegant in its design that you wonder why YouTube didn't launch it ten years ago, and then you remember that YouTube hasn't launched any creator tools of note since...ever.


What the Duet feature does, as described by how it would be done in a traditional non-linear editing program like Adobe Premiere, is the following:

  1. Copies the original file
  2. Inserts a new video track and a new audio track on top of the originals
  3. Allows you to lay down a new video on those new tracks
  4. Performs a whole series of steps to arrange the videos side-by-side on screen

TikTok abstracts a bunch of steps into a single function.


Yes, yes, some of these features in TikTok came from Musical.ly. But that's just a meta form of the theme of this piece! TikTok sampled from Musical.ly and improved upon it. They remixed a remix app.

But also, isn't this how innovation happens? We stand on the shoulder of giants and all that? Good artists copy, great artists steal?

TikTok enables, for video and audio, the type of combinatorial evolution that Brian Arthur describes as the underlying mechanism of the tech industry's innovation.


How many truly original ideas are there in Silicon Valley? Very few. Most have been tried umpteenth times in the past. Much of finding product-market fit in tech is context and timing. And people always underestimate the market side of product-market fit. When something fails, people tend to blame the product, but we should blame the market more often. The pull of the market is usually as important, if not more so, than the push from a product.

One day, the conditions are finally right, and an idea that has failed ten times before suddenly breaks out. Sometimes it's a tweak in execution, maybe it's an advance in complementary or enabling technology, sometimes it's a cultural shift.

Most of the best ideas in tech first appeared in science fiction books in the 1960s, and many of those are still waiting for their time to come. This is why rejecting companies that are trying something that's been tried before is so dangerous. It's lazy pattern-matching.

I do like Jeff Bezos' principle on when he decides to finally give up on an idea: "When the last smart person in the room gives up on the idea." But it also implies that you should bring some ideas back when a new smart person, or maybe a naive overconfident one, enters the room and champions the idea.


Given we know innovation compounds as more ideas from more people collide, it's stunning how many tech firms, even ones that ostensibly tout the value of openness, have launched services that do a better job of letting their users exchange ideas than any internal tool does for their own employees’ ideas.

How many employees join a firm and then spend a week in orientation learning where to get lunch, how to file expense reports, mundane trivialities like that. How many sessions are led by random trainers who don't even work at the company?

If you think of a company as an organism, and new employees as new brain cells, it's staggering how many join the company and begin from an absolute cold start. It's as if the company has chronic amnesia. What has the company learned from its past, what is its culture? When employees take months or even years to get up to speed at a company, companies should be embarrassed. Instead, it's treated as normal.

The free flow of ideas outside a company shouldn't, or in apps like TikTok, shouldn't exceed the rate at which knowledge flows inside a company, but I see it happen time and again.


The toughest job for any creative is the cold start. The blinking cursor on the blank page in a new document. Granted, writing a tweet, or even shooting an Instagram photo, isn't like composing the great American novel. But we tend to underrate the extent to which new users often churn without having ever posted anything to a social network because we only focus on those who do.

Now imagine trying to make a TikTok from scratch if you're older than, say, 19. The creative bar is high, you don't know how to dance, you're not up on the latest memes or popular music. Even if you're a teen, it's not easy to come up with a 0 to 1 TikTok.

But the beauty of TikTok's FYP algorithm and the Discover page is that you don't have to create a TikTok from scratch. The vast majority of TikToks are riffs on memes and trends that other users originate. It's no shame to be a 1 to n TikToker. Many on the platform achieve their first viral hit riffing on an existing meme.

Charli didn't invent the Renegade dance, Jalaiah Harmon did, but Charli made it famous. A lot of Charli and Addison's most popular TikToks are their interpretation of dances other people choreographed to songs other people composed.The ongoing debate on cultural appropriation seems to have no end in sight, but at least on TikTok there is a chance, with time stamps and some of the literal links the app creates between videos, to trace the origin of memes more easily.


Richard Dawkins introduced the term meme in his classic The Selfish Gene, defining it as a unit of information that spreads via imitation. He noted that memes evolve via natural selection just as in evolution. This memetic evolution happens via the same mechanisms as biological evolution, via variation, mutation, competition, and inheritance.

The internet writ large has always been fertile ground for the accelerated breeding of memes (cue toothless old prospector: "Back in my day sonny boy we had to spread memes via email chain letters"). But the TikTok app is perhaps the most evolved meme ecosystem to date.


Assisted evolution occurs when humans intervene to accelerate the pace of natural evolution.

TikTok is a form of assisted evolution in which humans and machine learning algorithms accelerate memetic evolution. The FYP algorithm is TikTok's version of selection pressure, but it's aided by the feedback of test audiences for new TikToks.

Memes can start from almost anything on TikTok. It can be the lyrics of a song, or just the vibe of a track, or both. A user can post a question or a challenge. In a single session on TikTok, you'll find videos of all types, most being riffs on existing memes (the variation).

Regardless of the provenance, any video, once loaded into TikTok, is subject to the assisted evolutionary forces in the app. Software tools like the Duet or Stitch feature and all of TikTok's other video editing tools assist in mutation and inheritance, and each remix of a source video becomes a source video for others to remix, generating further variation. Meanwhile, the competition on the FYP feed is fierce, and the survivors of that extreme selection pressure are memes of uncommon fitness.


In this assisted evolutionary ecosystem that is TikTok, and with an...umm...assist from the pandemic that kept hundreds of millions of people locked inside scrolling their phones, we've seen a marked contraction in the half-life of memes.

Memes used to dominate TikTok for what felt like weeks, and now it seems the memetic zeitgeist on TikTok shifts every few days, if not nightly. If I don't check back on TikTok every day, I find myself scrambling to catch up to the meta when I finally do open the app.


Of course, people grab TikToks and share them on YouTube or Twitter or as Reels on Instagram, but those apps receive flattened video files and can’t break them into component parts to be remixed the way you can on TikTok. Those other services are fine endpoints for distribution, but the creativity happens on TikTok. Don't get me started on apps like Triller (which feels like a Ponzi scheme).

People will litigate Instagram copying Snapchat's Stories feature until the end of time, but the fact is that format wasn't ever going to be some defensible moat. Ephemerality is a clever new dimension on which to vary social media, but it's easily copiable.

This is why TikTok's network effects of creativity matter. To clone TikTok, you can't just copy any single feature. It's all of that, and not just the features, but how users deploy them and how the resultant videos interact with each other on the FYP feed. It's replicating all the feedback loops that are built into TikTok's ecosystem, all of which are interconnected. Maybe you can copy some of the atoms, but the magic lives at the molecular level.

TikTok has a a series of flywheels that interconnect, and there isn't any single feature you can copy to recreate the ecosystem. Meanwhile, Reels has to try to compete while being one of like a half dozen things jammed into the Instagram app.


Markets in the internet and technology age are conducive to winner-take-all effects thanks to preferential attachment. This means that if you are first to stumble upon some flywheelMany like my friend Kevin use the term loops. I use flywheel merely to indicate I'm referring to positive feedback loops since loops can also be negative feedback. Also, I had to make that damn Amazon flywheel diagram for way too many presentations back in the day, it's mounted on a wall in my brain. in your business, the returns are even greater and accumulate more quickly than they would've in any other era in history.

Building a flywheel, though, often requires connecting a series of features at once. When I advise various companies, big and small, I often run into objections to my recommendations because of the popularity of agile or other incremental development philosophies. We end up at loggerheads on the V of MVP (minimum viable product), V having always been contextually determined.

If a flywheel requires three or four or even more things to connect in your app, it takes more work to ship all of them at once, and that feels like a riskier expenditure of your team's time. But, I'd counter: 1) often, testing a flywheel by definition means you have to build multiple features that work together 2) the returns of achieving a flywheel are often so high as to be worth the risk and 3) if you don't achieve any flywheels you are, as investor updates are so fond of saying, default dead.


Instagram famously has never had its version of resharing (e.g. retweeting). This reduced the velocity of photos and later videos on the service, a sort of brake on spam and misinformation and other possible such downsides.

But after using TikTok, it does feel odd to go through Instagram and not be able to grab anyone's photo to remix. Imagine you could grab someone's photo and apply your own filters, or grab just one element of the photo and use it in your photo.

Once we all live in the metaverse, this type of infinite replication and remixability will be something we take for granted, but even now, we're starting to see an early version of it on TikTok. This type of native remixability feels like it will be table stakes in future creative networks.


Fanfic is one text version of sampling and remixing. It doesn't require much more than your imagination.

It's always been really expensive, in both time and legal costs, to sample and remix film and television. TikTok has, with its short video format and tools, made remixing of premium video easier and safer. In Harry Potter TikTok, and its sub-genus Draco Malfoy TikTok, creators pull from the repository of the Harry Potter film universe as if it were on GitHub and merge themselves into branching storylines in which, well, creators become students at Hogwarts and catch the romantic interests of one Draco Malfoy.


The Discover page acts as the Fed in the central economy of memes on TikTok, while the FYP algorithm is the interest rate on meme distribution.

The Discover Page features hashtags. By the very act of featuring a hashtag, they signal to creators that if they create using that hashtag, they will get the distribution boost of that hashtag being featured on the Discover Page. Which raises the age-old conundrum, which came first, the Discover Page hashtag placement, or the hashtag's trending? The answer is yes. It's circular, an ouroboros of virality.

TikTok also posts the number of collective views on videos with that hashtag, helping creators gauge the potential distribution value of climbing aboard that trend.

TikTok is a mix of a centrally planned economy and a free market, much like many multiplayer video games where the game publisher manages the price and availability of various assets like weapons and armor while the players put them to use in the virtual economy.

The Discover Page is also where TikTok will feature corporate challenges. Yes, it's a paid placement, but the creative output is collective and distributed.


Because the most popular memes get super-distribution via the FYP algorithm, you can assume common knowledge of the meme among your viewers and just cut to the punchline. You don't need a bunch of what would be the video equivalent of exposition upfront. This keeps the majority of videos on TikTok compact, critical to the high cadence of the FYP feed. TikTok feels fast. Almost manic.

It also gives viewers that hit of in-group dopamine when they already know the references in your video.


If you don't understand a TikTok video and its references, you can trace the provenance from within the app in any number of ways. You can follow the hashtags in the caption or tap the sound icon and see all the other videos which have been made in that meme branch. Often that's enough to derive the context.

Or you can just read the comments. You'll find you usually aren't alone, someone will almost always have posted a comment like "In here before the smart people arrive" and then below that will be comments that explain the video to everyone else.


The internet, and the assumption of the internet, allowed for more complex and long linear narratives in television, shows like Lost and Game of Thrones. The assumption of Know Your Meme, or just knowledgeable commenters in the TikTok comments, allows for less expository and more compact, obscure TikToks. TikTok comments are a form of distributed annotation.


This technique of offloading the setup for a joke to the internet allows TikTok's, or even Tweets or Instagram posts to take on a form of what I call compressed narrative.

The old format of a joke, with a setup—A man walks into the doctor's office wearing only underwear made of Saran wrap—and then the punchline—and the doctor said "I can clearly see your nuts."—is dead. The internet killed the "joke."

Instead, the internet is mostly punchline, with the barest of setup, if any. It's on you to know the context. Go Google it.

And if you still don't get it, you weren't meant to.


An example is the "I ain't ever seen two pretty best friends" meme that went around on TikTok for a hot minute and has since just become a base trope of the TikTok creative universe. Videos started taking more and more circuitous routes to end with that punchline, throwing all sorts of sleight of hands before dropping it, out of nowhere, like an M. Night Shyamalan film twist.

If you hadn't heard of the meme or didn't know the reference, these videos would be complete mysteries. Even now, if you don't know what I'm talking about, this section will make no sense. Nor will comments like "We found the two pretty best friends" on various videos.


One of the better pieces I read last year was this on the death of political humor in the age of Trump. My favorite turn of phrase from the piece is that "Irony in politics, meanwhile, has reversed its polarity." David Foster Wallace predicted the death of irony, of cynicism, after an initial boom when the internet was coming of age. The lament of the humorist is that figures like Trump are beyond the reach of irony because they are already satires of themselves.I often lament when I refer to as fortune cookie Twitter, and to combat this, I think Twitter should set up a GPT-3 bot that constantly trains on each account, and the moment most of your followers can no longer distinguish between the GPT-3 spoof of your account and your actual account, you should be forced to vacate your account and allow the GPT-3 bot to replace you. You will have literally become a parody of yourself. Also, if for some reason I ever hacked my way into a famous person's account, my goal would not to be to request BTC or post something offensive. Instead, my goal would be to post a tweet that so resembles their voice that no one, not even the person who owned that account, could tell. They'd just think, wow, that's strange, I don't remember posting that, but it is something I'd post, so ¯_(ツ)_/¯

To me, humor has always depended on creating a gap and then helping your audience to hurdle it. In a traditional joke, the gap is the space between the setup and the punchline. When the audience's mind comprehends the joke, they soar across that gap, and the exhilaration is released as laughter. You don't want to carry them across, you want to do just enough to let them take that last leap themselves.

A comedian like Chris Rock will take something from real life and just point out the hidden social truth beneath it, and your mind gets that dopamine hit of acknowledging a social fiction that you'd otherwise observe without question. Like Moses, comedians part the sea of taboo and let you stroll through, laughing all the way at being able to get away with it.

Pre-cancellation Louis C.K. also lived in this space, exposing something of your nature that you were embarrassed to acknowledge. Either he'd absolve you of your shame by absorbing it all himself in a performance of self-loathing, or he'd just forgive you that fault by making it seem universal. Comedians let you look at yourself from outside yourself, creating a gap between you and your own nature.

Trump killed humor by closing that gap entirely, becoming such a parody of himself that shows like Veep seemed less dark satire than some form of fatuous cosplay even though they came first.

But humor is not so easily killed. You just need new ways to restore the comedic gap. Much of TikTok humor is oblique in form, making references that flatter you if you understand them and puzzle you if you don't. But for the latter, you then must set off on a journey to traverse that gap. And when you've completed that journey, you get the delayed satisfaction of getting the joke but also the pleasure of now being in the in-group.

But more sophisticated creators can also play with that expectation, setting off on what seems like a familiar meme, then subverting audience expectations.


Douyin, the Chinese version of TikTok, from the same parent company Bytedance, provides an interesting contrast in the styles of humor between China and America.

A lot of comedic videos in China use a laugh track sound effect. I can't remember the last time I heard a laugh track in a TikTok.

I want to draw some conclusion here, but I don't feel confident enough. Someone more familiar with the cultural differences in Chinese and American humor might clarify this for me.


Netflix brings international programs to the U.S. TikTok brings some Chinese programming to the States also.

TikToker @funcolle makes a sort of hyper-compressed episodic detective series that is filmed in China and spoken in Mandarin, but it works on U.S. TikTok thanks to onscreen subtitles. The sound she uses is, by now, as memorable to me as the theme song to any number of popular TV series like Game of Thrones.

If you can't solve these really short single TikTok video mysteries, you can turn to the comments section to get help from all the other viewers who've pored over the videos in detail and raced to post the solution.

One measure of a platform's power is the number of things people make with it that you had never been made before. Every week, I find videos on TikTok that I can't imagine having been made on any other app.

Funcolle (@funcolle) has created a short video on TikTok with music original sound. | Anything wrong with this room? Come on, my detectives!#foryou


On TikTok, the comments have become creative terrain in their own right. Somewhere along the line, riffing on someone else's TikTok no longer required you to make a TikTok. Instead, you can just go into the comments and tack on a punchline to the punchline of the video and rack up hundreds of thousands of likes. Writing the most clever comment on a TikTok video has become its own art form.

I can't remember the last time I watched a good TikTok video without then opening up the comments to see what the peanut gallery came up with. Sometimes I read the comments before even finishing the video. TikTok's method of ranking comments almost always surfaces the best and most relevant comments to the top. However you feel about a video, it's uncanny how often one of the top five comments encapsulates it perfectly.

It's difficult in a video to feel the presence of other viewers in a tangible, meaningful way. The Twitch comment bar gives you a visible if somewhat bewildering waterfall of text as evidence of their presence, and the hearts on something like an Instagram Live or the bullet comments on Bilibili videos do the same.

TikTok comments, though, feel ike they extend the canvas of the video. Just as talent shows like The Voice require both contestants and voices to work, more and more it feels as if the TikTok experience is about watching the performers and then listening to the judges (all of us viewers) render their opinions via the comments. There isn't one Simon Cowell on TikTok, but in any comments section of any TikTok video, someone will play that role.

Never read the comments. Unless you're on TikTok, in which case, always read the comments.


Reading the comments on TikTok serves a communal function. It's like hearing the laughter of the crowd at a comedy show.

One of the existential challenges of life is truly connecting with other people's thoughts. Who can ever know that series of emotions and thoughts and dreams we call our consciousness? True human connection seems always out of grasp.

The pandemic exacerbates that sense of isolation. When most of our interactions are with flat faces on video screens, it feels either like we're living in a simulation or some solipsistic nightmare.

Before I check the comments on a TikTok I've just watched, I almost always have a strong reaction to that video. That's why opening the comments and finding that one of the first few comments perfectly encapsulates your reaction, then seeing it already has tens or hundreds of thousands of likes, is so comforting. This confirmation of a shared response creates, asynchronously, a passing score on a form of the Voight-Kampff test. It's a checksum on your humanity.

Many comments have begun using the inclusive second person singular, literally speaking for the rest of the viewers. These comments often begin with "POV:" as in "POV: You're lying bed at 2am scrolling TikTok." It's presumptive, and yet the best TikToks evoke such a consistent multiple-choice checklist of responses that it's rare the times I can think of an original comment that isn't already posted above the fold.


The sense of collective response in TikTok comments and the publicly visible view and like counts have been around long enough that users now assume enough others have encountered enough of the same memes despite everyone's FYP algorithm being tailored to their individual tastes. Many a comment on a viral TikTok will read like "Oh we're back here again."

Though I have said that TikTok isn't a social network—I don't know most people on the app, I don't have to follow anyone to have a good experience—the algorithm does create, through its efficient sorting, a sense of traveling through subcultural neighborhoods as you scroll down one TikTok at a time.

Users have adopted spatial or geographic language to describe this sense of shared viewing spaces. Various subcultures are described by appending -tok or TikTok behind a descriptor. Someone commenting on a particularly high-quality video might say "I've finally gotten Premium TikTok." People share weird niches they're on by saying things like "I'm deep into carpet cleaning tok" or "I don't know how but I've found music theory tok." Sometimes it's just one word, like "Sportstok or Liberaltok." Tok has almost come to be a suffix meaning "neighborhood" or "community," almost like Disney uses -land to describe themed areas in its parks like Frontierland or Tomorrowland.

Of course, we're all just in our FYP feeds, which just scrolls up endlessly, so it isn't an actual space. But we trust the visible view counts as evidence FYP is doing its job getting many of us with the same tastes in front of the same videos, and so this evidence of common knowledge creates a liminal third place that exists [waves hands at the air in front of me] out there.

I’ve tended to think of social networks as being built by people assembling a graph of people bottoms up, but perhaps I’ve been too narrow-minded. TikTok might not qualify by that definition, but it feels social, with FYP as village matchmaker.


There's been a lot written on Warner Media's decision to move some films from theatrical only windows to having a concurrent release on HBO Max. A lot of conclusions were drawn about theatrical's future based on Wonder Woman 84’s Christmas premiere in theaters and on HBO Max day-and-date. A lot of it is the usual knee jerk extrapolation that the internet is famous for, despite confounding circumstances like a pandemic, and despite Wonder Woman 84 being a single data point.

But one thing I'm confident of is that something is lost in not having the audible feedback of a hundred or more humans around you when you watch something, especially from genres that are built to elicit frequent emotional feedback, like comedies and horror films. At some point, perhaps we'll crack the nut on social viewing and how to make it more, umm, social, but for now, pre-VR metaverse, it's a shoddy facsimile of a crowd.

Look, I've streamed my share of concerts during this past year, and I don't miss standing for an hour between sets in a crowded club or bar, nursing a $9 beer in a plastic cup, waiting for my band of choice to get on stage.

And yet, I miss standing in that bar, my shoes sticking to the beer-soaked floor, trying to look at ease in my own skin while gawking at other humans.


In a year where we've been trapped inside for nearly a year now, there's something about the chaotic collectivist media art form that is TikTok that felt most joyful and genuine.

Thumbing through the FYP feed one portrait-oriented rectangle at a time felt like swiping from one bedroom window to the next on a tall skyscraper, peering into one user's bedroom after another (literally, as the bedroom is the most common space in which teens do their creative work). It's like a Chris Ware comic strip, with its architectural design, navigated one window pane at a time.

Because it's full screen, it can feel like my phone screen is literally a rectangular porthole. As if one user after another is hijacking my rear-facing camera and turning it into their rear or front-facing camera.

There's something about media like TikTok or ChatRoulette or Omegle, where so much of what you see is a creator directly addressing the camera, breaking the fourth wall from the start, that is immediately intimate.


One thing I wish TikTok would do is make it easier to trace multi-part videos from creators. Nothing drives me crazier than videos that end with "Stay tuned for Part 2" or "Like for Part 2" and then you spend like ten minutes browsing their profile trying to find the second part.

I understand that it's a sort of view count hack on the part of creators, but some videos do need to be broken up across installments. TikTok needs to add some sort of concept of a pointer or link to make it easier to jump directly to the next installment in a series. Perhaps it could be done via a playlist feature.

For now, the best way to trace linked videos is to visually scan the thumbnails on a person's profile and search for onscreen text reading "Part #" or just click on every video with the same visual grammar, the user in the same outfit in the same room with the same lighting.

(Since I wrote the note above, the app has added a way to highlight, on a creator's profile page, the video you just watched, and since videos are sorted reverse chronologically by creation date on the profile, often part II can be found next to the video you just watched, which is handy).


In another example of the community coming up with creative solutions, commenters on the first in a multi-part video series where the next part has yet to be published will now leave a comment saying that they'll promise to tag people who like their comment once the next installment is posted. In other words, users are serving as Mechanical Turk notification bots.


Another feature I wish TikTok would add is the ability to sort by descending popularity on any grid of videos, like on sound or profile pages. Please.


TikTok's needs to improve its search ranking algorithm. Trying to find popular TikTok's I remembered seeing back in the day was much harder than it should have been using TikTok's native search. A couple that I wanted to use I just couldn't locate, and even Google and YouTube didn't turn them up (a thing you realize after trying to do it more than once is how hard it is to create a comprehensible search query for certain TikTok's).


Network effects are powerful, but there are so many distinct types. It's important to understand exactly what type of network effect you have because they all scale and operate differently.

For example, Dunbar's number is just one form of limit on a very specific type of network effect. But there are dozens and dozens of network effects, all with their distinct quirks. Someone could make a lot of money just making a reference book of the taxonomy of network effect varietals in the world.

TikTok is an extreme experiment in not only making creative network effects endogenous to its app but to the medium of video. Like some video Minecraft, almost everything in the app is a replicable chunk of bits that you can combine into a larger configuration of bits, and the resulting creation becomes, itself, a chunk that anyone can take and splice or mutate or combine however they want.

This is anathema to old media, most of whom have spent hundreds of millions of dollars to lock up their content behind copyright law, DRM, and any number of other mechanisms meant to slow the rate of reproduction and iteration of their work. It has the effect of slowing the evolutionary feedback loops on all of that work.

TikTok's "OODA loop" is collective and distributed, and it spins thousands of times faster than that of big media.


When I first joined the Amazon Web Services team in 2003, it was still a small Jeff Bezos-sponsored project. There were only some 15 people or so on the team at the time under the leadership of now Amazon CEO Andy Jassy.

A book Jeff had us read, one which he said should serve as an inspiration for how we'd design AWS, was Creation: Life and How to Make It by Steve Grand. It's a book about programming artificial life, but the core principle that Jeff wanted us to take from it was the idea that complex things like life forms are built from very simple building blocks or primitives. It's the same thesis as that in Stephen Wolfram's A New Kind of Science.

The key implication for AWS from the book was about how to design the first AWS primitives. Jeff urged us to include only what was necessary and nothing more. If you were designing a storage service, like S3, you'd need functions like get, write, delete, but you wouldn't want to layer in things that weren't part of storage, like security. That should be a separate primitive.

The reason to design your primitives with the utmost elegance is to maximize combinatorial optionality.


This is one of the most elegant things about TikTok's design. It includes a ton of primitives, and they are almost all ones you can combine or link.

More than that, every element in a TikTok is a building block you can replicate and use in your own TikTok. The most important of these is the soundtrack or sound of your TikTok.


Be careful of taking this idea of building primitives too far. In many ways, choosing what level of abstraction to stake your ground on is one of the most important questions any company must answer.

The answer is contextual. Abstract at too high a level and someone can come in beneath you, with something like AWS. In some ways this is a form of disruption.

Build at too low a level, however, and often someone will abstract at a level above you and siphon all the value of that market above your product. Many of TikTok’s filters are abstractions of a lot of things, almost like Lightroom Presets. As many of us learned early in this pandemic, maybe paying a few bucks for a loaf of bread is preferable to having to spend hours of our free time mastering baking.


When I think about modern remix culture and apps like TikTok, I often think back to Mixel, an app designer Khoi Vinh launched years ago. It was an iPad collage app.

In his blog post introducing Mixel, Vinh wrote:

Because of the componentized nature of collage, we can add new social dimensions that aren’t currently possible in any other network, art-based or not. Mixel keeps track of every piece of every collage, regardless of who uses it or how it’s been cropped. That means, in a sense, that the image pieces within Mixel have a social life of their own. Anyone can borrow or re-use any other piece; you’re free to peruse all the collages (we call them “mixels”) and pick up literally any piece and use it in your own mixel. If you don’t like the crop, the full, unedited original is stored on the server, so you can open it back up in an instant and cut out just the parts you like. Mixel can even show you everywhere else a particular image has been used, so you can follow it throughout the network to see how other people have cropped it and combined it with other elements.

The thread view turns collaging into a visual conversation, where anyone can remix anyone else’s work.


Though Mixel is no longer around, what he describes presages modern meme culture and TikTok.

Inherent to digital culture is the remix.


In Mark Ronson's TED Talk on How Sampling Transformed Music, he says:

That's what the past 30 years of music has been. That's the major thread. See, 30 years ago, you had the first digital samplers, and they changed everything overnight. All of a sudden, artists could sample from anything and everything that came before them, from a snare drum from the Funky Meters, to a Ron Carter bassline, the theme to "The Price Is Right." Albums like De La Soul's "3 Feet High and Rising" and the Beastie Boys' "Paul's Boutique" looted from decades of recorded music to create these sonic, layered masterpieces that were basically the Sgt. Peppers of their day. And they weren't sampling these records because they were too lazy to write their own music. They weren't sampling these records to cash in on the familiarity of the original stuff. To be honest, it was all about sampling really obscure things, except for a few obvious exceptions like Vanilla Ice and "doo doo doo da da doo doo" that we know about. But the thing is, they were sampling those records because they heard something in that music that spoke to them that they instantly wanted to inject themselves into the narrative of that music. They heard it, they wanted to be a part of it, and all of a sudden they found themselves in possession of the technology to do so, not much unlike the way the Delta blues struck a chord with the Stones and the Beatles and Clapton, and they felt the need to co-opt that music for the tools of their day. You know, in music we take something that we love and we build on it.


One of the most revolutionary aspects of TikTok is how effortless it makes it to sample or interpolate any other TikTok video.


Anyone who's used a non-linear editor like Adobe Premiere, Final Cut Pro, or Avid Media Composer knows the standard multi-pane interface. And any editor knows that editing begins with importing all the media, the shots from dailies, the temp music, and so on, into your media bin. From there, you drag elements onto the timeline to compose the edit.

Much of the pain of creating memes is gathering all the components, like images, from the web. In the modern networked age, though, the media bin should really just be the entirety of the internet. Anything you want should just be a short search away. We're starting to get closer, though the library of material is still sparse, and many pieces, especially video, still require chasing down.

Someday, any sort of remix will just be a GPT-3 like interface away from composing. You'll just be able to write "This is Fine cartoon but the dog's face is Donald Trump" and it will just spit it out for you. If you're building this, please let me know, I'll write you a seed check.


The Verge interviewed a TikTok beatmaker named Ricky Desktop.

What makes a great TikTok beat?

You need concrete, sonic elements that dancers can visually engage with on a person-by-person basis. I know that sounds super scientific, but that is how I think about it. If you’re trying to make a viral beat, it’s got to correspond with the viral dance.

In order to lock in on that, you need elements of the music to hit. So for example, I have this beat called “The Dice Beat.” I added a flute sound, which in my head was like, “Okay, people will pretend to play the flute.” And then there’s the dice sound, where they’ll roll the dice. It was super calculated. I would create the music with the dance in mind.

I developed this little pattern. I pioneered the “triple woah” thing where in all the beats there’s three kicks — bum-bum-bum. So typically, when the bass drop hits, the dancers will do the woah (ed: an accentuated arm and elbow movement popular in TikTok dances) to emphasize the bass drop. Usually, the beat will keep going after that. But what I did, I would add three more bass hits, super calculated, so that dancers could do the woah three times or do three concrete dance accents.


The woah inspired Ricky Desktop to develop a score for the triple woah which then actually inspired dancers to choreograph and perform an actual triple woah.

Can you program human movement with music? It turns out you can. You use an API called TikTok. That's delightful.


TikTok beatmaker Ricky Desktop pictured, in his head, dancers performing some movement. Then he wrote a piece of music that included a musical cue intended to elicit that exact movement.

Then, later, some dancers on TikTok performed the movement he had pictured, exactly at the moment he had inserted the musical prompt. It's not just that he choreographed the human body via music, but how he did it. Ricky Desktop is a marionettist manipulating human bodies not via strings but music.


Ricky Desktop:

So I would post my beat and say, “Anyone trying to help me make this beat go viral?” Or I would say, “Who’s gonna create a dance to this new banger?” I’m giving an action item to whoever’s watching. And that’s important because it gives the person watching something to do.


The message is in the medium. That is, Ricky Desktop issues these to-dos inside of the video he uses to release his various sounds.


Ricky Desktop:

What makes a great TikTok beat?

You need concrete, sonic elements that dancers can visually engage with on a person-by-person basis. I know that sounds super scientific, but that is how I think about it. If you’re trying to make a viral beat, it’s got to correspond with the viral dance.

In order to lock in on that, you need elements of the music to hit. So for example, I have this beat called “The Dice Beat.” I added a flute sound, which in my head was like, “Okay, people will pretend to play the flute.” And then there’s the dice sound, where they’ll roll the dice. It was super calculated. I would create the music with the dance in mind.


In filmmaking, when you want a score for your film, you bring the latest cut of your film to a composer's studio, and they start riffing based on what they see on screen, incorporating some of the themes you're trying to evoke in that scene.

What Ricky Desktop talks about above is a different process in which he scores to visuals that only exist in his imagination, generic dance tropes like "pretend to play the flute".

This is a form of "inverted scoring." Or, if you prefer to go from the other direction, what TikTok dancers do with sounds is "visualizing."

The program WinAmp used to do software visualizations of music. TikTok is like Mechanical Turk for visualizing music.


If you've watched any amount of TikTok, you've doubtless seen someone answering questions by dancing and pointing to floating text overlays.

Now, they could easily just speak the questions and answer them verbally. There's no reason to have to dance to music while answering the questions.

To which I say, no one knows what it means, but it's provocative, it gets the people going!

Kylie Jenner (@kyliejenner) has created a short video on TikTok with music original sound. | i'm still a supermodel on the inside | INSTAGRAM MODEL | SUPERMODEL | DADS FAV MOMS FAV | ...

This is one of many TikTok survey or poll formats, all devised by the users. On one hand, there are simpler ways to share this information. On the other hand, this is much more entertaining than a Twitter poll.


On the other hand, maybe all this choreographed dancing is something more of us should be doing to make our messages land. A teacher went viral on TikTok this year for filming herself trying to teach her class remotely over Zoom. Seeing her precise and broad gestures paired with her sharply articulated speech, you couldn't help but feel empathy for what a burden we've placed on our teachers, trying to make remote classes engaging over Zoom.

But perhaps we just lose some of our childlike exuberance and joy expressiveness as we age? Perhaps if we were more animated in our delivery, more people would remember what we said.

One of the most common weaknesses among managers and leaders is the illusion of transparency, though it is a problem for most people. It is the tendency to overestimate how much people know what you're thinking. It can ruin marriages or relationships, and it leads to a healthy market for therapy.

Young children have the a strong form of this illusion which is why in early childhood they are so frustrated when you don't understand why they're upset (and parents are likewise just as exasperated that their children can't verbalize why they're freaking out). Until later in life, children think you should know exactly what they're feeling, and it takes a bit of coaxing to tease out their inner emotional state. Ironically, despite their illusion of transparency, kids tend to be much more emotionally transparent and thus expressive.

It's when they finally realize that no one can see into their heads that they learn to lie. It's then that you wish they still had the illusion of transparency. When they become teenagers, the battle over transparency into their lives becomes literal: parents yell at their teenagers to keep their bedroom doors open, and those same doors slam shut after heated arguments. Their bedrooms become, like their thoughts, spaces they wish to protect from prying eyes.

This is all a roundabout way to say that a CEO communicating a company's top goal for the coming year in a TikTok dance, pointing to on-screen captions, isn't the worst idea in the world? Maybe this is the new Amazon 6-page memo.


Study any high-level memory competitor and they'll all say the same thing. Humans' visual memories are far superior to their memories for abstractions. It's one of the core lessons from the great book Moonwalking with Einstein. It's the reason people who have to try to memorize a thousand digits of pi or the order of a deck of cards turn numbers and letters into images which they place spatially in memory palaces.

In its heyday, which coincided with my childhood, MTV was dominated by music videos, and each of those was essentially a visualization of a musical track. To this day, I can't hear a song like A-Ha's "Take on Me" without picturing its music video. I haven't seen it in decades, but its cartoon sketches come to life are forever how I "see" the song. Likewise, I can't hear Michael Jackson's Thriller without conjuring its epic music video of nearly 14 minutes.

It doesn't even have to be a music video. A song incorporated in a film can permanently bond with the moving images on the screen. For example, I can't hear three tracks, one each by Huey Lewis, Genesis, and Whitney Houston, respectively, without picturing Christian Bale and quoting Patrick Bateman, and then being filled with a sense of self-loathing for having been indicted as someone who turns to the appreciation of cultural artifacts as a substitute for personality. If I mention Celine Dion's song "My Heart Will Go On," what do you see in your mind's eye?

TikTok is the modern MTV because (1) it increases consumption of music tracks that go viral on its platform as sounds and (2) any number of songs will forever summon the accompanying meme and visual choreography from my memory.


When Charli and other TikTokers formed the Hype House in Los Angeles, they were experimenting with IRL creative network effects. They created what was efffectively a commune to produce the D'Amelio TikTok Universe with Charli at the center as, I don't know, Tony Stark or something.

They started guest-starring in each other's TikTok's, some of them started dating and hooking up, and soon, to follow the entire extended narrative, you had to follow each other's accounts. Studios have tried to push out fictional versions of such networked series, but Charli et al just created it bottoms up, with TikTok as the distributor.

The Kardashian-Jenner clan are the clear predecessors who ran this type of crossover mindshare grab, but they're family. This new generation of influencers often aren't related, their common bond is just that they're young and famous in the age of social media and so they already all live together in a virtual universe held together by the gravity of popularity.


In Status as a Service, I wrote about how social networks require some proof of work to gain status.

A lot of TikTok's have the caption "I spent way too long on this" as a sort of plea for likes, but that wouldn't land if the proof of work wasn't visible on the screen. It is, and even non-creators can see it. Some TikToks seem like they took days to produce.


Have you tried using the in-app TikTok video editor? In some ways, it's loaded with really first-rate filters and effects, but in many ways, its user interface is inscrutable. I went to editing school and have used a variety of non-linear editors (NLEs) like FCP, Premiere, and Avid to edit video in a previous life, and I still tear my hair out trying to use TikTok's native editor.

The easiest videos to make are just ones where you film yourself live and apply a filter, but if you want to bring in pre-recorded video and mix them with other graphical elements, like text boxes, it is very painful to assemble them properly. My kingdom for a persistent timeline with a scrubber in the TikTok editor.

In one sense, it's staggering to ponder how many more videos TikTok would have if its video editor were more usable. On the other hand, every video that does make it onto the app feels like a miracle. The proof of work is in the pain.


If you're a movie star like Will Smith and you get a VFX studio to produce some whiz-bang TikTok for you, it will feel off, like driving a Ferrari down the street in Omaha. Authenticity or at least the sheen of one's own sweat equity is part of the TikTok aesthetic, and the canonical backdrop for any TikTok video is always some teenager's somewhat messy bedroom, just as it was in the heyday of the YouTube vlog.

On Instagram, you can get away with proof of wealth, but the TikTok aesthetic is proof of creative labor. The verdict is a bit more mixed on proof of hotness, though. I still think Instagram is a more welcoming home for pure thirst trap content than TikTok, where, if you want to honeytrap the simps, you're going to have to dance for it.


Something about a feed that can hit you with such a variety of styles and moods in such quick succession makes TikTok feel like the most modern of media channels. One second you're watching a dog communicate with their owner using a language mat, the next second some high school girl is roasting one of her classmates, the next you see a teen making an earnest confessional deprecating their own looks (only to have thousands of commenters offering affirmation), and then you might see a boat chase that you later realize is some drug cartel member filming a TikTok as police boats give chase (even Narcos be chasing them likes). At times it feels as if the FYP feed is a pastiche generator.

It is equal parts ironic and earnest, having long since surpassed its label as the cringey social network.

Whereas Instagram is performative, TikTok is performative and self-aware. It’s not that any single creator is self-aware, but that the Greek chorus in the comments will descend on anyone with the slightest bit of hubris like a pack of harpies.

In this rectangular proscenium that is TikTok, the fickle god of Zeus is played, of course, by the FYP algorithm. Everyone offers up their sacrifices of time and labor in the hopes of being graced by its favor, but its whims remain just capricious enough to keep everyone grinding.


If your FYP feed is dialed in to your tastes, you start to pre-react to videos purely based on the like count visible on the right hand side of the screen.

If a video has a high like count, even if it starts slowly I'll tend to give it the benefit of the doubt and stick around to the end, simply because this statistic has proved, in my experience, reliable evidence of a worthwhile payoff. The larger the figure, the more I anticipate a strong punchline or close. I'm like Tom Cruise in Minority Report, already having seen the precog verdict printed on that ball.

Conversely, when I'm the test audience for a little-seen video (a dead giveaway is it has almost no likes yet), I tend to be merciless in skipping ahead if it doesn't hold my attention after a few seconds.

This creates a ruthless rich-get-richer dynamic, but that's by design. Bytedance as a company has built its products around pitiless algorithms enforcing a high Gini coefficient economy of entertainment. It's a marketplace in which the supply side—the TikTok videos from creators—can be shown to an unlimited number of viewers. Much of the content is evergreen, so there is almost no end to the leverage TikTok can get off any single good video.

Imagine if YouTube's key metric was to show every good video in its entire catalog to every viewer that would enjoy it. If you view the TikTok mission that way, even if no one submitted another new video for the next year, its FYP algorithm would still have an almost infinite supply of short videos to show to hundreds of millions of users for that entire dry spell.


Because sounds become the genesis of particular memes, when you start watching a TikTok video and hear a familiar sound, you anticipate the moment of that sound when the punchline will happen. It's Pavlovian.

The kismet shoe transition, for example, causes you to anticipate the pleasure of that exact moment when the performer will go from looking plain to looking EXPENSIVE. There are only so many plots in Hollywood, but we go see genre films precisely for the story beats we know are coming.

On TikTok, sounds and memes are almost inseparable. The sound is the meme is the sound.

TikTok sounds are often the most pleasing snippets from pop songs, and listening to one catchy loop after another is like listening to a pop radio channel that doesn't play entire songs, only plays bass drops and choruses. The time between anticipation and payoff is so short that scrolling the feed can feel like pressing the button on some sonic IV drip over and over. Just inject it into my ears.


In Infinite Jest, David Foster Wallace describes a film called Infinite Jest which is so entertaining people lose all will to do anything except watch it until they die. He had often written about the addictiveness of television and may have been extrapolating to the future, projecting the entertainment value of entertainment increasing until it surpassed some threshold where you'd lose all will to do anything except consume. In that way, he predicted binge watching.

But the earliest form of entertainment that conjured the addictive properties of his fictional film (referred to in the novel as "the Entertainment") was video games. I read stories about players who died after playing games for so long without eating and, recalling some game binge sessions from my youth, could imagine myself trapped in a similar dark loop.

TikTok is the second form of entertainment that brings DFW's fictional entertainment to mind. In hindsight, it seems obvious that a personalized feed of video, tailored to your tastes, would be the addictive end state of entertainment. And, considering the rise of social media and the smartphone, it would make sense that the videos might all be short, like pellets of rain, sliding comfortably into every spare pocket of time in our day, of which we have so many.

One of my favorite paragraphs of recent years was one describing the miracle that are Cheetos:

To get a better feel for their work, I called on Steven Witherly, a food scientist who wrote a fascinating guide for industry insiders titled, “Why Humans Like Junk Food.” I brought him two shopping bags filled with a variety of chips to taste. He zeroed right in on the Cheetos. “This,” Witherly said, “is one of the most marvelously constructed foods on the planet, in terms of pure pleasure.” He ticked off a dozen attributes of the Cheetos that make the brain say more. But the one he focused on most was the puff’s uncanny ability to melt in the mouth. “It’s called vanishing caloric density,” Witherly said. “If something melts down quickly, your brain thinks that there’s no calories in it . . . you can just keep eating it forever.”

TikTok is entertainment Cheetos. Each video requires so little cognitive exertion and reaches its climax so quickly that it feels like we could keep watching forever, each punch line scored to the most satisfying bass drop or stanza from every pop song. TikTok delivers dopamine hits with a metronomic rhythm, and as soon as we swipe up the previous one melts in our memory.


It's always been the case, but especially in this networked age, that every piece of entertainment is its own social network. The network effects of a story arise from shared consumption. The more people watch Star Wars, the more people I can talk to about particular scenes or compare costumes with at a convention. The more people that watch Game of Thrones, the more my Game of Thrones memes will land.

TikTok is personalized, yet through its algorithm it creates shared stories of real scale. Some of these shared stories occur on the creative side in duets and trims that connect creators to each other literally and metaphorically. The FYP algorithm also aggregates large communities of viewers for the hottest TikTok videos. It's not uncommon now for me to send a TikTok to a friend who's already seen it, or vice versa. Not always, but enough that the audience now assumes enough common knowledge to foster that sense of shared experience.

Despite having what must be a gazillion videos in its catalog, watch TikTok enough and you'll be able to refer to something like Sea Shanty TikTok and feel reasonably confident other TikTok addicts get the reference. In contrast, people regularly send me YouTube videos with like millions of views that I've never even heard of.

It is algorithms that may be tearing us apart. But maybe it's also algorithms that reassemble us, albeit in smaller unit sizes. 330M Americans feel like too large an optimal governance size if we're going to let social media algorithms just run amok, but I find some comfort sometimes when I find some TikTok that feels so catered to my tastes that it must be a micro-niche and then see it has millions of likes.


The term binge-watching typically refers to watching multiple episodes of a series in one sitting, but perhaps the act of watching dozens of TikTok videos in a row is the purest form of this type of entertainment gluttony.

Other types of social media like Instagram and Twitter are also series of really compact units of media. When I scroll Twitter or Instagram, I often feel like an elephant, standing there placidly, as various people toss individual packing peanuts at my forehead (let’s call these people the peanut gallery?).

TikTok videos are, for the most part, a bit longer. Their compressed narratives are still, nevertheless, complete, with some full story arc to traverse. In its rhythm, binge-watching TikTok reminds me of watching a standup comedy set, but instead of watching one comedian, I’m watching a whole series of them, each on stage just long enough to tell one joke. And if they bore me, I can press a button and, like a Looney Tunes cartoon, a cane whisks them off the stage and a new comedian pops out from the floor to take their place and start right into their joke.

Someone told me that if you watch TikTok for over an hour it posts a warning asking you to consider taking a break. I'm not sure if that's the case, but I'm glad I've never encountered it yet.


TikTok can only match you with videos it has, and for some people, there may not be enough relevant content in the TikTok catalog to sustain a feed. But that pool of videos has grown by an astonishing amount in a short amount of time.

I'm an easy mark for the sort of wry, sometimes savage humor of TikTok, especially when it skews almost post-modern in its awareness of its own form. It's both a community that constantly tries to legislate its own social norms of decency—any video of someone making fun of how they look using a supposed beauty filter will be flooded with comments like "You're a queen", the comments section being sort of a rolling floor vote on what the acceptable response is—and also a bloodbath of Gen Z violence. The kids will be alright, but that's in part because they're savage. Every generation learns it has to fend for itself.

During a pandemic when most of social media feels even more nakedly performative than usual, as we sit inside day after day for month after month, my occasional sessions on TikTok have been one of the only pastimes to reliably make me laugh, and it's not particularly close.

Twitter has reached a crest of fortune cookie thinkboi bait when it's not subsumed in petty high school lunchroom culture war fistfights. Seemingly every day, a playground brawl breaks out and we all form a circle to gawk, but at the back of our minds is always the threat that we'll be the next to be sucker-punched and forced to throw down. Outrage porn is exhausting and also not that fun?

When viewed from the eye of a global pandemic, Instagram feels like a horrifying Truman Show of idyllic capitalist showboating. Life must go on, influencers gotta influence, but I'm also not weeping any tears when people get chastised for renting private islands and posting photos of themselves partying during a pandemic.

Andrew Niccol, the screenwriter of The Truman Show, once said, "When you know there is a camera, there is no reality.” The most absurd but popular tag on visual social media is #nofilter, a hashtag that aspires to a pretense of truth when there is almost nothing on an app like Instagram that isn’t production-designed within an inch of its life.

TikTok, by virtue of its high bar to even produce a video that anyone will see (FYP algo is like "That's a no for me dawg" on almost every video), is upfront about what it is: a global talent show to entertain the masses. In a pandemic where much of the U.S. lives in eternal lockdown, TikTok is the 24/7 channel where the American Idle entertain each other from their bedrooms. I laughed, and then I laughed some more.

TikTok and the Sorting Hat

I often describe myself as a cultural determinist, more as a way to differentiate myself from people with other dominant worldviews, though I am not a strict adherent. It’s more that in many situations when people ascribe causal power to something other than culture, I’m immediately suspicious.

The 2010’s were a fascinating time to follow the consumer tech industry in China. Though I left Hulu in 2011, I still kept in touch with a lot of the team from our satellite Hulu Beijing office, many of whom scattered out to various Chinese tech companies throughout the past decade. On my last visit to the Hulu Beijing office in 2011, I was skeptical any of the new tech companies out of China would ever crack the U.S. market.

It wasn’t just that the U.S. had strong incumbents, or that the Chinese tech companies were still in their infancy. My default hypothesis was that what I call the veil of cultural ignorance was too impenetrable a barrier. That companies from non-WEIRD countries (Joseph Henrich shorthand for Western, Educated, Industrialized, Rich, and Democratic) would struggle to ship into WEIRD cultures. I was even skeptical of the reverse, of U.S. companies competing in China or India. The further the cultural distance between two countries, the more challenging it would be for companies in one to compete in the other. The path towards overcoming that seemed to lie in hiring a local leadership team, or sending someone over from the U.S. who understood the culture of that country inside-out.

For the most part, that has held true. China has struggled, for the most part, to make real inroads in the U.S. WeChat tried to make inroads in the U.S. but only really managed to capture Chinese-Americans who used the app to communicate with friends, family, and business colleagues in China.

In the other direction, the U.S. hasn’t made a huge dent in China. Obviously, the Great Firewall played a huge role in keeping a lot of U.S. companies out of the Chinese market, but in the few cases where a U.S. company got a crack at the Chinese market, like Uber China, the results were mixed.

For this reason, I’ve been fascinated with TikTok. Here in 2020, TikTok is, for many, including myself, the most entertaining short video app going. The U.S. government is considering banning the app as a national security risk, and while that’s the topic du jour for just about everyone right now, I’m much more interested in tracing how it got a foothold in markets outside of China, especially the U.S. with its powerful incumbents.

They say you learn the most from failure, and in the same way I learn the most about my mental models from the exceptions. How did an app designed by two guys in Shanghai managed to run circles around U.S. video apps from YouTube to Facebook to Instagram to Snapchat, becoming the most fertile source for meme origination, mutation, and dissemination in a culture so different from the one in which it was built?

The answer, I believe, has significant implications for the future of cross-border tech competition, as well as for understanding how product developers achieve product-market-fit. The rise of TikTok updated my thinking. It turns out that in some categories, a machine learning algorithm significantly responsive and accurate can pierce the veil of cultural ignorance. Today, sometimes culture can be abstracted.


TikTok's story begins in 2014, in Shanghai. Alex Zhu and Luyu “Louis” Yang had launched an educational short-form video app that hadn’t gotten any traction. They decided to pivot to lip-synch music videos, launching Musical.ly in the U.S. and China. Ironically, the app got more traction across the Pacific Ocean, so they killed their efforts in their home country of China and focused their efforts on their American market.

The early user base consisted mostly of American teenage girls. Finally, an app offered users the chance to lip synch to the official version of popular songs and have those videos distributed to an audience for social feedback.

That the app got any traction at all was progress. However, it presented Alex, Louis, and their team with a problem. American teen girls were not exactly an audience Alex and Louis really understood.To be fair, most American parents would argue they don't understand their teenage daughters either.

During this era where China and the U.S. tech scenes have overlapped, the Chinese market has been largely impenetrable to the U.S. tech companies because of the Great Firewall, both the software instance and the outright bans from the CCP. But in the reverse direction, America has been almost as impenetrable to Chinese companies because of what might be thought of as America’s cultural firewall. Outside of DJI in dronesI'd argue one reason DJI had success in America was that drone control interfaces borrow heavily from standard flight control interfaces and are not culturally specific. Thus DJI could lean on its hardware prowess which was formidable., I can’t think of any Chinese app making real inroads in the U.S. prior to Musical.ly. To build on its early traction, Musical.ly would have to overcome this cultural barrier.

It’s been said that if you ask your customers what they want, they’ll ask for a faster horse (attributed to Henry Ford, though that may not be true). Frankly, that’s always been half horses***, and not just because horses are involved. First of all, what if your customers are horse jockeys?

Secondly, while you can’t listen to your customers exclusively, paying attention to them is a dependable way to build a solid SaaS business, and even in the consumer space it provides useful signal. As I’ve written about before, customers may tell you they want a faster horse, and what you should hear is not that you should be injecting your horses with steroids but that your customers find their current mode of transportation, the aforementioned horse, to be too slow a means of getting around.

Alex and Louis listened to Musical.ly’s early adopters. The app made feedback channels easy to find, and the American teenage girls using the app every day were more than willing to speak up about what they wanted to ease their video creation. They sent a ton of product requests, helping to inform a product roadmap for the Musical.ly team. That, combined with some clever growth hacks, like allowing watermarked videos to easily be downloaded and distributed via other networks like YouTube, Facebook, and Instagram, helped them achieve hockey-stick inflection among their target market.

Still, Musical.ly ran into its invisible asymptote eventually. There are only so many teenage girls in the U.S. When they saturated that market, usage and growth flatlined. It was then that a suitor they had rebuffed previously, the Chinese technology company Bytedance, suddenly looked more attractive, like Professor Bhaer to Jo March at the end of Little Women. In a bit of dramatic irony, Bytedance had cloned Musical.ly in China with an app called Douyin, one that had taken off in China, and now Bytedance was buying the app that inspired it, Musical.ly, an app conceived and built in China but that had failed in China and instead gotten traction in the U.S.

After the purchase, Bytedance rebranded Musical.ly as TikTok. Still, if that’s all they had done, it’s not clear why the app would’ve broken out of its stalled growth to the stunning extent it has under its new owner. After all, Bytedance paid just $1B for an app that’s rumored to sell now, if the U.S. government approves the transaction, for anywhere from $30 to $70B.

Bytedance did two things in particular to jumpstart TikTok’s growth.

First, it opened up its wallet and started spending on user acquisition in the U.S. the way wealthy Chinese used to spend on American real estate (no, I’m not still bitter at all the Chinese all-cash offers that trounced me repeatedly when condo-hunting six years ago). TikTok was rumored to have been spending a staggering eight or nine figures a month on advertising.

The ubiquity of TikTok ads lent the theory credence. I saw TikTok ads everywhere, on YouTube, Instagram, Twitter, Facebook, and in mobile gamesTikTok ads are bizarre. The video ads I see for the app in mobile games convey nothing about what the app is or does. One ad I've seen dozens of times has an old lady doing lunges in her living room, another has a kid blow drying his hair, and as he does, his hair changes colors. I feel like the ads could do a better job of selling the app, but what do I know?. If Bytedance could have purchased ads on the back of my eyelids at sub $20 CPMs I don’t doubt they would have done so.

It didn’t look like a wise investment at first. Rumors abounded that the 30-day retention of all those new users poured into the top of its funnel was sub 10%. They seemed to be lighting ad dollars on fire.

Ultimately, the ROI on that spend would turn the corner, but only because of the second element of their assault on the US market, the most important piece of technology Bytedance introduced to TikTok: the updated For You Page feed algorithm.

Bytedance has an absurd proportion of their software engineers focused on their algorithms, more than half at last check. It is known as the algorithm company, first for its breakout algorithmic “news” app Toutiao, then for its Musical.ly clone Douyin, and now for TikTok.

Prior to TikTok, I would’ve said YouTube had the strongest exploit algorithm in video,The exploit versus explore conundrum is sort of a classic of algorithmic design, usually mentioned in relation to the multi-armed bandit problem. For the purposes of this discussion, think of it simply as the problem of choosing which videos to show you. An exploit algorithm will give you more of what you like, while an explore algorithm tries to broaden your exposure to more than just what you’ve shown you like. YouTube is often described as an exploit algorithm because it tends to really push more of what you like, and then before you know it, you’re looking at some alt-right video that’s trying to redpill you. but in comparison to TikTok, YouTube’s algorithm feels primitive (the top creators on YouTube have long ago figured out how to game YouTube’s algorithm’s heavy dependence on click-through rates and watch time, one reason so many YouTube videos are lengthening over time, much to my dismay).

Before Bytedance bought Musical.ly and rebranded it TikTok, its Musical.ly clone called Douyin was already a sensation in the Chinese market thanks in large part to its effective algorithm. A few years ago, on a visit to Beijing, I caught up with a bunch of former colleagues from Hulu Beijing, and all of them showed me their Douyin feeds. They described the app as frighteningly addictive and the algorithm as eerily perceptive. More than one of them said they had to delete the app off their phone for months at a time because they were losing an hour or two every night just lying in bed watching videos.

That same trip, I had coffee with an ex-Hulu developer who now was now a senior exec in the Bytedance engineering organization. Of course, he was tight-lipped about how their algorithm worked, but the scale of their infrastructure dedicated to their algorithms was clear. On my way in and out of this office, just one of several Bytedance spaces all across the city, I gawked at hundreds of workers sitting side by side in row after row in the open floorplan. It resembled what I’d seen at tech giants like Facebook in the U.S., but even denser.The mood was giddy. I could tell he was doing well. He took me and my friends to a Luckin Coffee in their office basement and told us to order drinks off an app on his phone. I reached in my pocket for some RMB to pay for the drinks and he put his hand on my arm to stop me. “Don’t worry, I can afford this,” he said, laughing. He didn’t mean it in a boasting manner, he seemed almost sheepish about how well they were doing. Afterwards, as we waited outside the office in their parking lot, he walked past and asked me if I needed a ride. No, I said, I’d be taking the subway. A Tesla Model X pulled up, the valet hopped out, and he jumped in and drove off.

It’s rumored that Bytedance examines more features of videos than other companies. If you like a video featuring video game captures, that is noted. If you like videos featuring puppies, that is noted. Every Douyin feed I examined was distinctive. My friends all noted that after spending only a short amount of time in the app, it had locked onto their palate.

That, more than anything else, was the critical upgrade Bytedance applied to Musical.ly to turn it into TikTok. Friends at Bytedance claimed, with some pride, that after they plugged Musical.ly, now TikTok, into Bytedance’s back-end algorithm, they doubled the time spent in the app. I was skeptical until I asked some friends who had some data on the before and after. The step change in the graph was anything but subtle.

At the time Musical.ly got renamed TikTok, it was still dominated by teen girls doing lip synch videos. Many U.S. teens at the time described TikTok as “cringey,” usually a kiss of death for networks looking to expand among youths, fickle as they are about what’s cool. Scrolling the app at the time felt like eavesdropping on the theater kids clique from high school. Entertaining, but hardly a mainstream entertainment staple.

That’s where the one-two combination of Bytedance’s enormous marketing spend and the power of TikTok’s algorithm came to the rescue. To help a network break out from its early adopter group, you need both to bring lots of new people/subcultures into the app—that’s where the massive marketing spend helps—but also ways to help these disparate groups to 1) find each other quickly and 2) branch off into their own spaces.

More than any other feed algorithm I can recall, Bytedance’s short video algorithm fulfilled these two requirements. It is a rapid, hyper-efficient matchmaker. Merely by watching some videos, and without having to follow or friend anyone, you can quickly train TikTok on what you like. In the two sided entertainment network that is TikTok, the algorithm acts as a rapid, efficient market maker, connecting videos with the audiences they’re destined to delight. The algorithm allows this to happen without an explicit follower graph.

Just as importantly, by personalizing everyone’s FYP feeds, TikTok helped to keep these distinct subcultures, with their different tastes, separated. One person’s cringe is another person’s pleasure, but figuring out which is which is no small feat.

TikTok’s algorithm is the Sorting Hat from the Harry Potter universe. Just as that magical hat sorts students at Hogwarts into the Gryffindor, Hufflepuff, Ravenclaw, and Slytherin houses, TikTok’s algorithm sorts its users into dozens and dozens of subculturesThe Sorting Hat is perhaps the most curious plot device from the Harry Potter universe. Is it a metaphor for genetic determinism? Did Draco have any hope of not being a Slytherin? By sorting Draco into that house, did it shape his destiny? Is the hat a metaphor for the U.S. college admissions system, with all its known biases? Is Harry Potter, sorted into Gryffindor, a legacy admit?. Not two FYP feeds are alike.

For all the naive and idealistic dreams of the so-called “marketplace of ideas,” the first generation of large social networks has proven mostly unprepared and ill-equipped to deal with the resulting culture wars. Until they have some real substantial ideas and incentives to take on the costly task of mediating between strangers who disagree with each other, they’re better off sorting those people apart. The only types of people who enjoy being thrown into a gladiatorial online arena together with those they disagree with seem to be trolls, who benefit asymmetrically from the resultant violence.

Consider Twitter's content moderation problems. How much of that results from throwing liberals and conservatives together in a timeline together? Twitter employees speak often about wanting to improve public discourse, but they’d be much better off (and society, too) keeping the Slytherins and Gryffindors apart until they have some real substantive ideas to solve the problem of low trust conversation.The same can be said of NextDoor and their problem of racist reporting of minorities just walking down the sidewalk. They’d be better off just removing that feature. At some point, NextDoor needs to face the fact that they aren’t going to solve racism. Tweak that feature all you want, put up all the hoops for users to jump through to file such a report, but adverse selection ensures that those most motivated to jump through them are the racist ones.

After some time, new subcultures did indeed emerge on TikTok. No longer was it just teenage girls lip-synching. There are so many subcultures on TikTok I can barely track them because I only ever see a portion of them in my personalized FYP. This broadened TikTok’s appeal and total addressable market. Douyin had followed that path in China, so Bytedance at least had some precedent for committing to such an expensive bet, but I wasn’t certain if it would work in the U.S., a much more competitive media and entertainment market.

Within a larger social network, even subcultures need some minimum viable scale, and though Bytedance paid dearly to fill the top of the funnel, its algorithm eventually helped assemble many subcultures surpassing that minimum viable scale. More notably, it did so with amazing speed.

Think of how most other social networks have scaled. The usual path is organic. Users are encouraged to follow and friend each other to assemble their own graph one connection at a time. The challenge with that is that it’s almost always a really slow build, and you have to provide some reason for people to hang around and build that graph, often encapsulated by the aphorism “come for the tool, stay for the network.” Today, it’s not as easy to build the “tool” part when so much of that landscape has already been mined and when scaled networks have learned to copy any tool achieving any level of traction.In the West, Facebook is the master of the fast follow. They struggle to launch new social graphs of their own invention, but if they spot any competing social network achieve any level of traction, they will lock down and ship a clone with blinding speed. Good artists borrow, great artists steal, the best artists steal the most quickly? Facebook as a competitor reminds me of that class of zombies in movies that stagger around drunk most of the time, but the moment they spot a target, they sprint at it like a pack of cheetahs. The type you see in 28 Days Later and I Am Legend. Terrifying.

Some people still think that a new social network will be built around a new content format, but it’s almost impossible to think of a format that couldn’t be copied in two to three months by a compact Facebook team put in lockdown with catered dinners. Yes, a new content format might create a new proof of work, as I wrote about in Status as a Service, but just as critical is building the right structures to distribute such content to the right audience to close the social feedback loop.

What’s the last new social network to achieve scale in recent years? You probably can’t think of any, and that’s because there really aren’t any. Even Facebook hasn’t been able to launch any really new successful social products, and a lot of that is because they also seem fixated on building these things around some content format gimmick.

Recall the three purposes which I used to distinguish among networks in Status as a Service: social capital (status), entertainment, and utility. In another post soon I promise to explain why I classify networks along these three axes, but for now, just know that while almost all networks serve some mix of the three, most lean heavily towards one of those three purposes.

3-axis.png

A network like Venmo or Uber, for example, is mostly about utility: I need to pay someone money, or I need to travel from here to there. A network like YouTube is more about entertainment. Amuse me. And some networks, what most people refer to when they use the generic term “social network,” are more focused on social capital. Soho House, for example.

TikTok is less a pure social network, the type focused on social capital, than an entertainment network. I don’t socialize with people on TikTok, I barely know any of them. It consists of a network of people connected to each other, but they are connected for a distinct reason, for creators to reach viewers with their short videos.Bytedance hasn't been successful in building out a social network to compete with WeChat, though it's not for lack of trying. I think they have a variety of options for doing so, but as with many companies that didn't begin as social first, it's not in their DNA. Facebook is underrated for its ability to build functional social plumbing at scale, that is a rare design skill. Companies as diverse as Amazon and Netflix have tried building social features and then later abandoned them. I suspect they tried when they didn't have enough users to create breakaway social scale, but it's difficult to imagine them pulling that off without more social DNA. But having a social-first DNA also means that Facebook isn't great at building non-social offerings. Their video or watch tab remains a bizarre and unfocused mess.

One can debate the semantics of what constitutes a social network forever, but what matters here is realizing that another way to describe an entertainment network is as an interest network. TikTok takes content from one group of people and match it to other people who would enjoy that content. It is trying to figure out what hundreds of millions of viewers around the world are interested in. When you frame TikTok's algorithm that way, its enormous unrealized potential snaps into focus.

The idea of using a social graph to build out an interest-based network has always been a sort of approximation, a hack. You follow some people in an app, and it serves you some subset of the content from those people under the assumption that you’ll find much of what they post of interest to you. It worked in college for Facebook because a bunch of hormonal college students are really interested in each other. It worked in Twitter, eventually, though it took a while. Twitter's unidirectional follow graph allowed people to pick and choose who to follow with more flexibility than Facebook's initial bi-directional friend model, but Twitter didn't provide enough feedback mechanisms early on to help train its users on what to tweet. The early days were filled with a lot of status updates of the variety people cite when criticizing social media: "nobody cares what you ate for lunch."I talk about Twitter's slow path to product market fit in Status as a Service

But what if there was a way to build an interest graph for you without you having to follow anyone? What if you could skip the long and painstaking intermediate step of assembling a social graph and just jump directly to the interest graph? And what if that could be done really quickly and cheaply at scale, across millions of users? And what if the algorithm that pulled this off could also adjust to your evolving tastes in near real-time, without you having to actively tune it?

The problem with approximating an interest graph with a social graph is that social graphs have negative network effects that kick in at scale. Take a social network like Twitter: the one-way follow graph structure is well-suited to interest graph construction, but the problem is that you’re rarely interested in everything from any single person you follow. You may enjoy Gruber’s thoughts on Apple but not his Yankees tweets. Or my tweets on tech but not on film. And so on. You can try to use Twitter Lists, or mute or block certain people or topics, but it’s all a big hassle that few have the energy or will to tackle.

Think of what happened to Facebook when it’s users went from having their classmates as friends to hundreds and often thousands of people as friends, including coworkers, parents, and that random person you met at the open bar at a wedding reception and felt obligated to accept a friend request from even though their jokes didn’t seem as funny the next morning in the cold light of sobriety. Some have termed it context collapse, but by any name, it’s an annoyance everyone understands. It manifests itself in the declining visit and posting frequency on Facebook across many cohorts.

Think of Snapchat’s struggles to differentiate between its utility— as a way to communicate among friends—and its entertainment function as a place famous people broadcast content to their fans. In a controversial redesign, Snapchat cleaved the broadcast content from influencers into the righthand Discover tab, leaving your conversations with friends in the left Chat pane. Look, the redesign seemed to say, Kylie Jenner is not your friend.

TikTok doesn’t bump into the negative network effects of using a social graph at scale because it doesn't really have one. It is more of a pure interest graph, one derived from its short video content, and the beauty is its algorithm is so efficient that its interest graph can be assembled without imposing much of a burden on the user at all. It is passive personalization, learning through consumption. Because the videos are so short, the volume of training data a user provides per unit of time is high. Because the videos are entertaining, this training process feels effortless, even enjoyable, for the user.

I like to say that “when you gaze into TikTok, TikTok gazes into you.” Think of all the countless hours product managers, designers and engineers have dedicated to growth-hacking social onboarding—goading people into adding friends and following people, urging them to grant access to their phone contact lists—all in an attempt to carry them past the dead zone to the minimum viable graph size necessary to provide them with a healthy, robust feed. (sidenote: Every social product manager has heard the story of Facebook and Twitter’s keystone metrics for minimum viable friend or follow graph size countless times.) Think of how many damn interest bubble UI’s you’ve had to sit through before you could start using some new social product: what subjects interest you? who are your favorite musicians? what types of movies do you enjoy?The last time I tried to use Twitter’s new user onboarding flow, it recommended I follow, among other accounts, that of Donald Trump. There are countless ways they could onboard people more efficiently to provide them with a great experience immediately, but that is not one of them.

TikTok came along and bypassed all of that. In a two-sided entertainment marketplace, they provide creators on one side with unmatched video creation tools coupled with potential super-scaled distribution, and viewers on the other side with an endless stream of entertainment that gets more personalized with time. In doing so, TikTok, with a product team and infrastructure mostly located in China, came out of left field and became a player in the attention marketplace on the same playing fields around the world as giants like Facebook, Instagram, Snapchat, YouTube, and Netflix. Not quite a Cinderella story...maybe a Mulan story?

TikTok didn't just break out in America. It became unbelievably popular in India and in the Middle East, more countries whose cultures and language were foreign to the Chinese Bytedance product teams. Imagine an algorithm so clever it enables its builders to treat another market and culture as a complete black box. What do people in that country like? No, even better, what does each individual person in each of those foreign countries like? You don't have to figure it out. The algorithm will handle that. The algorithm knows.

I don’t think the Chinese product teams I’ve met in recent years in China are much further ahead than the ones I met in 2011 when it comes to understanding foreign cultures like America. But what the Bytedance algorithm did was it abstracted that problem away.One of the concerns about CCP ties with Bytedance is that they might use it as a propaganda tool against the U.S. I tend to think that problem is overrated because my sense is that many in China still don't understand the nuances of American culture, just as America doesn't understand theirs (though I speak Mandarin, some of the memes on Douyin fly way over my head). However, perhaps an algorithm that abstracts culture into a series of stimuli responses makes it more dangerous?

Now imagine that level of hyper efficient interest matching applied to other opportunities and markets. Personalized TV of the future? Check. Education? I already find a lot of education videos in my TikTok feed, on everything from cooking to magic to iPhone hacks. Scale that up and Alex and Louis might finally realize their dream of a short video education app that they set out to build before Musical.ly.

Shopping? A slam dunk, Douyin and Toutiao already enable a ton of commerce in China. Job marketplace? A bit of a stretch, but not impossible. If Microsoft buys TikTok, I’d certainly give the TikTok team a crack at improving my LinkedIn feed, which, to be clear, is horrifying. What about personalized reading, from books to newsletters to blogs? Music? Podcasts? Yes, yes, yes please. Dating? The world could absolutely use an alternative to the high GINI co-efficient, high inequality dating marketplace that is Tinder.

Douyin already visualizes much of this future for us with its much broader diversity of videos and revenue models. In China, video e-commerce is light years ahead of where it is in the U.S. (for a variety of reasons, but none that aren’t surmountable; a topic, again, for another piece). Whereas TikTok can still feel, to me, like a pure entertainment time-killer, Douyin, which I track on a separate phone I keep just to run Chinese apps for research purposes, feels like much more than that. It feels like a realization of short video as a broad use case platform.

There’s a reason that many people in the U.S. today describe social media as work. And why many, like me, have come to find TikTok a much more fun app to spend time in. Apps like Facebook, Instagram, and Twitter are built on social graphs, and as such, they amplify the scale, ubiquity, and reach of our performative social burden. They struggle to separate their social functions from their entertainment and utility functions, injecting an aspect of social artifice where it never used to exist.

Facebook has struggled with its transition to utility, which would’ve offered it a path towards becoming more of a societal operating system the way WeChat is in China. To be fair, the competition for many of those functions is much stiffer in the U.S. In payments, for example, Facebook must compete with credit cards, which work fine and which most people default to in the U.S., whereas in China AliPay and WeChat Pay were competing with a cash-dominant culture. Still, in the U.S., Facebook has yet to make any real inroads in significant utility use cases like commerce.I speak so often about how much video as a medium is underrated by tech elites. In an alternate history of Facebook, they would've made a harder shift to becoming a video-only app, moving up the ladder from text to photos to videos, and maybe they would've become TikTok before TikTok. If they had, I think their time spent figures would be even higher today. For as quickly as Facebook moved to disrupt itself in the past, there's a limit to how far they're willing to go. I plan to compare the Chinese and U.S. tech ecosystems in a future post, and one of the broadest and most important takeaways is that China leapfrogged the U.S. in the shift to video, among many other things. This doesn't mean the U.S. won't then leapfrog China the next time around, but for now, the U.S. is the trailing frog in several categories.

Instagram is some strange hybrid mix of social and interest graph, and now it’s also a jumble of formats, with a Stories feed relegated to a top bar in the app while the more stagnant and less active original feed continues to run vertically as the default. Messaging is pushed to a separate pane and also served by a separate app. Longer form videos bounce you to Instagram TV, which is just an app for videos that exceed some time limit, I guess? And soon, perhaps commerce will be jammed in somehow? Meanwhile, they have a Discover tab, or whatever it is called, which seems like it could be the default tab if they wanted to take a more interest-based approach like TikTok. But they seem to have punted on making any hard decisions for so long now that the app is just a Frankenstein of feeds and formats and functions spread across a somewhat confused constellation of apps.

Twitter has never seemed to know what it is. Ask ten different Twitter employees, you’ll hear ten different answers. Perhaps that’s why the dominant product philosophy of the company seems to be a sort of constant paralysis broken up by the occasional crisis mitigation. One reason I’ve long wished Twitter had just become a open protocol and let the developer community go to town is that Twitter moves. At. A. Snail's. Pace.

The shame of it is that Twitter had a head start on an interest graph, largely through the work of its users, who gave signal on what they cared about through the graphs they assembled. That could have been a foundation to all sorts of new markets for them. They could’ve even been an interest-based social network, but instead users have mostly extracted that value themselves by pinging each other through the woefully neglected DM product.Of course, Twitter also once purchased Vine and then let it wither on the, uh vine. Of all the tech companies that could purchase TikTok, maybe Twitter is the one that least deserves it. At a minimum, they should be required to submit a book report showing they understand what it is they're buying.

A few other tech companies are worth mentioning here. YouTube is a massive video network, but honestly they may have shipped even less than Twitter over the years. That they don’t have any video creation tools of note (do they have any?!) and allowed TikTok to come in and steal the short video space is both shocking and not.

Amazon launched a short video commerce app some time ago. It came and went so quickly I didn’t even have time to try it. Though Amazon is good at many things, they just don’t have the DNA to build something like TikTok. That they have failed to realize the short video commerce vision that China led the way on is a shocking miss on their part.

Apple owns the actual camera that so many of these videos are shot on, but they've never understood social.iMessages could be a social networking colossus if Apple had the social DNA, but every day other messaging apps pull further away in functionality and design. But I guess they're finally adding threading in iMessages with the next iOS release? Haha. At least they'll continue to improve the camera hardware with every successive iPhone release.

None of this is to say TikTok is anywhere near the market value of any of these aforementioned American tech giants. If you still think of it as a novelty meme short video app, you're not far from the truth.Are there flaws with TikTok? Of course. It’s far from perfect. The algorithm can be too clingy. Sometimes I like one video from some meme and the next day TikTok serves me too many follow-up videos from the same meme. But the great thing about a hyper-responsive algorithm is that you can tune it quickly, almost like priming GPT-3 to get the results you want. Often all it takes to inject some new subculture into your TikTok feed is to find some video from it (you can easily find them on YouTube or via friends whose feeds are different from your own) and like it. Another problem for TikTok is that a lot of other use cases are being jammed into what was designed to be a portrait mode lip synch video app. Vertical video is good for the human figure, for dance and makeup videos, but not ideal for other types of communication and storytelling (I still hate when basketball and football highlight clips can’t show more of the horizontal playing field, and that goes for both IG and TikTok; in many highlights of Steph Curry hitting a long 3 you can’t see him, or the basket, only one of the two, lol). Stepping up a level, the list of opportunities Bytedance and TikTok have yet to capitalize on in the U.S. is long, and it wouldn’t surprise me if they miss many of them even if they stave off a ban from the U.S. government. Much of it would require new form factors, and it’s unclear how strong the TikTok product team would be, especially if divested out of Bytedance. Under Microsoft, a company with a fairly shaky history in the consumer market, it's unclear that their full potential would be realized.

Still, none of that product work is rocket science. Much of it seems clear in my head. More importantly, TikTok, if armed with the Bytedance algorithm as part of a divestment, has a generalized interest-matching algorithm that can allow it to tackle U.S. tech giants not head on but from an oblique angle. To see it as merely a novelty meme video app for kids is to miss what its much greater disruptive potential. That an app launched out of China could come to the U.S. and sprint into cultural relevance in this attention marketplace should be a wake-up call to complacent U.S. tech companies. Given how many of those companies rely on intuiting user interests to sell them things or to show them ads, a company like TikTok which found a shortcut to assembling such an interest graph should raise all sorts of alarm bells.

It surprises me that more U.S. tech companies aren’t taking a harder run at trying to acquire TikTok if the rumored CFIUS hammer stops short of an outright ban. I can’t think of any of them that shouldn’t be bidding for what is a once-in-a-generation forced fire sale asset. I’ve seen prices like $30B tossed around online. If that’s true, it’s an absolute bargain. I’d easily pay twice that without a second thought.

I could cycle through my long list of nits, but ultimately they are all easily solvable with the right product vision and execution. TikTok has figured out the hardest piece, the algorithm. With it, a massive team made up mostly by people who’ve never left China, and many who never will, grabbed massive marketshare in cultures and markets they’d never experienced firsthand. To a cultural determinist like myself, that feels like black magic.


On that same trip to China in 2018 when I visited Bytedance, an ex-colleague of mine from Hulu organized a visit for me to Newsdog. It was a news app for the Indian market built by a startup headquartered in Beijing. As I exited the elevator into their lobby, I was greeted by a giant mural of Jeff Bezos’ famous saying “It’s Always Day One” on the opposite wall.

A friend of a friend was the CEO there, and he sat me down in a conference room to walk me through their app. They had raised $50M from Tencent just a few months earlier that year, and they were the number one news app in India at the time.

He opened the app on his phone and handed it to me. Similar to Toutiao in China, there were different topic areas in a scrollbar across the top, with a vertical feed of stories beneath each. All of these were stories selected algorithmically, as is the style of Toutiao and so many apps in China.

I looked through the stories, all in Hindi (and yes, one feed that contained the thirst trap photos of attractive Indian girls in rather suggestive outfits standing under things like waterfalls; some parts of culture are universal). Then I looked up from the app and through the glass walls of the conference room at an office filled with about 40 Chinese engineers, mostly male, tapping away on their computers. Then I looked back down at page after page of Hindi stories in the app.

“Wait,” I asked. “Do you have people in this office or at the company who know how to read Hindi?”

He looked at me with a smile.

“No,” he said. “None of us can read any of it.”


NEXT POST: Part II of my thoughts on TikTok, on how the app design is informed by its algorithm and vice versa in a virtuous circle.

Beware the lessons of growing up Galapagos

In All the old rules about movie stardom are broken, part of Slate's 2017 Movie Club year end review, Amy Nicholson writes:

Lugging my $10 masterpiece back to the hotel, I thought about how most of the famous faces who represent the movies have been dead for 50 years. Marilyn’s smile sells shot glasses, clocks, calendars, posters, and shirts in stores from Sunset Boulevard to Buenos Aires, Tijuana to Taiwan. What modern actor could earn a seat at her table? The biggest stars of my lifetime—Julia Roberts, Brad Pitt, Nicolas Cage, Sandra Bullock—never graduated past magazine covers to souvenir magnets.

If Hollywood played by its old rules, I, Tonya’s Margot Robbie and Call Me by Your Name’s Armie Hammer should be huge stars. They’re funny, smart, self-aware, charismatic, and freakishly attractive. Yet, they feel like underdogs, and I’m trying to figure out why. Robbie has made intelligent choices. Her scene-stealing introduction as Leonardo DiCaprio’s trophy wife in Wolf of Wall Street. Her classic romantic caper with Will Smith in the underseen trifle, Focus. She even survived Suicide Squad with her dignity intact. In I, Tonya, she can’t outskate being miscast as Tonya Harding, but bless her heart for trying. As for Hammer, Kameron, your review of Call Me by Your Name called him, “royally handsome,” which seems right. He’s as ridiculously perfect as a cartoon prince, and I loved how Luca Guadagnino made a joke of how outlandish the 6-foot-5 blond looks in the Italian countryside. Whether he’s unfurling himself from a tiny Fiat or stopping conversation with his gangly dance moves, he can’t blend in—and good on him and Guadagnino for embracing it.
 

But even if Robbie and Hammer each claim an Oscar nomination this year, I suspect they’ll stay stalled out in this strange time when great actors are simply supporting players in a superhero franchise. I’m fascinated by Robbie and Hammer because they’re like fossils of some alpha carnivore that should have thrived. Does anyone else feel like the tectonic plates under Hollywood have shifted and we’re now staring at the evidence that everything we know is extinct? It’s not just that the old rules have changed—no new rules have replaced them. No one seems to know what works.

Nicholson goes on to cite Will Smith, who once had huge hits seemingly with every movie he made and who is now on a long cold streak.

I'm wary of all conclusions drawn about media in the scarcity age, including the idea that people went to see movies because of movie stars. It's not that Will Smith isn't charismatic. He is. But I suspect Will Smith was in a lot of hits in the age of scarcity in large part because there weren't a lot of other entertainment options vying for people's attention when Independence Day or something of its ilk came out, like clockwork, to launch the summer blockbuster season.

The same goes for the general idea that any one star was ever the chief engine for a film's box office. If the idea that people go see a movie just to see any one star was never actually true, we can stop holding the modern generation of movie stars to an impossible standard.

The same mistake, I think, is being made about declining NFL ratings. Owners blame players kneeling for the national anthem, but here's my theory: in an age of infinite content, NFL games measure up poorly as entertainment, especially for a generation that grew up with smartphones and no cable TV and thus little exposure to American football. If I weren't in two fantasy football leagues with friends and coworkers, I would not have watched a single game this season, and that's a Leftovers-scale flash-forward twist for a kid who once recorded the Superbowl Shuffle to cassette tape off a local radio broadcast just to practice the lyrics.

If you disregard any historical romantic notions and examine the typical NFL football game, it is mostly dead time (if you watch a cut-down version of a game using Sunday Ticket, only about 30 minutes of a 3 to 3.5 hr game involves actual game action), with the majority of plays involving action of only incremental consequence, whose skill and strategy on display are opaque to most viewers and which are explained poorly by a bunch of middle-aged white men who know little about how to sell the romance of the game to a football neophyte. Several times each week, you might see a player hit so hard that they lie on the ground motionless, or with their hands quivering, foreshadowing a lifetime of pain, memory loss, and depression brought on by irreversible brain damage. If you tried to pitch that show concept just on its structural merits you'd be laughed out of the room in Hollywood.

Cultural products must regenerate themselves for each successive age and generation or risk becoming like opera or the symphony is today. I had season tickets to the LA Phil when I lived in Los Angeles, and I brought a friend to the season opener one year. A reporter actually stopped us as we walked out to interview us about why we were there, so mysterious it was to see two attendees who weren't old enough to have been contemporaries of the composer of the music that night (Mahler).

Yes, football has been around for decades, but most of those were in an age of entertainment scarcity. During that time the NFL capitalized on being the only game in town on Sundays, capturing an audience that passed on the game and its liturgies to their children. Football resembles a religion or any other cultural social network; humans being a tribal creature, we find products that satisfy that need, and what are professional sports leagues but an alliance of clans who band together for the network effects of ritual tribal warfare?

Because of its long incubation in an era of low entertainment competition, the NFL built up massive distribution power and enormous financial coffers. That it is a cultural product transmitted by one generation to the next through multiple channels means it's not entirely fair to analyze it independent of its history; cultural products have some path dependence.

Nevertheless, even if you grant it all its tailwinds, I don't trust a bunch of rich old white male owners who grew up in such favorable monopolistic conditions to both understand and adapt in time to rescue the NFL from continued decline in cultural relevance. They are like tortoises who grew up in the Galapagos Islands, shielded on all sides from predators by the ocean, who one day see the moat dry up, connecting them all of a sudden to other continents where an infinite variety of fast-moving predators dwell. I'm not sure the average NFL owner could unlock an iPhone X, let alone understand the way its product moves through modern cultural highways.

Other major sports leagues are in the same boat though most aren't as oblivious as the NFL. The NBA has an open-minded commissioner in Adam Silver and some younger owners who made their money in technology and at least have one foot in modernity. As a sport, the NBA has some structural advantages over other sports (for example, player faces are visible rather than hidden under helmets), but the league also helps by allowing highlights of games to be clipped and shared on social media and by encouraging its players to cultivate more authentic public personas that act as additional narrative fodder for audiences.

I remember sitting in a meeting with some NFL representatives as they outlined a long list of their restrictions for how their televised games could be remixed and shared by fans on social media. Basically, they wanted almost none of it and would pursue take-downs through all the major social media companies.

Make no mistake, one possible successful strategy in this age of abundant media is to double down on scarcity. It's often the optimal strategy for extracting the maximum revenue from a motivated customer segment. Taylor Swift and other such unicorns can only release their albums on CD for a window to maximize financial return from her superfans before releasing the album on streaming services, straight from the old media windowing playbook.

However, you'd better be damn sure your product is unique and compelling to dial up that tactic because the far greater risk in the age of abundance is that you put up walls around your content and set up a bouncer at the door and no one shows up because there are dozens of free clubs all over town with no cover charge.

Sports have long had one massive advantage in production costs over scripted entertainment like TV and movies, and that is that their narrative engine is a random number generator (RNG). If you want to produce the next hot streaming series, you have to pay millions of dollars to showrunners and writers to generate some narrative.

In sports, the narrative is embedded in the rules of the game. Some players will compete, and someone will win. It's the same script replayed every night, but the RNG produces infinite variations that then spin off infinite variations of the same narratives for why a game turned out one way or the other, just as someone has to make up a story every day to explain why the stock market went up or down. At last check, RNG hadn't found representation with CAA or WME or UTA and thus its services remain free.

Unfortunately for major sports, this advantage is now a weakness as sports narrative is much more brittle than its entertainment counterparts. Narrative is a hedge against disaggregation and unbundling, and that is a critical moat in this age of social media and the internet.

One way to measure entertainment value on this dimension is to ask whether you can read a summary of a narrative and enjoy it almost as much as consuming the original narrative in its native medium. My classic test of this is for movies and TV shows. If you can enjoy a movie just as much by reading the Wikipedia plot summary as by watching it, or if you can enjoy a TV shows almost as much by reading a recap than by bingeing it on your sofa, then it wasn't really that great a movie or TV show to begin with.

Instead of watching the entire last season of Game of Thrones when it returns in 2019, I offer you the alternative of just reading textual recaps to your hearts content online. Is that as enticing an alternative as actually watching all six or seven episodes? You'll ingest all the plot details either way, but for the vast majority of fans this would be a gut-wrenching downgrade.

My other test of narrative value is a variant of the previous compression test. Can you enjoy something just as much by just watching a tiny fraction of the best moments? If so, the narrative is brittle. If you can watch just the last scene of a movie and get most or all the pleasure of watching the whole thing, the narrative didn't earn your company for the journey.

Much more of sports fails this second test than many sports fans realize. I can watch highlights of most games on ESPN or HouseofHighlights on Instagram and extract most of the entertainment marrow and cultural capital of knowing what happened without having to sit through three hours of mostly commercials and dead time. That a game can be unbundled so easily into individual plays and retain most of its value to me might be seen as a good thing in the age of social media, but it's not ideal for the sports leagues if those excerpts are mostly viewed outside paywalls.

This is the bind for major sports leagues. On the one hand, you can try to keep all your content inside the paywall. On the other hand, doing so probably means you continue hemorrhaging  cultural share. This is the eternal dilemma for all media companies in the age of infinite content.

Two nights ago, I watched a clip of multiple angles of Tua Tagovailoa ripping a laser beam of a pass to win the National Championship for Alabama. I didn't watch it live, or on ESPN. I watched it on HouseofHighlights on Instagram, where, instead of hearing some anchor on Sportscenter basically tell me what I can see with my own eyes, the video spins around after a moment to reveal the stunned face of the fan who just witnessed the pass live, reaction videos being a new sort of genre which allows a person in the video to act as the emoji reaction caption from within the video itself, speaking a visual language that most young people of this YouTube/Snapchat generation are already familiar with but which traditional media doesn't notice, let alone grok.

This disaggregation problem extends to ESPN, currently still the 400 pound gorilla in the sports media jungle (reminder, there are no 800 lb gorillas). The network suspended Jemele Hill for tweeting something negative about Trump, using the same playbook as the NFL, who threatened players with suspension for kneeling for the national anthem. Both believed these actions on the part of their talent were harming the value of their product.

The irony is that if both ESPN and the NFL had let these things play out naturally, I suspect at worst it would have been neutral, and at best it might have increased their ratings. For the NFL, the ties to modern movements for social justice might have kept the league and its games in the national conversation and made it tangentially relevant to the next generation. The most culturally relevant bit of Sportscenter today may just be the Sportscenter Top 10, as athletes who make a stunning play routinely tell reporters they are excited to see if they'll be featured on that evening's roundup of the top 10 plays.

Unfortunately, many athletes already see an appearance in HouseofHighlights as the social media alternative to appearing in the Sportscenter Top 10. If you follow top athletes on Instagram, you can see which of them favorite posts on HouseofHighlights. Lebron James routinely favorites posts, as do many other stars. Since many of those athletes follow each other on Instagram, that feature of Instagram produces common knowledge. It's not just that Donovan Mitchell knows that Lebron James favorited a HouseofHighlights clip of him dunking, it's that Mitchell knows that James knows that Mitchell knows and so on.

For ESPN, hewing to the idea that only highlights presented dispassionately or games broadcast respectfully are key to their value is a risky one. Not that they haven't generated a ton of wealth from doing so, and not that TV broadcast rights to major sports aren't still extremely valuable, but those are much more fixed commodities, available to the highest bidder, and ones whose value are close to their peaks, if not past them. This can't be a complete surprise within the four walls of their corporate offices given how much salary and air time they devote to blowhards like Stephen A. Smith and Skip Bayless, but their hesitance to lean into cultivating more original voices will haunt them in the long run. The average caption on an Instagram clip of a major sports league highlight is about twice as likely to be fresh and contextually humorous to a young person than any amount of generic sportscaster hooey spouted on ESPN.

This vulnerability extends to their online presences. I still visit ESPN.com on the web and on my mobile devices to get my sports news roundup each day, but sometime in the past few years, the designs of all these presences shifted dramatically. Gone was a hierarchical layout with different sized headlines and groupings of stories. In its place is a long center gutter of updates from a variety of sports leagues, in modern news feed style.

One can see why they went this way, it made ESPN more current, allowing them to push the latest stories to the top of the page to compete with people getting more current updates from Twitter and other social media sites. For a smartphone, in particular, with its limited screen space, it's not easy to block content into multiple sections on one page.

However, the moment you copy someone else's design, you've shifted the terms of the debate in their favor. In a previous era, ESPN's visually distinct information hierarchy set itself up as the authority on what stories mattered. In the new design, what matters skews towards what's the last thing to happen. It's all flow.

To some extent, in our hyper-personalized world, the era of any media entity deciding what stories matter more than others was always going to decline from what might be seen now as a temporary heyday. I care more about Chicago sports teams and Stanford given my background, so having those elements given more prominence was a notable improvement in the site's newly personalized design. Still, what is lost is that sense of authority, that ESPN sets the terms of the debate. Humans remain a social animal, and we take cues about what matters from our each other, including our media entities. ESPN has ceded more and more of the work of determining our sports Schelling points to other entities.

While this may sound grim, the major sports, their respective leagues, and ESPN all have a fairly solid near term window. For one thing, sports is still the highest volume, highest popularity real-time entertainment. As such, it remains a linchpin of many entertainment packages including cable bundles, and so we'll see various media companies pouring money into it until it can't hold things together anymore. We may even see the prices bid even higher for some time as often happens for assets being milked for their last but fleeting window of cultural scarcity.

A second and less discussed factor is that most young tech CEO's don't know the first thing about sports. They, like a sizable part of Silicon Valley (the group that tweets #sportsball whenever Twitter is inundated with reactions to some notable sports event), grew up with other interests. Without that intuitive sense of sports' place in culture, they aren't as attuned to the opportunities in that category.

This provides the leagues opportunities to swindle the tech companies for a while longer, an example being the rights to stream Thursday Night Football, which a series of tech companies from Yahoo to Twitter to Amazon have (probably) overpaid for the last few seasons. As Patrick Stewart said in L.A. Story, "You think with a statement like this you can have the duck?!" The chef says, "He can have the chicken!" Thursday Night Football is zee chicken of the NFL broadcast portfolio, but the restaurant is still called L'Idiot.

This happened for tech companies when they tried to add film and television to their portfolio, too. They routinely paid fortunes for the rights to back seasons of shows that are no longer relevant anymore. When I was at Hulu, I could only shake my head when I heard the asking price for all the back seasons of Seinfeld. Years later, long after I'd left, Hulu paid multiples of that. The cultural decay curve for content in this age of abundance is accelerating by the day, and there is no equivalent of botox to ward it off.

Given market feedback, however, such temporary arbitrage never lasts long. The days of the NFL strong-arming its partners to overpay for the most meager of rights are coming to an end. The thing about setting up a moat around your content is that the moment your cultural value crosses its peak, the moat becomes a set of prison bars. The flywheel loop can turn just as furiously counter-clockwise as clockwise.

And one of these days, a tech company will look at ESPN's homepage and notice how much it looks like their own. If they just put a bit more structure around it, could they satisfy that sports itch for their captive audience which already check in with them multiple times a day?

It seems implausible today, but look at what happened in film and television. For the longest time, so many tech companies were guilty of exactly what Hollywood accused them of, not understanding how film and television is made and marketed, how that industry creates demand for its product. Like all engineering led-cultures, Silicon Valley suspected Hollywood of not being data-driven enough, and many suspected that upstream process failures were responsible for failed releases. Half a film's budget is spent on prints and marketing? What a waste! (Engineers despise marketing.)

Forget that most of these people in tech had never been on a film set, or sat inside a writer's room, or seen the volumes of market research done before any film's release. It's all just content, let's just crowd source some alternatives. Or, if we produce some premium content, what's needed is earlier crowd-sourced feedback. Hundreds of millions of dollars were wasted before Silicon Valley realized they didn't know what they were doing.

Fortunately, all it cost them was some money and some time, something most of the incumbents have a surplus of. Now they write checks to creatives in Hollywood and leave them alone to do what they do very well already. Machine learning improves with data even when the algorithms are off, and so do most tech companies.

I am a lifelong lover of media in all its forms, and sports in particular was central to how I assimilated into America. It has long served as cultural connective tissue between me and friends, family, and strangers. But if I had an easy way to short all the major sports leagues over the next decade, I would. Nostalgia serves many purposes, but its most dangerous one is wrapping us in a memory of a time when we were still relevant.

Optimal pricing for bread and circuses

A survey (pdf) by Anthony Krautmann and David Berri has found that most fans in many popular sports pay less for their tickets than conventional economic theory would predict.
 
Which poses the question: are team owners therefore irrational?
 
Not necessarily. There are (at least?) four justifications for such apparent under-pricing.
 

Lots of things in the real world are underpriced. Most popular concerts and sporting contests lose some volume of revenue to aftermarket transactions on sites like StubHub and SeatGeek. It's nearly impossible to get a reservation at some of the most popular restaurants in San Francisco like State Bird Provisions. There's a waiting list for NOMA Sydney that's 27,000 people long.

If you were pricing to maximize revenue, to match supply and demand exactly, you'd boost prices or perhaps auction off all the seats. What would NOMA Sydney have to charge until its waiting list dropped to zero? I can't even begin toguess, but would it surprise you if it was well north of $2,000 a head for dinner?

Given all of that, I was curious to see what this author thought might explain football ticket underpricing.

The first argument is that underpricing tickets leaves more revenue to be gathered through ancillary sales like souvenirs or overpriced concessions. Without data, I'm skeptical. My instinct is that concession and souvenir sales are less elastic with ticket prices than hypothesized.

The second point is that it's better to have a full stadium for team morale and to influence the officiating. But again, you could sell tickets via a mechanism like a Dutch Auction and maximize revenue while still filling the stadium.

The other two arguments are more convincing.

Thirdly, higher ticket prices can have adverse compositional effects: they might price out younger and poorer fans but replace them with tourists – the sort who buy those half-and-half scarves and should, therefore be shot on sight. This increases uncertainty about longer-term revenues: a potentially life-long loyal young supporter is lost and a more fickle one is gained. It also diminishes home advantage: refs are more likely to give dodgy decisions in front of thousands of screaming Scousers than in front quiet Japanese tourists.
 

I went to a couple games at the old Chicago Stadium, during Jordan's early years with the Bulls, and that place was loud. When they moved to the United Center and the ticket prices went way up, the crowd felt different. More wealthy, and definitely not as loud. It could just be the acoustics of the new space, but anecdotally, I saw fewer fans standing and screaming. Also, the rise of the smartphone means more of the dead moments in a game are filled with people scrolling on their phones, quietly.

Fourthly, high ticket prices can make life harder for owners. They raise fans’ expectations: if you’re spending £50 to see a game you’ll expect better football than if you spend just £10: I suspect that a big reason why Arsene Wenger has been criticised so much in recent years is not so much that Arsenal’s performances have been poor but because high prices have raised expectations. 
 

It's hard to lower prices. Some sports teams may have done it at some point, but I've never seen it. You can raise prices when the team is good and on the rise, but those prices tend to stick when the team declines, and that's when stadiums start to empty out.

Saison is the restaurant in San Francisco that feels closest to pricing to match supply and demand. When I first moved to San Francisco, I had a meal there for $79. The next time there, the meal price had jumped over $100. Then the next time, it was up to $149. Later I heard the tasting menu had risen yet again to $248. The last time I went, thankfully on some banker's expense account, the price was $398 for dinner.

The dining room is usually full, but it's usually possible to get a table the same week. It feels like they've finally reached a price that about as close as you can get to where the supply and demand curves meet. Since the number of seats and turns is limited each night, perhaps this is revenue maximizing pricing, but the margin of error is razor thin.

My guess is that optimal pricing is somewhere below the price that matches supply and demand perfectly. Always being sold out adds a feeling of exclusivity, and no one knows how sold out you are, so being just sold out may be as good from a perception standpoint as being having a massive waiting list.

At the same time, I have a sneaking suspicion continuing to raise the price of a dinner would actually raise demand at some high end restaurants. There may be some Veblen-like qualities to restaurant pricing.

Miss American Dream

Each residency is a reflection of the demographic the property is going for—the Mirage made a play for the affluent and not-quite-debaucherous late 30s/early 40s crowd with Boyz II Men, these boyz who are now patchily gray men, who remain pure in their desire to romance you, to make gauzy, romantic, sweet, consensual love to you, and quickly retreat when you give the nod, who wear sequined letter sweaters and overestimate the impact their music had on our sex lives (“I bet there are some Boyz II Men babies in here!”). The Venetian very much wants the Midwestern, soft-country audience of Tim McGraw and Faith Hill. Planet Hollywood, with its hot pink accents and movie-themed rooms, was built for the Britney fan.

It’s only late in the evenings that Vegas visibly becomes what the tourism board says it is: young and saturated with sex—and not the Boyz II Men-sanctioned lovemaking kind, either. Out on the Strip, aging women wear shirts that say “Girls! Girls! Girls!” A man working for a competing strip club has a shirt that says “Orgasim Clinic: Accepting New Patients.” (Sic on that tragic typo.) Single-named DJs pump their skinny arms as women in tight tube dresses and Lucite heels they bought online a year ago straddle mouth-breathing men on VIP couches like they just heard there was an asteroid headed toward earth or just took a handful of Ecstasy; platonic girlfriends decide to make out at no urging at all because we’re in Vegas bitchez! One does not have to go far to feel the erection of a stranger in the rear of one’s jeans. It is in these small, handsy hours of the night that Caesars’ hope for Britney was born.
 

Taffy Brodesser-Akner on Britney Spears residency at Planet Hollywood in Las Vegas.

I enjoyed this economic breakdown from AEG Live president John Meglen before her residency began. It's telling about the economics of Vegas.

“Even if you believe in Britney, that gives you 50 shows [per year], great, what are you going to put in there your other 200 nights a year?” Meglen told me, in his office in L.A. “If all they have in there is Britney Spears and she is sold out for 50 shows, they have failed. They need Britney Spears and the Spice Girls and Jennifer Lopez and Pink or whoever, okay?” That said, even if the theater is sold out and the seats filled, that doesn’t quite fulfill the residency’s mission, which is to say: Vegas may claim to want youth, but young people aren’t actually good for business.

“You have to ask, ‘Are those kids buying tickets yet?’” Meglen continued. “Because most of them still are seven in a carload driving out from Southern California, they all sleep in one room, they spend the day at the pool and at night they go to the clubs. They’re great at using the workout room, that comes with your ticket. They don’t get the body scrubs or the facial wraps, you know? They don’t gamble and they don’t eat at restaurants and right now, in my opinion, it’s fucking tanking the whole fucking city.”