Wish We Wrote #2: The Flattening of Culture
How chasing the YouTube meta has burnt out its best creators and flattened its cultural footprint
Hi, welcome to Dilemmas of Meaning, a journal at the intersection of philosophy, culture, and technology. This is the another post of ‘Wish We Wrote’ in our Discovery series. Here I’m examining Youtube’s AI tools in light of Ted Chiang’s pronouncement of AI as a tool for management, capitalism, and the status quo.
Will A.I. Become the New McKinsey? - Ted Chiang
In 2022, at the midst of the Amber Heard v Johnny Depp trial, NBC News ran a story: It told of Jacob, 15, a video games YouTuber, who consistently saw videos about the trial in his recommendations. These videos had millions of views from channels with barely any subscribers. Looking at his video game videos getting only a few hundred, he put 2 and 2 together and pivoted. In a week, his new videos about the trial had over 10 million views.
Anybody who has tried their hands at online content creation has had to learn how their specific system works—what is rewarded and what is buried, how should you package your content, thumbnails, captions, should you mouth be closed or open. Those with experience call this 'learning the meta', a term popularised in competitive games where you by figuring out the optimal strategies and way of playing, you gain an advantage over less informed players. It is apt as playing the online content creation game becomes that, a game, and one that is zero-sum.
If you aren't playing the meta, you'll be left behind.
It is within this context that a cottage industry of content creation consultants has arisen. Claiming to have studied the many algorithms that promise fame and threaten obsolescence, they offer insights. Whether these be successful creators sharing their path to success, or analysts. As becoming a YouTuber, for example, becomes a coveted career choice by the young, learning the meta has gone from being an edge to being a necessity. By sheer number, the possibility of being buried is near guaranteed and making it like threading a needle.
If you aren't playing the meta, you'll be left behind.
At their recent Made On YouTube event, YouTube announced a suite of tools, designed to ‘help push the boundaries of creator innovation.’ From generating video and image backgrounds for YouTube Shorts to its own mobile editing app. Some are long overdue, like Aloud, a way to dub videos and increase the global reach for channels. However, others are curious:
Ever get stuck on what videos to make next? AI Insights is designed to help spark your next idea and outline suggestions based on what your audience are already watching on Youtube. In our initial test, more than 70 percent of those surveyed said it's helped them develop and test ideas for videos.
Called AI Insights, YouTube hopes to not reveal the meta, but bring the industry of analysis in-house. Optimistically, it removes the zero-sum nature of the status quo. Making AI Insights accessible to all ought to level the playing field, nobody should be left behind.
Harnessing the potential of AI to offer insights and consult people, makes real the metaphor in Ted Chiang's Article, Will A.I. Become the New McKinsey? He argues that in the same way that consultants like McKinsey & Company work to make possible the wishes of those it consults, so should we see AI. Not as an independent or neutral force but an agent of its employers; therefore, an extension of those in charge:
Social-media companies use machine learning to keep users glued to their feeds. In a similar way, Purdue Pharma used McKinsey to figure out how to “turbocharge” sales of OxyContin during the opioid epidemic. Just as A.I. promises to offer managers a cheap replacement for human workers, so McKinsey and similar firms helped normalize the practice of mass layoffs as a way of increasing stock prices and executive compensation...
Ultimately, Chiang concludes that as it stands, like consulting corporations, AI will aid management rather than workers, supporting the capitalist system where 'the people who have lots of money get to wield power over people who actually work'. The way of being—the status quo—will be maintained under a world of AI consultants, nothing subversive or indeed novel will arise. It is this connection, between AI tools and consulting corporations that spurred these questions: will AI Insights hinder YouTube culture by limiting the types of videos that fit the meta? How do we square YouTube's mission to 'help push the boundaries of creator innovation' and Chiang's prediction that these tools will close the horizon of possible futures?
In this edition of Wish We Wrote, I'm asking whether this is how YouTube becomes boring?
As a YouTuber, you are given a narrow path and told to sprint. It is no surprise that many quit and those that stay end up making the same kind of videos.
Let's take the premise and argue against it first. What if I'm wrong and this is a good thing that will push the innovative spirit of creators? AI Insights is framed by YouTube, as a tool for the individual creator. In its copy describing and selling the product, it defers to the second person. This is not for creators, but for you, dear reader, the individual artisan-like creator. The modern-day job of the YouTuber is to wear many hats: editor, writer, social media manager, business manager. It is common for YouTube channels to be run by teams rather than the cliché of the single creator. There is a concern that the professionalisation and grind of YouTube has transitioned many from creators to managers. You cannot be an artisan creator when the demand of posting has burnt out the platform’s longest tenants. By providing YouTubers with analytically backed ideas—a peek at the meta—and outlining tools, YouTube offers a way to make easier the many hatted job of creating videos. The grind is outsourced, and video ideas can be tested and planned with ease.
Yet, many people who might call themselves creators will readily admit that the idea generation, that which this system focuses, is the easiest part. Ideas overflow. Billy and I, when starting this journal wrote down a list of topics we'd like to explore. Months later and many articles and essays published, that list has grown. It is the execution—the practice—that where the grind lies. Concerning burnout, it is in the feeling of working hard on an idea, one new but with belief and inspiration, planning, editing, recording, and publishing this video before seeing it not even reach your subscribers. Profiling top YouTubers on why they feel burnt-out, former entertainment reporter, Julia Alexander highlighted three reasons:
Constant changes to the platform’s algorithm, unhealthy obsessions with remaining relevant in a rapidly growing field and social media pressures are making it almost impossible for top creators to continue creating at the pace both the platform and audience want — and that can have a detrimental effect on the very ecosystem they belong to.
These findings are corroborated by other investigations: frequency, consistency, and a black-box algorithm with rules they fear falling foul of. Burnout doesn't come from not having enough ideas or taking too long to plan them, it comes from YouTube itself. Success on YouTube turns you into a content factory, forced to churn out the right type of videos in too short a time frame with an optimistic attitude. You become the product, the commodity. Take too long working on a novel video and you'll fall foul of the recommendation algorithm. Make a video that isn't what people are searching for and you'll be buried. As a YouTuber, you are given a narrow path and told to sprint. It is no surprise that many quit and those that stay end up making the same kind of videos. My read of events is this: If you make a lot of low-quality videos but make them current and consistent, you'll probably make it as a Youtuber, but if you take your time and drop a high quality or novel video every month, there is almost no chance you make it.
In arguing against the best-case scenario of AI Insights, you might be wondering why YouTube would be in favour of this. Allow me to make my case:
The birth of platform economies—Uber, Deliveroo, YouTube, Twitch—has created an economy where individual agents operate and earn a living within the auspices of the platform owners. These agents own neither their audience nor the platform's tools. A Twitch streamer moving to a rival platform, or their own streaming set up cannot transfer their subscriber base they might have spent years growing. Many a YouTuber have vocalised the fear that, should YouTube disappear, so too their livelihoods. In addition, any dramatic change to how the platform works, such as algorithm tweaks, has real material consequences. Thus, the emphasis on the meta, thus the quasi-social Darwinist nature of burnout or adapt. A force that allows you to earn a livelihood via working under their rules and guidelines, without the ability to get up and leave with a continuation of service? What is this if not the 21st Century equivalent of the classical factory boss. Many a platform owner, chiefly Uber and oligarchs of the Gig Economy, have spent the GDP of a small nation fighting that fact; that they might be in practice, if not in name, employers of their workers. Might we not say the same for YouTube? With this leading question, we come right back to Chiang's thesis that AI will assist management over workers, YouTube is the management while the creators, the workers. So, what does YouTube as management want? Watch time. (To qualify for the Partner Program and start earning, it is not enough to reach a subscriber count. You must achieve an amount of watch-time within a set time frame). It doesn't matter if the watch time is on something good or bad, one great video or multiple terrible videos. This makes sense, if YouTube makes their money by placing adverts next to or in-between videos, they'd want channels that can guarantee a lot of eyeballs. Not merely views but time spent on the platform. The valuable consumer is one that will spend time watching hours of YouTube and therefore, see a lot of ads. From this perspective, a system that suggests video ideas based on what people already want to see is a way of nudging channels to make videos that will keep people watching. Quality or not.
With AI Insights, everyone knows what that current video idea will be, tilting the calculus further in the direction of cheap, quick, and low quality. YouTube is a place of some of the most creative people in the world, turning all manner of ideas into videos and AI Insights won't stop them. It'll just bury them to such an extent that looking across the virtual cultural landscape, all you will see is a blur of sameness, a flattening of culture. YouTube is a microcosm for a larger problem, as solving the meta moves from an edge to the motivation to posting on entire internet, our entire World Wide Web becomes, well, boring.
While working on this piece, I went looking for other YouTubers who pivoted to the Heard v Depp trial. In truth, I initially set out to write about the turning of a serious case—pertaining to slander, domestic violence, and the elite capture of social justice causes by celebrity–into content. Where onlookers care more about who's winning and losing and less about nuance. That idea can join the everlasting list of ideas at Dilemmas of Meaning HQ, as what I found prompted this pivot you've just finished.
I found a YouTube channel, name withheld, that made their career on compilation of MCU actors being charming or funny or silly. A few of their ~8-minute videos went viral and reached the millions but the majority of them averaged out at a few hundred thousand. Good work if you can get it. During the trial, they dropped a compilation of a cross-examination in the Heard v Depp trial and reached views that were 10 million plus. They pivoted hard. Suddenly, reaching millions of views was no longer a rarity and their monthly viewership shot up. Scanning their subsequent videos is to take a nostalgic stroll of pop cultural obsession of time gone by: Jenna Ortega and Wednesday; Pedro Pascal; Queen Charlotte: A Bridgerton Story. The views never reached the heights of the trial and more than that, they dropped to the low hundreds of thousands and then the tens of thousands. The pace of uploads remained rapid but eventually, they stopped for a month. Trying to recapture their original niche, burnt-out by chasing trends and their diminishing returns, the channel rises after their rest, and in a move that must have taken time and effort, pulls up their computer and uploads a grand and well edited compilation consisting of decades worth of footage.
Four months later, that remains their last video. They played the game and still got left behind.