The Post-YouTube World Will Be Machine-Made
To get it out of the way right off the bat, I think YouTube is ultimately done for, at least in its current form. The massive scale and complexity of the service don’t bode well for it in the long-term. At least Netflix has narrowed its focus to a fairly small number of high-quality productions, and building a reliable distribution service. YouTube is like a blunderbuss; spewing content in every direction in the hope of hitting the mark. It’s worked so far, but it isn’t going to work forever.
So what is going to replace it? Will it be an evolutionary transition like broadcast TV and cable? Or will it be radical transition like cinema and TV? My money is on the former since the distribution platform of the internet is pretty much established at this point. The big developments are over (mobile being the last one) and now it’s all about refining service and evolving existing technologies.
Will YouTube evolve though? It has up until now, but the biggest change the service witnessed was the creation of user channels; one that spurred the multi-channel network (MCN.) Other than that, watching YouTube videos has remained remarkably similar to when it was launched over ten years ago.
What has evolved are viewing habits. Once upon a time they were simple drive-by hits of funny videos. Now they seek professional-level production quality and a connection to the creators. Creating content for YouTube has become big business for everyone involved, with Google quietly skimming their cut from every penny made.
The problem with broadcasting via YouTube is that it’s ineffective. Creators’ work does not get seen unless they really hustle, or jump in with an MCN or other moneyed entity. Yet even this does not guarantee success, or views. Anyone hoping to actually make a go of it on YouTube today is in for a tough time and a lot of hard work if they hope to avoid drowning in an ocean of competition.
The other issue is that we are rapidly approaching what could be termed ‘peak audience’. There is more content uploaded to YouTube ever day than could ever be watched by someone over their entire lifetime. Consumers are drowning in content, and with the explosion in the number of creators, they don’t wont for lack of choice. However, there are still only so many hours in a day, and at some point, consumers will hit the wall of how much time they can devote to viewing content amongst other daily necessities.
This will drive consumers to only focus on, and watch, the best content. Here again, they will not wont for something to watch. In times gone past they subscribed to channels that they liked, or MCNs that produced the kind of stuff they were interested in. We’re already witnessing this become a reality as the latest trend is curators; trusted individuals who pick the best content for you.
Such an arrangement will exist for the foreseeable future, but looking past the curator is something else. Given the massive volumes of data that Google and other tech companies retain on their users, it seems quite likely that content creation and consumption will move beyond mere discovery and recommendations and into more custom-made solutions.
Knowing a person’s specific set of preferences would enable a service (or producer) to profitably create content suited specifically to them instead of trying to appeal to a broad mass of people in the hope of acquiring an audience. Machine learning is already being used by Spotify to build specific playlists for listeners. Motion pictures are a bit more complicated than music, but it’s not out of anyone’s reach.
Google’s algorithms already suggest content based on your viewing history. What if they could also create the content and deliver it to them as well? Contemporary attempts at machine-authored content have been less than stellar, but this is just the beginning, and they will only get better. Given a group of users’ behavioural activity and preferences, it’s entirely possible that a machine could generate content for them and match or exceed humans.
This is not as implausible as you may believe. Any of the latest crop of superhero films has followed a remarkably similar array of tropes designed to elicit the right emotions from the target audience. While no computer has written or produced the films, there are few reasons to believe that given the correct information and programming they could not.
Professionals remain mired in a dedication to the ‘art’ while YouTube ‘stars’ have shown that audiences really don’t care for art at all. The quality of storytelling on the internet is woefully low, and is unlikely to improve unless revenues do. This scenario means that kids are slowly being raised with different opinions of storytelling, and a markedly different set of standards when it comes to judging quality. Much like how MTV gave baby boomers plenty to roll their eyes at, YouTube will similarly make plenty of people decry the tastes of the youth of today and mourn the dying art of yesteryear.
Content will become a much more targeted affair. It won’t be created with a take it or leave it approach, but rather as something that will seem to be exactly what you were looking for. For a bit of rampant speculation, consider your smartwatch monitoring your higher-than-normal pulse, the type of coarse language used in your texts that day, and what you had for lunch (fast food, gleaned from the Android Pay account that you used). Google would know that you had a rough day at the office, and that you would like nothing better than watching a goofy comedy to unwind based on data from previous instances when these data points were similar . You plop down in front of the TV (or computer) and your favourite show is ready and waiting for you, with a brand new episode too!
How customised could content get? We’ve seen target audiences getting ever smaller having dwindled from the tens of millions that the American broadcast networks used to enjoy, all the way down to less than a million for some of the more lucrative YouTube channels. There’s a real possibility that content could survive on viewer rates in the low hundred’s of thousands. Of course, there is the possibility that the same content could be shot and edited in ways that appeal to specific audiences similar to how Netflix creates a number of trailers and aims them at different viewer demographics despite the show itself remaining the same.
Ironically enough, this is an area where animation can succeed over live-action despite the latter being far more prevalent on YouTube today. Animation already relies upon a large amount of technology for its production, and programming it with the various animation techniques would not be as large a leap as for live-action. Character rigs, sets, sounds, music and synthetic voices together with a script can be stored in a database being retrieved and used as necessary. That isn’t to say that live-action could not also be similarly programmed, but rather that animation has the potential to be entirely automated while live-action will always require the human factor; even if they are reduced to being pretty faces.
Whether Google’s YouTube is the service that survives the coming transition is uncertain. I’m not putting my money against them, but Google is not infallible and tech companies have a bad habit of leading rather short lives. Machine learning is a nascent technology that will take a lot of refining before it becomes successful in interpreting the finer points of human desires, but it’s entirely within reach, and will have an enormous impact on popular culture when it does.