Shush Baby, The Algorithms Are Talking

The last 20 years have been a media boomerang

This post was sent to free members of Death By Tech and is available on the web for free.

If you like what you’re reading and want to encourage me to do more, please consider becoming a paying subscriber. You will get at least one more, subscriber only, post per week. I will even give you 50% off the first year because I’m cool like that. Click below!


Back in the early aughts consumer internet became fast enough that it became feasible to download large files. In those days, the biggest files on the internet were usually audio and video files. Streaming wasn’t a thing yet, so this milestone ushered in a massive change in how we consumed content, and it freed us from the 250 channels of crap that the major media houses were jamming down our throats. It also allowed small voices to speak, neutering a powerful lobby of a corrupt mainstream media which was partially funded by special interest groups. It finally put the power back in our hands to consume the content we wanted, and that power largely came from a little thing called podcasts.

The early days of podcasting were filled with names nobody had heard of before. The podcast ecosystem was complex, it was not as simple as just recording an audio file and putting it on the web. In order for the “cast” part of “podcast” to work, creators had to mimic the broadcast feel of the media we were used to consuming. In effect, pushing these audio files to listeners automatically. To do that, the Really Simple Syndication (RSS) specification had to be updated to include the ability to deliver media files instead of just plain text, and a class of applications had to be created which were dubbed “podcatchers”. Despite the fact that a lot of time has been spent on cleverising the “cast” and “catch” part of the ecosystem, nobody has satisfactorily explained the “pod” part. But I digress.

Now that these technologies aligned, we found ourselves in a world where any person can create a show, and anyone can listen to it. We went from 250 channels of crap to an infinite and ever-changing number of podcasts of crap. Podcast proponents heralded this as a win because it removed the ability of mainstream media to control what we listened to. It put the power back into the hands of the individual listeners to critically examine and choose the content they want to consume. And, to some extent, the podcast crowd was right. But, as the saying goes, be careful what you wish for.

Humans are not designed to sift through mountainous piles of information searching for interesting bits of content. The media houses of the 20th century knew this and became the de facto gatekeepers of knowledge. With that responsibility comes some bias and many media houses emerged as partisan organizations with political biases. Podcasting solved this issue in theory because it gave everyone an equal voice. But, overwhelmed with choices, listeners eventually devolved back into accepting gatekeepers.

The early days of podcasting were similar to the dot com bubble.

The podcast gatekeepers took the mantle of “podcast networks” and promised viewers that their network would provide the best quality content so listeners did not have to be burdened with intermittent publishing schedules and poor audio quality. There were hundreds, perhaps thousands, of podcast networks in the aughts, and most of them are gone these days, replaced by existing media houses that have the cash to push out their regular programming in podcast format to be consumed upon demand. While there are indie podcasts still, and podcasting has definitely made some stars, it has utterly failed to live up to the promise of democratizing the media.

The reason podcasting has essentially failed to change the landscape is because the landscape shifted underneath it. Podcasting sought to destroy an enemy that no longer exists. The mainstream media houses that podcasting took aim at are gone or withered and powerless. Google and Facebook have destroyed that industry and in doing so created a vacuum that sucked podcasting in from the fringes and now the internet giants are the new media gatekeepers.

Podcasters hoping to get any significant audience must post their shows on social media sites. Almost nobody uses podcatchers anymore (had you even heard the term before reading this article?), they just go to their favourite social media site to download or stream new episodes. This puts those social media sites squarely into the role of gatekeepers. Which is almost universally understood to be worse than what we started with.

Social media is a fundamentally misaligned business model

Mike Elgin said something on This Week in Tech recently. I am liberally paraphrasing, but it was something like: internet companies get in trouble when they’re misaligned with their users. For example, Google doesn’t get in trouble with search because Google’s goal of presenting relevant search results aligns perfectly with the human sitting at their computer typing search terms into Google. But Facebook gets in trouble all the time because it’s trying to sell ads and sell personal data to companies, but none of its users are there for that. Users are there to chat with their friends, not have their personal data collected and sold. That’s a fundamental misaligned business model and that’s where problems crop up.

He’s right - the social media sites are fundamentally misaligned with all their users. There is no Twitter or Facebook user that uses those sites in the hopes of having their personal data harvested and sold. But doing so is the lifeblood of those sites which means social media is a fundamentally flawed business model and that’s why those guys are always in crap, being fined, being called to testify in front of Congress, and are generally hated by all humanity.

Picking up a magazine doesn’t exploit people

Contrast that with the goals of the previous regime, mainstream media. That industry is also fundamentally misaligned, but it is also almost completely powerless to harvest personal data from its users. How much information can a newspaper or magazine publishing house possibly glean from a subscription? A name, an address, possibly a few more little bits, but that is about it. The user base of yesteryear was almost entirely guessed at; we think middle-aged white men like to fly fish, therefore we will try to sell ads in Field & Stream to companies that have products they want to sell to middle-aged white men. But they don’t actually know that because they have very little idea about who actually buys the magazine. Factor in the non-subscription purchases from supermarket aisles, and the multiple-reader subscriptions such as magazines in doctor waiting rooms, and you’ve got an incoherent mess that makes it impossible to identify the reader base. But, going back to my second sentence, it’s not a big deal because nobody is being exploited by having their user data stolen by picking up a magazine. That was an acceptable trade-off; magazines put ads in front of us, and we were OK with that because we were not disadvantaged in any way by that ad. Today, that is not the case.

Today, the ads you see in front of you on the social media sites aren’t shots in the dark. They’re carefully crafted for you because the social media sites know who you are and they’ve built very sophisticated algorithms to exploit that. The algorithms know you’re a 30-year old female making $40K a year living in Manitoba with a dog you inherited from your last breakup and a car wash job you hate but you have to keep while you finish your degree in modern history online at Athabasca Univesity even though you’re hopelessly behind because you’ve been binge-watching Letterkenny instead of doing your homework.

Facebook doesn’t care what you want. It has its own wants.

Further, it’s not just about the ads. It’s about the actual content that is put in front of you. I think we all know this by now but try it. If you and your friend are both on Facebook sitting beside each other, you will have different content shown on your timelines (notwithstanding your different friend lists). The content these sites chose to show you isn’t like a newspaper where everyone reads the same thing. If your friend belongs to a conservative Facebook group, they’re going to see a right-leaning article about the economy. You, a sane person, are not conservative and you’ll see articles about the economy as well, but they’ll have a left-leaning tone. This is another “feature” of the fundamental misalignment of social media. You’re not there in the hopes of being put into an echo chamber that supports all your views and thoughts (I hope). You’re there to read the “news” but Facebook isn’t interested in your wants. Facebook wants you to stay on the site for as long as possible and interact as much as possible so it can continue to harvest ever-more minute data about you to sell. To do that, Facebook has to create a place you like, a place you want to stay, a place you feel safe. The “news” is never universally loved by all, but news slanted to your particular viewpoint? Oh yes, everyone loves that.

In the first few decades of this century we’ve gone from the dawn of a new democratized media promise to trading mountains of personal data to internet corporations and in exchange we are right back where we were last century: a world with a small number of gatekeepers that control everything we see and read. But this time around it is more insidious because there isn’t even any pretense that social media is trying to inform us, it’s an all-out adversarial relationship between us and the algorithm. We say “give us the facts!” and the algorithm pets our heads and replies “shush…you’ll like this better. Trust me.”