Like a hearty stew on a brisk autumn day in the country, I have prepared a hot take for you: all online discourse is algorithmically generated. I don’t mean that it’s amplified or created by bots, but literally that the trending discourse of any given day on social media is chosen essentially by quirks in an algorithmic model that shines a spotlight at random on people who may or may not be prepared for it.
You can game the algorithm, sure. There are ways to avoid being the Twitter Main Character, and ways to appease the YouTube recommendations algorithm. After all, this is essentially how Mr. Beast has operated for years, his success is due in large part to his obsession with learning how to work the algorithm and staying at the vanguard of viral trends. But at the end of the day, a lot of what we talk about on social media, we talk about because a recommendations algorithm showed it to us. I, for instance, started watching niche electronics repair videos. YouTube noticed I might be crafty and suddenly my feed was full of big name clickbaity makers all trying the newest trends. My special interest was a tributary that led me into the main stream, whether I liked it or not.
Every video has an intended audience. And the nature of the algorithm is that it is trying to match you with videos that are intended for people like you. This matching process means that successful content must be specialized but ubiquitous. This turns makeup channels into drama channels, or art channels into lifehack debunking channels. And as the framing for these videos (clickbaity thumbnails, exaggerated titles, or even trend-hopping premises) strip the content of the context of their community in order to be more attractive to a recommendation algorithm, the conversations around the content are stripped of context as well.
And because this isn’t all meta enough, as the discourse grows beyond the group originally recommended the content through people with larger audiences giving their opinions on the matter, others are pressured to weigh in in order to feel like they belong, even if they don’t know the context of the discourse.
This is how social media works. This is the doomscrolling cycle. This is the ouroboros, the snake endlessly eating its own tail as the content becomes the discourse becomes the discourse about the discourse, becomes the niche blog post about the discourse about the content. That cycle is what fuels the addictive nature of social media; the dopamine drip of Being On The Right Side of an issue, no matter how inconsequential, the feeling of being Part Of The Moment, and the crushing fear of missing out on a communal event.
If you’ve subscribed to a cooking youtuber or two, ANTI-CHEF has probably come across your recommendations page recently. The YouTube algorithm has been pushing him as of late — his “Cooking Julia Child’s Cassoulet Should Be An Olympic Sport” video made its way to YouTube’s highly coveted Recommended Videos carousel for me, my girlfriend (who is co-writing this piece with me), several people I follow on Twitter, and presumably many other folks who follow channels like Cooking With Dog, Alton Brown, and Binging with Babish. [Editor’s note: This video indeed also snuck up on me after watching a bunch of Joshua Weissman videos in a row.]
Jamie Tracey, the titular ANTI-CHEF of the channel, has built his brand on showing how messy, confusing, frustrating, and rewarding cooking can be. His videos track the process of cooking a dish from start to finish, from the gruesome process of killing and cleaning a lobster for bisque, to meticulously weaving a pattern out of puff pastry. The videos are chaotic, almost by design, as if to acknowledge the low-grade stress of chopping vegetables while you’re checking to be sure the steak you’re reverse-searing isn’t overcooking.
Tracey has skills in the kitchen, but that’s immaterial to the point he’s making with his videos: that anyone can cook if they’re willing to read recipes carefully, not take shortcuts, and put the work in. For this reason, the videos on the channel always include moments of relatable cooking panic: adding all the vegetables to the pan at once instead of one at a time, realizing you bought the wrong kind of breadcrumbs, having to stop what you’re doing to do quick math because the recipe is for a 5-pound roast and all Whole Foods had was an 8-pound roast and now you have to figure out how long to cook it. Anybody who’s spent hours in the kitchen making a particularly labor-intensive recipe has had hundreds, thousands of these small moments, and seeing them displayed and documented is important, especially when very few other channels or cooking shows recognize this truth of cooking; it’s both stressful and joyful at once.
The YouTube thumbnail for Tracey’s Cassoulet video shows him, head bowed, grabbing the bridge of his nose. He has the look of a man who has been utterly defeated. An exuberant black-and-white image of Julia Child sits to his right, she wears an irrepressible smile, as if reveling in Tracey’s failure to cook her recipe. This style of thumbnail sends a clear message: this is not just a cooking video, it’s a challenge video too.
It’s a thumbnail tailor-made to appease an algorithm. You want to watch the video. You want to see how this insidious dish from the French countryside could so thoroughly defeat a person. You want to pause the video, yelling at the screen telling Tracey he’s doing everything wrong. You want to get to the end of the video and be assured in your belief that actually, you could do that a lot better.
We don’t know how the algorithm works, exactly, but this thumbnail was a perfect offering to the mysterious recommendation gods, and so the video started being recommended heavily. And because it got picked up by the YouTube algorithm, people started to respond and talk about it on Twitter. And then, people who hadn’t seen the original video started to weigh in. Sound familiar? It should be; if it isn’t, reread the first half of the piece ok thanks.
Now, this wasn’t going to be in anyone’s Twitter Moments. ANTI-CHEF is a big channel with around 140,000 subscribers, but he’s certainly not one of the platform’s biggest. The discourse came and went without much fanfare. But it serves as a perfect primer on how important media literacy is in an algorithmically-driven world.
The response I encountered to the video began with a hilarious (if potentially ableist (shaming someone for complaining about having to spend 11 hours on their feet doing active cooking isn’t a good look)) tweet from lauren, (a must-follow on Twitter) who actually watched the video and is a good cook, at least judging from her posts. Then, the replies started coming in, each less context-driven and more unhinged than the last. The worst of them went as far as to say that if this is his reaction to making a complex recipe, he doesn’t belong in the kitchen — which begs the question if these folks have ever woken up early on the last Thursday of November to solo prep a holiday dinner for a dozen people. It all essentially boils down to telling Jamie to quit complaining, and that he shouldn’t be bashing Julia Child’s recipe when he’s the one who’s at fault.
But if you watch the video, you’ll notice something interesting: Jamie never bashes Julia Child’s recipe, and doesn’t even complain much for someone who painstakingly picked pork skin out of beans, even though Julia Child assured him that the skin would “disappear” as it cooked. All of his wounds are self-inflicted, and he says as much! Jamie’s video was shot in one day, over the course of 11 grueling hours of active cooking. He admits at the beginning of the recipe that Julia Child has a disclaimer at the top of the recipe, saying that it doesn’t take shortcuts in the pursuit of flavor, and should be cooked over 3-4 days of casual cooking instead of in one marathon session.
Twitter dunks flatten context, and as here, can actively invert context. The point of the video, per ANTI-CHEF’s body of work, is about how yeah, cooking is hard, but you can pretty much cook anything if you’re willing to work at it. And also, listen to the recipe because Julia Child knows better than you. When you get a few degrees of separation away from the actual video, when you only have the thumbnail for context, you can see how easy it is for these kinds of reframings to happen. It’s so common, and it happens in contexts much more harmful than this one. It’s a kind of forced recontextualization, a maybe-intentional-maybe-not twisting of intentions, meant to allow a social media user to read bad-faith intentions when none may exist, specifically so that the user can both have moral superiority and feel as though they’re part of the moment.
The impulse to weigh in is relatable and deeply human, and YouTube comments certainly exploit that fact. It’s certainly one of the top tier uses of Twitter. But being wary of what we are saying about a piece of content vs an assumption based on a headline, thumbnail, title, or snarky tweet is the subtle difference between bandwagon jumping and meaningful consideration. So departing from the now worn out plea to “read past the headline,” I am here to beg you to watch the damn video before you tweet about it.