Since this one runs a bit longer than what usually goes out here, here's a quick recap / TL;DR. Though the whole point of the piece is exactly what we lose by summarizing:
A recurring theme when you listen to creators of every size and subject talk about their own creative process (especially the ones who make content about making content) is the worry about figuring out what the algorithm is rewarding. The topic, the aesthetic (YouTube thumbnails as maybe the best example), songs that are trending (on TikTok especially), hooks that might be a catchphrase or a quick camera move to make sure whatever got made survives those few critical seconds where the viewer decides whether to stick around, and so on.
If pleasing the machine matters that much and is the baseline condition for getting wide distribution, who are we actually serving? Precisely because winning over the algorithm is such a big part of the visibility of so many things, instances of manufacturing consensus through active manipulation of the algorithms while posing as public opinion are more and more frequent:
In other times, this more pernicious use of the tech got weeded out by the platforms themselves. In the early days of SEO, keyword jamming (invisibly stuffing keywords into a page) worked, but with constant updates to the ranking system, Google gradually turned many of those practices into something that hurts rather than helps whoever uses them. Now, when even scams the companies themselves already recognize are booked as revenue lines and therefore get de-prioritized as problems to fix, not to mention the complicity with things even worse, it's hard to believe in big voluntary pro-user changes.
This increasingly coordinated, commercial and staged use, far from the networks' original utopias, is a mercy shot to two convenient lies:
Seen through that lens, they look a lot like ”legacy” media: the algorithm plays the role of the major-label executive who sets the form, tone and language of what reaches the market. The fundamental difference is that when control sat with the labels, there was a whole ecosystem of independents and niche where artists who don't fit the mass-market mold could still exist and sustain themselves without sacrificing their own work. Today's system rewards ”performed spontaneity,” like that video shot while someone makes a coffee or gets ready to go out, or other tactical tricks that get quickly copied (and soon saturate), while whatever isn't made with that dynamic in mind disappears. Wouldn't the truly authentic thing be to not give a damn about the algorithm? That makes the idea of authenticity as a differentiator, in a context where everyone has to pass through this filter, kind of absurd.
These ideas are very rarely challenged by people who work in marketing and communications, partly because there's a whole set of commercial interests in this world wanting to convince us that the media is ”the pulse of culture,” treating relevant topics in tiny contexts as if they were universally Brazilian, generational, or even global. It doesn't help at all that the platforms are direct beneficiaries of our collectively believing they still work today the way they worked 12 years ago, because that keeps the aspiring creators working for free and props up part of the perceived value of social listening. There's nothing wrong with wanting to take part in conversations that already exist, monitoring crisis signals, public complaints and customer requests. The problem is treating as a thermometer of a society what the platforms simply can't represent. ”Media” as a concept plays a key role: nobody treats a home-shopping channel as representative or as a movement, we know it's content with narrow appeal and a clearly commercial purpose. Why is it different with social media? Shouldn't the individual who has ”become media” be seen more as media than as a consumer?
Now for the other part of the problem. Quality journalism, scientific papers, deeper studies with a goal bigger than landing a meeting, and other denser forms of ”content,” all have tolls: tolls of money (the paywalls) and also of time, which, partly through conditioning (short video and carousel versus longread and long video, the medium is the message), create natural barriers for things of objectively higher quality.
Except things start getting more worrying, even a bit dramatic, when it's not just more trivial subjects like a makeup tutorial or a movie review playing this attention-at-any-cost game with the tone and language adjustments the algorithms tacitly ask for.
A great example is how the personality and even the type of discourse of a century-old publication like the Economist is often very different out front (on the platforms) and behind the paywall, where it keeps its usual journalistic rigor, whether or not we agree with their points of view. On Instagram, for instance, the tone is often more ragebait, deliberately controversial or provocative toward specific groups, which brings a wave of hostile comments from those who take the bait and massively amplifies confirmation bias among those who agree, via likes.
It also becomes a kind of ideological Rorschach test for every political belief at once, painting the Economist as communist to fascist depending on who you ask and which post they saw. Except the people who actually read the articles are proportionally few (around 1.29 million paying subscribers versus more than 8 million followers on Instagram, of course there's overlap and this is back-of-the-napkin math, but just to compare orders of magnitude).
That provoking and antagonizing generates engagement and attention is nothing new, in fact there's a study showing that attacking an out-group is a critical driver of engagement on the platforms, at least when the subject is politics. Except when we're talking about building brands and credibility, it gets more complicated. The logic of wanting to maximize visibility to bring in more paying subscribers is legitimate. Now, if we think about equity, brand values and distinctive assets (this last one being high-level journalism with a specific tone of voice), isn't that dissonance of tone a shot in the foot, especially for someone who doesn't know the publication well and never subscribed? Is visibility at any cost (easier to measure) worth the potential hit to brand reputation (more expensive and harder to measure)?
A study showing that engagement is not brand memory already exists! And what if this collective desperation for visibility and engagement carries a huge cost that isn't being measured? What if, like what happened with NPS, once a metric gets tied to targets it stops being a good metric, when we're talking about views and engagement in particular?
Grabbing people's attention has always been part of the game. The problem is when the incentives align so that every publication or personal brand turns into a potential tabloid. Tabloids, at the height of print media, chose the tradeoff of losing credibility to grab more attention at the newsstand. If everyone plays the same game, what happens?
Philosopher Harry Frankfurt separates outright lying from bullsh*t: the liar knows he's lying, while the bullsh*tter is indifferent, as long as it serves his purpose, which is quite a bit more harmful, because a lie still assumes the truth exists and matters.
I'm calling bullsh*tmaxxing the many incentives that exist to echo dumbed-down, distorted and touched-up versions of things that might not even be outright lies, but are clearly being optimized for metrics that hurt us collectively.
Prioritizing wide reach almost always prostitutes the quality of the content because the shared language is more limited, especially on more technical subjects with more established knowledge, which is what opens the door to, for example, a fitness influencer saying that sunbathing naked ”raises testosterone,” forcing even public institutions to weigh in. The mechanics go roughly like this:
Because of this dynamic, we end up at something I'm calling...
The tragedy of the commons is when people, acting rationally in their own self-interest, exploit a shared resource until it runs out, hurting everyone in the process. So every unverified piece of bullsh*t, every sensationalist story, every ”non-story” that we, both people and institutions, carry forward thinking about reach and visibility instead of shared value and factuality, keeps poisoning our collective well.
The practical effect of this is already pretty visible. This year's Edelman Trust Barometer, one of the most important global studies on trust, shows a world getting more insular, more closed to different points of view, and more willing to trust individuals than institutions. Since 2020, the media has systematically been the least credible institution among those measured, in some years tied with government, which this year came in second-to-last.
At the same time, the people who trust influencers would consider trusting a company they don't trust today, if it were vouched for by an influencer in food or lifestyle (62% of the 48% who trust one) or a financial influencer (57% of the 44% who trust one). The creator-economy ecosystem will amplify the 62% and the 57%, except 62% of 48% is less than a third and 57% of 44% is a quarter of the whole. The other inconvenient side of the story is that the differentiator isn't the quality of the content, but the type of relationship (parasocial, manufactured intimacy) that the institution can't create. Then comes a hard question with no answer: does it help creators to have advantages over institutions if the digital environment they necessarily inhabit is getting more inhospitable and saturated? Are we putting up buildings on swamps?
Now, thinking more about the professional or the expert...
If we think about how misaligned the incentives are, we can't treat this as necessarily intentional or as bad faith from the people involved. But we also can't treat everyone as innocent. Can we chalk it all up to the urge to ride the wave and self-promote at any cost and the laziness to check, or is there strategic calculation involved? My argument is that there are three different levels of involvement (and of blame):
One could argue that a specialist who profits from other people's ignorance is as bad as the influencers who profit from their followers' gambling losses, except that this mechanism is only immediately clear to whoever has the familiarity to read what's happening, whether you're a doctor, a nutritionist, an investment analyst or work in market research, partly because in these markets the losses aren't immediately visible to the person being deceived. It's gotten so bad that some economists refer to today as the ”grift economy,” and I myself have written about whether we're collectively getting more dishonest or whether it's the grift being more visible and blatant that gives that impression.
In markets where quality isn't easily recognizable, buyers won't pay more for it. That information asymmetry stops the honest seller from charging more for what's superior and ends up driving quality out of the market. That, very briefly, is George Akerlof's market for lemons theory, which is a huge problem in our world of research and insights. In this deteriorated information context, isn't the market rewarding the opportunists disproportionately? How many services have you hired assuming that visibility is a certificate of quality?
Speaking of which...
Not long ago, the talk going around our world was about ”dopamine sites,” a supposed trend from South Korea where people place orders on sites that are simulations and never actually deliver, so they can feel the pleasure of buying without spending. Shall we do a fact-check?
The phenomenon: a Korean developer decided to build a fake food-ordering app. You go in, build an order, watch the little motorbike move on the map, and nothing arrives. The idea is that the pleasure of buying lives in the anticipation (something obvious to anyone who's window-shopped or planned a trip), so you play around there for two minutes, just without spending. The local press covered it and the piece had three characters: two people who used it and an expert who already ventured a correlation with the spirit of the age. Except the headline was already telling the ”generational phenomenon” story with two users.
How the press vouched for it: from an isolated case reported in a local paper, the thing became a ”global trend” in a few weeks and several big outlets kept reproducing it, including many in Brazil, eventually adding a little extra to the game of telephone. The heavy-hitting argument was that a single app hit a million visits (and most likely appeared because of the buzz around the original article, opportunity marketing in a literal sense). Yes, because it's an ad! And even if others showed up afterward, does that mean the phenomenon materialized, or that there's more opportunity marketing surfing the same wave, even if it's just to make a few bucks off ads?
Isn't that a lot of noise for very little event? Seriously, no expert took a closer look before posting and went straight to the pseudo-heady theorizing like ”(these sites) show that pleasure is no longer only in the acquisition, but in anticipating it”? More importantly: did it move the needle in any market? Will anyone still remember this in six months? Does anyone still remember reborn babies?
Except this story is far from an isolated case. I deliberately chose the -maxxing suffix for this article's title and the name I'm giving to what I'm describing because it is symptomatic of the hyperbole built on top of irrelevant things or old news repackaged, like Nonnamaxxing, recently described by a research institute as ”one of the most relevant trends today,” based on what exactly? More than the cost-of-living crisis, demographic transition, loneliness, mental health and other documentably more relevant things, is that it?
If the subject of the sentence is ”Gen Z,” you can almost automatically assume the piece is talking about a scattered handful of people (because it used views or engagement as a measure of widespread behavior, bullsh*t, right?), making outlandish generalizations about a group of millions of profoundly heterogeneous people, using American data to talk about the whole world, or citing a study whose methodology invariably never shows up in the piece.
NYU professor and cyber-ethnographer Ruby Thelot on made up phenomena. If both the guy who investigates where things on the platforms come from (Keith Presley, from the Vulture piece) and the ethnographer are saying the SAME thing, isn't that a sign it's time to be more skeptical and share less nonsense?
With this flood of bullsh*t and so many people replicating it thinking about their own personal brand, shouldn't that push us in a less exploratory and more fact-checking direction? Everyone wants to do the fun part, which is coming up with theory and doing exploratory research, but verifying and cleaning data, which is just as fundamental, nobody wants to know about? By the very nature of what our work is, shouldn't we be more investigative journalism and less tabloid press?
Now, how many of these ”phenomena” wouldn't disappear immediately if the language and format of social media didn't shape the discourse (but not necessarily what's actually happening) so directly? Aren't the side effects of this ultra-processed information diet already clear enough?
A year and a half ago, I wrote an article about a possible saturation of the attention economy. Since then, things have moved along quite a bit and the signs are out there:
Now, this rock bottom we're at looks a lot like what Cory Doctorow proposed as the final stage in the cycle of ensh*ttification: first the platforms are good for users (chosen content, organic reach, interests prioritized), then they abuse users in favor of B2B customers (cram in more ads and suggested content, open up more user data). The last stage before the end is when value is being sucked out of everyone at the same time (with more expensive ads and a high rate of fraud) and all that's left is a carcass vaguely resembling what the platform once was. Could we argue we're already there?
There are still people arguing that if you speak the platform's language even while talking about serious stuff, you get reach, using the (fantastic) biologist-turned-science-creator Mari Krüger as an example. The thing is you can't treat the exception as the rule. Not coincidentally, in general it's people who want to sell you ”posting like a creator” as a course, mentorship, etc. My reading is the opposite: if you're a specialist in a subject (and not a full-time creator), are you really willing to dance dressed up as a virus and spend a lot of money and time on production to be seen by a large number of people, most of whom aren't the natural audience for your work, and to let exactly those people, mediated by an algorithm that optimizes for the things I already mentioned, steer the agenda and direction of what you produce? Isn't that battle already lost by design? Isn't being forced to turn everything into accessible entertainment or outrage a sign of saturation in itself?
For at least ten years, trend bureaus, strategists and researchers have talked about curation as a (vague) way out of these problems, but it never quite materializes, because the problem is that curation that depends to any degree on algorithmic distribution to grow will invariably be contaminated by those same ills. The clearest example of this is maybe Substack, which showed up selling itself as an alternative to a digital information diet based on carousels and short videos, where begging for likes wouldn't be necessary anymore. They swore they'd never have an algorithmic feed until they changed their minds in 2023.
To understand how and why this decision was made, including with testimony from Substack's own leaders, this article breaks it down in detail. The most curious thing is that it arrives, by a completely different path from mine, at the same conclusion: from the moment an algorithmic feed exists on a platform, it's the thing that really controls the format and the language of what gets produced, with the results I argued in this piece.
Isn't it time we started managing the risks of our bets and investing, even gradually, in other ways of being found and recommended?
