It’s the Commodification of Attention, Stupid

29 minute read

One of my Big Serious Goals for 2023 is to become more socially connected and make more really good friends in my new city, where I hope to be for some time to come, so I’ve been doing a lot of reading about what seems to be making this harder in recent years. This leads naturally to a bunch of related topics, like why our kids and teenagers seem to be more unhappy on average lately, why the Internet and social media are becoming less of a positive in people’s lives, and so on. And I’m becoming convinced that one thing lies behind most of our recent social and informational problems.

“One thing” might be a bit of an exaggeration, because it’s not one specific cause but one idea, taken in a bunch of different directions at once by different people and products. I’m also not claiming that this idea is the entire cause of any issue I discuss, merely that it is a significant and often underappreciated contributing factor. But there’s still this one fundamental idea. That idea is commodification of attention – that is, a system of organizational, psychological, and financial innovations that enables one to make money and/or curry favor by simply convincing people to spend time looking at things, usually virtual ones. Any value provided to the viewers or to society by the things is secondary, and the value can even be negative, so long as the promulgator can (even momentarily) convince people they would feel good to look at. Unfortunately, humans have proven to be quite vulnerable to having their attention manipulated.

I’ll talk about five threads that play into this common theme.

1: The degradation of web search and content

Let’s get the flavor by starting with something less emotionally fraught that most of us have likely noticed in recent years: why are Google search results getting worse?

Kagi, a premium search engine I subscribe to, recently published a blog post called “The Age of PageRank is Over.” In it, they trace the gradual deterioration of search quality to a simple mismatch of incentives (which, as the post points out, anyone could have foreseen; in fact, Google did foresee this, but then changed their mind when they saw no other way to make money than to misalign their incentives). The mismatch is this: Users want to find the highest-quality and best-targeted material as quickly as possible, since that’s what they want out of a web search. Search engines want to make people click on as many ads and sponsored links as possible, since that’s how they make money.

The problem, of course, is that sponsored links and ads are typically less helpful to users than organic results; presumably the advertiser wouldn’t bother paying for a sponsored link if legitimate organic results for their site – those that the search engine predicted would be most useful on its own initiative – reliably appeared higher on the page than their sponsored link. And users are more likely to click on ads if they spend more time on the search engine. One could increase the amount of time people spent on the site by making the search engine so helpful that they searched more, but in the 2020s I think the market for traditional web searches is just about tapped out; people already search for about as many things as would be useful to search for. So the only way to get people to spend more time on the site is by making their search results worse, or at least by optimizing for things other than getting the user their information as quickly as possible. I think it’s less that Google et al. are intentionally making search results bad than that they simply don’t care whether they’re good, as long as their ad metrics keep looking good.

Note: That the market for traditional web searches is tapped out does not mean there’s no room for search innovation. There are kinds of combinatorial searches that you cannot generally run today, like “Give me a list of companies in my local area that are hiring for jobs that match my skills and that employ at least two people I know, along with the names and email addresses of those people,” that would be thoroughly useful but that nobody has yet figured out how to service. Recent AI advances may be bringing some of these within reach.

On the content side of things, in successfully commodifying users’ eyeball time, we have made it possible to earn more money by creating not good content, but simply content that people will look at (along with the ads on the site), regardless of whether it provides any value to them. Because search engines don’t aggressively filter these out (and in some cases perhaps can’t, due to an escalating SEO arms race), people end up looking at them whether they want to or not.

LLMs and other recent generative AI advances have made these clickbait sites even more of a problem by reducing both the barrier to entry as a content farm and the minimum amount of traffic you can receive for a given page while still paying back the production costs. You no longer need a person to write the (bad) content; you can just pick some topic a lot of people are searching for, feed a few lines of badly written description to the model, and copy and paste the output into a web page. (Remember, it’s irrelevant whether the content is any good; it just has to have a form close enough to useful content that it ranks well in the search engine.) Repeat a hundred times a day for $5 an hour. Creating such content provides negative value for society because this content is more or less useless and it buries the more useful content, but the commodification of attention means you can extract money from it, so of course people do.

It’s also worth pointing out that internet advertising is essentially a form of regressive wealth transfer, from less-technical users to more-technical ones. Anyone who knows how to spot the sponsored links in a search-engine results page and install an ad-blocker (not hard, by the way) pays comparatively few of the costs and gets free content, while everyone else spends more time using the web, risks getting malware from compromised ads, and occasionally gets convinced to buy worse products or things they don’t need at all. But even the most technical users can’t do much to fight the profusion of useless content and bad search engines.

Kagi has a painfully simple solution to the search side of the problem: charge users a monthly subscription fee. Kagi doesn’t have to sell out to advertisers to make money, and they realign their own incentives tightly with yours: they want to get you what you want as quickly as possible so you stop running searches, since every search costs compute time and you’re paying a fixed amount per month. I don’t think anyone would have paid a monthly subscription to search the web when web search was a new thing, but Kagi is counting on some reasonable fraction of people caring about this problem enough to subscribe now. It might work; the main challenge is similar to that faced by media organizations and Substacks, where people are happy to subscribe when the idea is new, but hoping average people will spend $300 a month in subscriptions to websites if everyone starts using this model seems like a bad gamble.

Of course, Kagi still has a limited ability to solve the problem of all the low-quality content out there. I do find their search results to be a little better-targeted than most other search engines out there, and you can promote or demote specific websites based on how useful you find them personally, which is a killer feature that makes your search results better over time. But it’s only a partial solution; keeping bad content made by motivated actors from reverse-engineering your search engine’s algorithm to bring themselves to the top is a never-ending battle.

2: The effects of social media

Having trouble searching for information makes life harder. But commodification of attention gets serious when it’s applied in domains that affect people’s personal relationships, so that’s what we’ll be focusing on for the rest of the post.

Unless you’ve been living under a rock for the last fifteen years, you’ll recognize immediately that extant social-media systems reward actions that get attention from other users, both financially and psychologically. They rarely reward any other kind of content; people can choose to behave on them in ways that aren’t seeking attention, but this generally requires a conscious choice and ongoing effort, as social media platforms are designed to steer you towards attention-seeking forms of interaction instead.

Goodhart’s Law strikes again. Just like with web search and content, we developed a whole social system that rewards a specific metric – how many people spend time looking at something and click buttons next to it – and unsurprisingly, people have started to optimize for it at the cost of anything else that matters. Posting content that few people click the “like” button on can make you feel bad and discourage you from posting more, even if it was thoughtful and the people who did look at it enjoyed it; yelling at other people or spreading misinformation can bring you attention and make you feel good, even when doing so is pointless or even harmful to society. In the marketing coup of the century, social media companies have tricked us into optimizing our own personal content for a metric that makes them money by making people stay on their platform and look at their ads, while bringing us no benefit at all (indeed, it often makes our lives worse).

And this optimization spreads into real life. In her recent book Hanging Out: The Radical Power of Killing Time, Sheila Liming explains in depth what happened at her college in 2005 when Facebook arrived (for those who weren’t plugged in to all this at the time, Facebook was initially institution-specific and rolled out to universities one by one, creating a number of fascinating natural experiments). For a while, there were essentially two kinds of parties on campus, which they even called “Facebook parties” and “non-Facebook parties,” depending on how people were invited to them. As Liming tells it, you could tell a Facebook party when you were there, even if you’d been brought along by a friend and didn’t know how the invitations had been dispersed, because everything people did was subtly angled towards performing so they could post pictures and other media on Facebook afterwards, demonstrating that they were cool and had gone to this party. Which, as she points out, was particularly silly because the Facebook parties were by their nature non-exclusive: anyone with a Facebook account who looked for the party was invited, so having been at the party hardly said much about your status.

In more recent news, the dissolution of polite political discourse, at least in the United States, provides a slightly contrasting view. Everyone has the perception that we are polarized, and that everyone who holds different views is crazy and evil. We don’t confine this assessment to Twitter, we also apply it to people we meet. Yet in the now-rare case that you physically sit down with someone you’d expect to think differently from you and have a real conversation, you still tend to find that you agree on more than you’d think, and that most people are fundamentally reasonable. Sure, you’d find some people to have legitimately radical and maybe even evil opinions once you talked to them, but they make up only a fraction of the people you’d think would fall into that category from a brief survey of Twitter or TikTok.

In some ways, though, this is a self-fulfilling prophecy; politics isn’t just about reasoning, it also has a strong social and signaling component, especially in the social-media era. If insane ideas keep spreading around social media and it looks like many other people believe them in real life, eventually people will start to believe them and act on them in real life, even if they don’t now; some people already have (see the January 6 riots). So it’s reasonable to think that not just political discourse but also actual political opinions are being damaged here.

I don’t think any of us legitimately want to optimize our lives around posting cool pictures on Facebook or saying angry things on Twitter (unless, of course, we are, ourselves, in the business of making money by doing so). But it happens anyway, and it wouldn’t have happened without commodification of attention.

3: Smartphone addiction

Smartphones and social media are engaged in a feedback loop. Neither could be nearly as addictive and effective at capturing people’s attention without the other: phones would have few apps which could instantly suck you in and relieve boredom for an effectively unlimited length of time (mobile games being the prime exception), and social media sites would be accessible only from computing devices you didn’t always have in your pocket, meaning you’d only use them when you deliberately decided to do so. (OK, many of us sometimes end up browsing to social-media sites when we’re bored and in front of our computers. But at least we’re only at our computers for part of our lives, and we’re usually not in front of them while trying to explore the world or eat dinner with our families.)

The smartphone, or, more specifically, the abstract worldwide platform that’s created by everyone having a phone in their pocket, has become centered around commodification of attention. Designing a killer app is about figuring out ways to make users interact with it all the time. Your app can do basically anything as long as it accomplishes this, even things that are actively harmful to the user, and you can make money. David Cain, one of my favorite bloggers, recently compared using your phone to smoking a cigarette, and he’s hardly the first person to have made the connection (that last Atlantic article is from 2010!).

The inevitable result, because companies making apps know how to do psychology research, is that we use their app instead of doing valuable things with our lives and connecting with other people. Using their app isn’t satisfying in any lasting way, but it hacks our brains’ reward systems so that it feels like the best thing to do in any given moment. I think few people would debate this anymore – we’ve seen it happen to ourselves or at least to someone we know – yet a solution seems elusive because of the all-encompassing nature of phones. On some days, throwing your phone out the window might seem attractive, but then you wouldn’t be able to, you know, get directions to anywhere, take pictures, wake up on time, log in to your accounts on the web, call an Uber, or talk to your friends. Of course, solutions to these problems exist, but solving them all at once, and dealing with the ongoing effort of keeping them solved as the world continues to go full steam ahead with smartphones, makes the problem feel so complicated that it doesn’t ever seem like a reasonable solution to most people.

N.B.: Apps that are sold once, up front, are usually not commodifying your attention. But most popular apps don’t make most of their money through app store sales anymore, but rather through in-app purchases, advertising, and getting you into other platforms that make money by getting your attention. This might be a useful heuristic for whether an app is likely to mess with your mind; if they charge you $10 and then you have the app for life, the developer’s incentive is to make an app that looks slick and useful on the app store page and draws good reviews, not to make you compulsively launch it fifteen times a day. Looking slick on an app store page is hardly a perfect incentive either, but it’s a lot better than “steal your attention.”

4: The decline of spontaneity

In his Slate article reviewing Hanging Out (mentioned earlier), Dan Kois tells a story of a time he was in college and having a boring afternoon. He wandered over to a friend’s apartment and went in, finding the door unlocked and the TV on. He could hear his friend in the bathroom, so he sat down and started watching whatever was on, and a minute or two later his friend came out, was unsurprised to find Kois there, and sat down to watch TV with him.

This would be completely bizarre today, wouldn’t it? I think I would scream if I walked out of my bathroom and found even my best friend sitting on my couch.

Overall, doing things with other people spontaneously, at least among people my age and in places I’ve lived, has become rare outside of specific communities, usually ones involving physical proximity of living spaces. If you’re leaving home to go shopping some afternoon, you might ask your family if anyone wants to come. Or if you’re living in a college dorm, you might try to knock on a couple of people’s doors and see if anyone wants to go get dinner with you. Maybe you’d text someone who didn’t live with you at the last minute…if they were your romantic partner or your closest friend. With anyone else, this would feel weird. But it used to be normal! You could even call them without asking in advance, or go over to their house and see if they were home!

Here’s my own story: Yesterday after church I caught a friend and asked if she’d like to hang out sometime this week. We ended up deciding that right then would be a great time. As we left together, it struck me that I couldn’t remember the last time I had agreed to spend an afternoon with someone without any advance notice. Like, the last time was probably pre-COVID (about which more in a minute).

I don’t want to make it sound like I disapprove of planning things. Plans make your life a little more predictable, ensure you get around to seeing each other sometime even if you forget in the meantime, and give you something to look forward to. But there’s something dull and even dark about running your whole life on a calendar, and something delightful about randomly asking someone you like to do something with you and finding them as excited about it as you are. And while I’m all for spending time alone (I’m an introvert!), at least some of the times you find yourself with no plans and wanting to do nothing in particular, it would be more fun and better for you to do nothing in particular with someone else. If you have no opportunities for social spontaneity, that’s barely ever going to happen. Instead you’ll quite probably find yourself scrolling through social-media sites, pretending you’re being “social” while sitting on your couch in an empty house.

What happened here? Yet again, I submit that much of it is related to attention.

First, I want to suggest, and this might sound bad at first, that most of your friends are not actually very good at grabbing your attention in comparison to most of the demands of modern life. Think about it: how often, in the course of a typical day, do you find your mind drifting randomly to someone you know and haven’t seen for a while? If you’re planning to do something with them later in the day, or you’re passionately in love with them at the time, sure. Otherwise, it can be pretty easy to forget until you hear from them, or run into something you’d like to tell them or do together. In comparison, how often do you find yourself taking out your phone or switching applications on your computer and doing something mindless there?

Of course, if you know them well at all, your friends can tap into your phone’s notification system to grab your attention. But if everyone’s feeling busy and nobody reaches out first, the days and even weeks can roll by with nobody really noticing. Meanwhile, I bet everyone is spending an hour a day on apps on their phones, no matter how busy they are, maybe even in a consolidated block of time, and even if everyone would rather spend that time with each other. Modern devices designed for the purpose – because lots of people are making big bucks from that design – are just better at getting your attention; friendships have never had to compete with TikTok or Candy Crush before, so we haven’t evolved a way to let them.

Second, I want to hypothesize that our brains treat expressions of social attention as a proxy for connection. (Just to be clear: this is an armchair theory. I think it’s compelling and explains a lot, but it could be totally wrong.) Think of this analogy: despite its constant necessity for continued life, we are not able to sense the level of oxygen in our blood. The sensation of feeling out of breath is due instead to the buildup of carbon dioxide. For whatever reason, this sense apparently proved easier to evolve, and it’s so closely correlated with oxygen levels under normal circumstances that measuring it instead of the actual target metric usually works fine. But there are edge cases where it doesn’t; for instance, if you hyperventilate for a while to clear all of the carbon dioxide, you can hold your breath for a ridiculous length of time, sometimes to the point that you black out from lack of oxygen before feeling like you need to breathe. (Obligatory disclaimer: If you try this at home, please use moderation and look up how to do it safely first! In particular, for reasons which are hopefully obvious, never, ever do this underwater.)

Similarly, in most situations our ancestors experienced, getting people to pay attention to you meant you were physically near them and interacting in meaningful ways that would probably lead to feelings of connection, so we haven’t developed any “no-genuine-connection” alarm bells that go off when we feel unfulfilled. We might get depressed, anxious, or worse, but the causes won’t be traceable unless we really take some time to think about it.

Unfortunately for us, the whole world created by the commodification-of-attention economy is one giant edge case. The ways we communicate through text and social media can easily bring us attention, both from people we know and from people we don’t know. That makes us feel like we’re doing OK on our basic social needs. But underneath, these are impoverished media and rarely actually fulfill them. When was the last time you had a deeply meaningful conversation with someone over text messages, for instance? It does happen, especially in cases where you both have feelings you want to express but can’t communicate any other way at the time, but it’s rare and won’t happen by accident, certainly not in the course of scrolling through a feed. No common platform optimizes for these kinds of interactions, because that doesn’t help with any metrics; genuine human connection, while a key component of a good life, is rarely addictive in the same way that social media and phone games are.

Yet because they make us feel like we’ve done everything we need to, social-media interactions and texts end up substituting for hanging out and spontaneous conversations and adventures. The side effect is that people get worse at normal social interactions and eventually become unhappy, but no platform study is going to measure that; even if Meta et al. seriously cared beyond mitigating the bad PR, all the bad outcomes fall into the metric black hole.

Lastly, most of us feel like we should be productive all the time. All the factors presented above are key drivers of this feeling as well. We need to have a life that looks good on social media and good to our future employers. We can cram something into all the blank spaces in our day, so waiting for something without “doing something” on our phones feels wasteful and even strange; boredom is so rarely experienced that it feels uncomfortable. The lack of boredom and free space to think, ironically, also makes it harder to recognize that we’re unhappy about it, or why we’re unhappy.

I’d like to point out that doing nothing in particular, especially socially, is productive, just in a different way that’s not readily measurable. You need space to be creative. You need to invest in relationships with other people so you can bounce ideas off of them and they can help you when you need it. When you look back on your life, most of your important “productive” outcomes likely wouldn’t have happened without things you did when you weren’t actively working. As a tiny example, take this post: one important part of it took shape in the course of a conversation with my friend yesterday afternoon, while we were sitting on the porch doing nothing besides talking and enjoying the spring sunlight, and several other parts during a walk I took this morning. Without those “unproductive” activities, you wouldn’t be reading this.

5: Suddenly, COVID

No discussion about recent social trends would be complete without mentioning the COVID-19 pandemic. COVID is definitely involved here, but blaming COVID, in itself, for any of these issues is missing the point, because every struggle was already well underway when it started.

Here’s a thing that happens in choral singing sometimes. There’s no accompaniment, and the group starts drifting a little bit off-key, a small enough amount that someone with an untrained ear wouldn’t notice. Gradually, continued error causes it to reach a point a little bit beyond midway between the current key and the next key. From here, the pitch could still be held steady and not get any worse. Or some dedicated and careful singers could even find their way back to the correct pitch.

But then there’s a silent moment, or a key change, or a transfer from one half of a double choir to the other, and the singers, trying to hear their next note in their heads, “correct” the key to the wrong, now-closer, key as they complete the changeover. As soon as the next chord sounds, everything settles into the new key – exactly on pitch, but with everything off half a step from where it’s supposed to be. Returning to the correct pitch is now hopeless; everything has re-centered in an entirely new state, and the only thing to do is to finish in the new key.

This is how I think about the social changes wrought by COVID. A bunch of trends caused by commodification of attention were headed in unpleasant directions, and then a moment of bizarre and unexpected transition suddenly made our attitudes and behaviors fall into a new state that we would have taken years to drift all the way into otherwise, during which time we could have still started moving back the other way with a little bit of resistance. (Videoconferencing is an unobjectionable example; it was already finding increasing use as the technology improved and the need created by continued globalization increased, and presumably it would have eventually become popular no matter what. But the onset of COVID lockdowns caused hundreds of millions, if not billions, of people to adopt it from one week to the next, instead of people adopting it in a slow trickle, as usually happens with new media. I continue to be amazed that the Internet didn’t disintegrate into a puddle of green goo that week.)

Specifically, I think COVID lockdowns and general pandemic-related caution caused the following settling effects, among others:

  • COVID gave us permission to feel socially miserable. In fact, for a brief period that became a virtuous thing to feel. Even now that the times of lockdowns are over almost everywhere, we still feel that people (including us) lost their social skills during COVID and should be cut some slack. While I approve of being nice to people, it’s also easy to use this as an excuse not to act and make yourself start feeling better.

  • COVID normalized bad feelings that many of us had. Maybe the commodification of attention (or other factors, for that matter) had begun to make us unhappy with our social lives. Then COVID happened and we couldn’t make them better for a year or two, so we said “this is fine for now” as a coping strategy. Now we’ve forgotten to pop out of it, and we don’t notice there’s anything we want to change at all.

  • COVID brought a lot of people into remote work for the first time, and massively changed the way work happened for almost everyone – even for people in professions that never worked from home and never will, since the rest of society shifted around them. To look at the experience of those whose workplace changed specifically, remote work certainly doesn’t have to mean you’re lonely or unconnected with your team. But unless you work in a bunker by yourself and never speak to another person, if you go into an office, your work is presumably a significant part of your social life, and suddenly uprooting that component is a system shock that’s almost certainly going to require you to make changes elsewhere to feel the same way you did before. People who did remote work before the pandemic spent years gradually figuring out how to do it right, both in terms of arranging interactions within their companies and in terms of designing the rest of their lives to meet their social needs. Everyone else was suddenly thrown into the deep end in 2020, and with everything going on at the time, most of us never took time to reflect on this at all.

    (There were more direct effects as well. For me, the months of isolation broke up a group of friends I had had lunch with every day at work for three years. I never saw some of them in person again, even once we were back in the office under heavy masking and distancing rules.)

  • COVID pulled us deeper into many of the worst commodification-of-attention platforms and patterns, because they were one of the only forms of social interaction we were allowed. It also caused nerve-driven overconsumption of attention-targeting material like COVID news, creating a habit which has likely persisted in more than a few people.

I’m not saying we’re all now permanently doomed to being lonely TikTok zombies with no friends, but it’s going to take a lot more deliberate work to fix; many, if not most, of us are going to have to explicitly recognize that we need to change and take specific actions. Furthermore, the problem is systemic; you can’t entirely fix your own relationships without cooperation from the other people in them.

What next?

I usually like to end these posts by suggesting some possible solutions, both personal and societal. But I haven’t been turning this one over for long enough to have any sensible solutions, so I’m going to beg off for now. I do promise that I’ll write another post, in at least the next few months, with some suggestions, and when I do, I’ll link it here! In the meantime, if you have suggestions, comments, or experiences of your own on this topic, I’d love to hear from you.