Derek Thompson interviewed Agnes Callard about the idea of the unicontext. I thought it was a useful framework for thinking through how we experience the world right now – and why it leads to certain behaviors and archetypes among people. It reminded me a lot of No Place for Grace by T. Jackson Lears on how people reacted to a change in how the world felt back in the 1900s.

Unicontext & Why Everything On The Internet Looks The Same Now

I’ve spent most of my career in digital marketing, and the single strangest thing I’ve watched happen is this: the better our data got, the more identical everyone’s work became.

Go back far enough and every marketer had their own weird private theories. You had your own sites, your own tests, your own hunches about what Google or Facebook YouTube or Twitter or just your own direct audience wanted. You argued with people at conferences because you genuinely could not verify who was right. It was a craft, and it was a craft practiced in a hundred little rooms that didn’t talk to each other much.

Then the tools got good. Ahrefs, Semrush, Moz, all of it. Suddenly every one of us was looking at the same keyword volumes, the same difficulty scores, the same top-pages reports on the same competitors. And within a few years, every piece of content in every industry started looking the same. Same H2s pulled from People Also Ask. Same 2,000-word minimum. Same “ultimate guide” framing. Same schema. I’ve written some of it. I wrote a very long guide to Ahrefs early on that is, in retrospect, partly a manual for how to do this to yourself.

I always filed that under “the tools got commoditized.” I don’t think that’s what happened anymore.

I listened to an episode of Derek Thompson’s Plain English with Agnes Callard, a philosopher at the University of Chicago, and she gave me a much better name for it. She calls it the unicontext, and once you have the word, you start seeing it in places that have nothing to do with search rankings.

What the unicontext actually is

Start with “context.” Callard defines a context as a normative situation — a set of circumstances that tells you how you’re supposed to act.

For almost all of human history, contexts were local and there were a lot of them. You knew how to behave by looking around. Am I in a field? In my house? In church? In a bar? At a funeral? Each room came with its own rules, and you didn’t have to think about it. You just responded to what was in front of you.

The unicontext is what happens when all of those rooms collapse into one room, and one set of norms applies everywhere, all the time, regardless of where you actually are.

This is not the same thing as context collapse, which is the idea you’ve probably heard – that your tweet is visible to your boss and your mom and your college roommate simultaneously. Callard’s point is that context collapse is about information, and hers is about norms.

Her example: a peasant village had total informational context collapse. Everybody knew everybody’s business. Families slept in one room. Kids knew things about their parents that no one today would know. But that world was nothing like ours, because the norms were still separate — different rules for children and adults, for the home and the field, for winter and harvest. Information collapse only became transformative once the normative collapse happened underneath it.

That distinction is the whole ballgame, and it’s the reason this isn’t just a repackaging of an idea from 2010.

1. one Metric, then homogenization

This is the part that explains my industry, and I think it’s the most useful piece of the theory.

Callard’s example is two school districts. District A goes to School A, District B goes to School B. You may have plenty of information about both schools. But you have no reason to compare them, because they’re not in the same space of choice. Then the rule changes and you can attend either one. Now you need to compare. So you reach for whatever is comparable — graduation rates, AP course counts, college placement.

And then the schools notice. The one without AP classes starts adding AP classes, because that’s the metric it’s being measured on, even though it was deliberately doing something else. The two schools converge.

That is exactly what happened to SEO. It’s exactly what happened to baseball, which Thompson raises in the episode — every team got the same statistics, every player entered the same comparative set, and now every team plays a nearly identical style of baseball and everyone complains that the sport got boring. Same pattern in restaurants chasing the same Instagram-legible dish. Same pattern in every DTC brand’s website looking like every other DTC brand’s website.

The sequence is: one evaluative field → pressure to compare → convergence on whatever is measurable → homogenization.

Nobody chose the homogenization. Everybody chose the comparison.

2. bad travels further than good

Callard’s explanation for why the internet is such a relentlessly negative place is the cleanest one I’ve heard.

Goodness is context-dependent. There is essentially nothing that is good for all people in all circumstances. What makes you happy depends on who you are, where you are, and what kind of day you’re having.

Badness is much more universal. Death, pain, illness, violence. You can see a person whose life you don’t understand at all, in a country you’ve never been to, and if they’re suffering, you get it immediately.

So if you’re posting into a room that contains everyone, and you want to be understood, you will drift toward the negative. Not because you’re a bad person. Because it’s the only thing that’s legible at that scale.

I keep coming back to my own trail posts as the test case. I’ve written detailed write-ups of hikes in the Oconee National Forest that maybe a few hundred people will ever want, because they’re for a person who is specifically going to that specific trailhead. That post is good in a narrow context and basically illegible outside it. If I instead wrote “here’s what the Forest Service is getting letting our trails fall apart,” that travels. Condemnation is portable. A nice creek crossing is not.

Callard pushes this further than I expected. It’s not just that trivial goods don’t travel — it’s that significant goods don’t either. Someone writes a genuinely great poem in a language most of us don’t read, and it needs years and translation and cultural assimilation before it can land. Meanwhile a video of something terrible needs zero setup. The form flattens the norm.

3. identity crowds out character

Character, in the old Aristotle sense, is a set of dispositions — how you handle fear, anger, generosity, temptation. The catch is that you cannot see someone’s character from one encounter. An irascible person isn’t angry all the time. And sometimes everyone is angry, so you can’t tell the irascible person apart. Reading character requires knowing somebody across a lot of different situations.

Identity categories are the opposite. There are the obvious ones like sex, nationality, race, etc – but there’s increasingly the “[X] type of person” created based on appearance, lifestyle, job, etc. You are that in every circumstance. Callard’s line: it’s a hat you never take off.

So in a world where we have to make sense of people we’ll never see in more than one context, we reach for the thing that’s visible in every context. Identity is legible at scale. Character isn’t.

I don’t think this means identity doesn’t matter — it obviously does. But it does explain the gap between how constantly we talk about identity categories and how rarely anyone talks about whether a person is courageous, or honest, or steady. We’ve mostly lost the vocabulary. Notice that you had to be reminded what “character” even means as a technical term, and you didn’t need any reminder about what identity means.

4. you didn’t used to “manage” your attention

The default human setup is bottom-up attention. You attend to whatever is salient. Being distractible is a feature if there might be a fire or a snake.

The unicontext replaces that with top-down attention management, and Callard has a detail here that stood out to me. In 19th century novels, there are all these scenes where a woman is reading a book and gets distracted by her own turbulent emotional life and puts the book down. Anna Karenina, Emma Bovary. Distraction is right there in the text.

What never happens is the character saying to herself: no, I need to get through another twenty pages.

That thought did not exist. Nobody was managing their attention with an iron fist, buying products designed to protect them from other products they’d already bought, treating focus as a resource to be optimized. We do that because we’re standing in a permanent choice relation to our own attention — the screen is in every room, so no room can tell us what to attend to anymore.

5. Caveats

A few honest reservations, because this is the kind of “this explains everything” theory that’s very easy to fall in love with.

It might explain too much. Any theory that accounts for negativity bias, identity politics, phone addiction, sports analytics, and the modernist novel is a theory I’m skeptical of – even if I want to believe it. When everything is evidence, nothing is. I’d want to know what the unicontext predicts that turns out to be wrong.

The technology question isn’t settled. In the interview, Thompson pushes on this and I think he’s right to. People said the radio destroyed local individuality in the 1920s. Then television. Then broadband. Callard’s answer is that it’s technology plus a human impulse — these particular technologies caught on because we wanted the unicontext, and there’s a dialectical thing going on. That’s a good answer. I’m not sure it’s a testable one.

She won’t say whether it’s good or bad, and I go back and forth on whether that’s wisdom or a dodge. Her reasoning, borrowed from Georg Simmel, is that we can’t judge it because we’re its product — it’s not something we stand outside of. She also makes a genuinely clever point that the question is rigged, because the unicontext can only see badness, so of course it renders itself as bad. Fair. But at some point you have to decide how to live.

Useful Takeaways

I don’t have a tidy program, and I’m skipping the obvious “put your phone away at dinner” advice, which Thompson himself calls cheap in the episode. But there are a few notes that I think are useful.

Write for the small room on purpose. My most useful posts are the ones almost nobody needs. The trail write-up for the guy who’s going to that trailhead next Saturday who wants more nuance than you can get from an AllTrails rating. The specific answer to the specific problem that needs experiential context. That work does badly in the universal room by design, and it’s the work I’m proudest of. Being illegible at scale is not the same as being worthless.

Be suspicious of any metric everyone can see. If a number is visible to your whole industry, it is already producing convergence, and the opportunity is almost always in whatever it fails to measure. This has been true in digital marketing for a decade and I keep having to relearn it. It reminds me of the Blue Ocean Strategy.

Notice when you’re only able to say the bad version. If the only way I can make a point travel is to make it a condemnation, that’s a signal about the medium, not about the point. That’s one of the many reasons that even with tech platforms, AI chatbots, and short form video…I still think the humble blog can win.

Judge character where you can still see it. You get context with your neighbors, your employees, your kids, the people you actually work with over years. Stay human.