If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
One of my favorite rhetorical and analytical moves is joining things together (showing that two different, seemingly unrelated ideas are aspects of the same phenomenon) and taking them apart (resolving a paradox by demonstrating that what appears to be one, contradictory thing is actually two different things that have been lumped together).
"Taking things apart" is a very useful framework for understanding AI. How do we resolve the (seeming) paradox that some skilled workers report wonderful results from their work with AI, while others are full of dire warnings about the lurking defects in their AI-assisted outputs? Simple: the first group are "centaurs" (humans who are assisted by machines) and the second are "reverse centaurs" (humans who have been pressed into service as peripherals for machines):
What are we to make of the people who've been fired by bosses who replaced them with AI, in light of the fact that AI is demonstrably not able to do their (former) jobs? Again, it's simple if you separate out two distinct phenomena: "AI can do your job" is the first. The second is: "Your boss is a credulous dolt who is infinitely horny for replacing lippy workers with pliable machines, which made him an easy mark for an AI salesman who convinced him to fire you and replace you with an AI that can't do your job":
This is also a useful move for understanding the AI investment bubble. It's not just billionaires who don't think other people are as real as they are and consequently their jobs can be done by chatbots. It's also billionaires who believe that bosses can be sold AI and don't care if the AI is defective, because that's your boss's problem after he buys the AI and fires you. They don't have to believe in AI in order to think it's a good investment: like an investor betting that Joe Rogan can sell millions of dollars' worth of peptides to desperate young men, they are assessing the sales potential, not the merits of the thing for sale:
https://pluralistic.net/2026/08/03/andor/#either
As useful as "taking things apart" is, "putting things together" is also a very important technique for assessing, critiquing and improving AI. In a stellar essay entitled "Temperature Zero for Culture: Why Everything Is Starting to Look the Same" by the data scientist Lauren Leek, we get a top-notch example of "putting things together":
Leek's essay is one of those fabulous, wide-ranging, cross-disciplinary pieces, touching on urban design, music trends, synthetic LLM crowds, Netflix recommendation algorithms, and several other subjects, all seeking to resolve a(nother) (seeming) paradox: how is it that we have so much potential variety, but everything is so manifestly the same?
The answer is complicated and nuanced, but Leek's foundational point is that in a data-driven society, "predictions" are self-fulfilling prophecies. As Leek puts it: "Once prediction shapes the choices in front of us, we lose the ability to tell the difference between what people wanted and what the system made easy to want."
This is a pervasive issue across many domains. Leek says that economists call it "performativity," while machine learning researchers call it "model collapse" and urbanists call it "placelessness."
"Performativity" describes how, once a market has been modeled by economists, that model becomes the foundation for economic policy, which pushes the market to conform to the model:
"Model collapse" describes how machine learning models that are trained on their own predictions become incredibly bland, with all variety disappearing from the system's predictions:
This is hugely consequential: it's why bias proliferates through predictive policing algorithms: train a model with data from racist stop-and-frisks and it will predict that all the weapons and drugs in a city are to be found in Black and brown peoples' pockets. Turn those predictions into recommendations telling cops where to go look for weapons and drugs and they will double down on racist stops, producing even more biased training data, which turns into still more bias in the predictions:
"Placelessness" is the urbanist's name for "when everywhere optimises toward the same template." I think of it as Flinstones Syndrome, where the same background is looped behind Fred and Barney as they drive through Bedrock. In New York City, it's Citibank-bodega-Chipotle-Walgreens; in the Chicago suburbs, it's the strip malls with a Chili's, a gas station, and a big box store.
Leek proposes that these are all expressions of the same underlying phenomenon, a failure mode of data science that takes a world of "granular personal data" and arrives at a world where "personalisation produc[es] more sameness."
To these excellent examples, I'd add another one, from the world of monetary policy: Goodhart's Law, which holds that "When a measure becomes a target, it ceases to be a good measure":
https://en.wikipedia.org/wiki/Goodhart%27s_law
Goodhart's Law captures a wide variety of phenomena. When Google first deployed Pagerank, they showed that by counting the inbound links to all the pages on the web, you could extract a signal about which pages were most important (because there was no reason to link to a page unless you found it noteworthy).
But once Pagerank became the dominant means by which web users found pages, counting links stopped being useful: first, because people used Pagerank to find the best pages and link to them, making it impossible for new pages to get the inbound links needed to supersede incumbent pages; and second, because it's easy for fraudsters to create inbound links for low-quality pages in bulk, once there's a reason to do so.
Counting inbound links was a world-beating retrospective way of predicting which page would best match a searcher's query, but once it shaped the world it sought to analyze, it ceased to be a good prospective way to predict which page would best match your queries.
Leek is a brilliant data scientist and an even better science communicator, with a knack for crisp, readily understood explanations. How can a world of granular, highly varied data turn into a world of homogeneous choices? Simple: start with a set of items ("cuisines, genres, shop types") and a standard algorithm for sorting them. Let users choose from those recommendations. The mode (average) of those choices "gets shown more, so it gets picked more, so the model grows more confident the mode is what people want, and the tails starve." Run this for a few rounds and the evenly distributed catalog of choices "collapses onto one dominant option."
This is intrinsic in the choices we make in designing recommendation algorithms, tilting them towards the likelihood of a successful recommendation. A recommender that wants to succeed every time will make the safest possible recommendations, "so an algorithm that is uncertain about you, and it is always at least a little uncertain, hedges toward the average."
Then she busts out a beautiful, perfect little statistics aphorism: "Personalisation under a standard loss function is regression to the collective mean with extra steps." That is to say, "regression to the mean" (the tendency of varied things to become more standardized) cannot be avoided with the standard personalization algorithm. That algorithm is going to play it safe, showing you things that are broadly palatable, and because your choices are constrained to the average, you will choose average things.
This is how recommendation systems – and other analytical tools that produce predictions that are then turned into action – force so many diverse phenomena (streets, markets, media recommendations) into sameness. The fact that these recommenders are self-fulfilling prophecies means that "they don't have to be right," only "listened to."
This explains the sameness of so many of London's high streets. Leek examines 640 shopping streets, characterizing 18,000 food places spread out across them, flagging all the chain restaurants. Her analysis shows that any two London streets will, on average, share about half of their "food profile."
Obviously, this is most pronounced on streets with chain outlets, and it doesn't take that many chain outlets before a street's sameness shoots up: "A relatively small number of repeated names is enough to make otherwise different streets resemble one another more." So why do streets with chains resemble one another so much? Because the chains use an algorithm (weighting footfall, proximity to train stations, demographics, and competitors) to decide where to put their restaurants. If a street with a Gail's Bakery on it feels like every other street with a Gail's Bakery, that's because Gail's only puts its restaurants in places that have highly similar characteristics, measured to a high degree of accuracy and controlled by a narrow set of tolerances.
In other words, every street that feels like it should have a Gail's will eventually get a Gail's, whereupon that street will feel even more like all the other streets that have a Gail's, because it will share one more common factor with those other streets (a Gail's).
Leek points here to her earlier work on pub closures in the UK. The UK has experienced an epidemic of pub closures, with thousands of pubs disappearing since 2016:
Her research found that the biggest predictor of a pub surviving was its similarity to the median pub; which is to say that the more distinctive a pub was, the more "character" it had, the more likely it was to close. Pubs that are different from the average pub are harder to categorize, which means they're harder for a bank manager to assess for creditworthiness or for a landlord to justify extending a long-term lease to. The algorithms used to allocate capital and real estate are also recommenders, and they also drive variety out of the system.
This same phenomenon acts on culture. In an age of music recommendation algorithms, hit songs are changing; today's songs use a smaller vocabulary of unique words and repeat those words more often:
Vocabulary richness, distinct words relative to length, has fallen by more than a quarter since the early 1960s, while the share of repeated lines has climbed by nearly a third. The modern hit says less and says it more often, because the hook that works gets repeated.
But that's not the whole story! While each song resembles itself more ("saying less more often"), within that constraint, there's far more variety today than before: a given song's (constrained) vocabulary has grown more distinct when compared to all the other songs' vocabularies. Songs repeat the words they use, but the words repeated in songs are getting more different.
For Leek, this is the key to understanding the whole phenomenon and (more importantly) doing something about it. Music recommendation systems optimized for a singable hook, but did not optimize on any of the other variables in songs, so those dimensions acquired a broader range, even as the optmized variable got flatter and narrower.
This means that the tendency of recommenders to "flatten the world" isn't a single blunt outcome: it depends on which dimension we choose to flatten through recommendation, and who chooses to flatten that dimension.
A media recommender optimizes for consumption, showing you a tractable set of things it believes you'll watch, read or listen to. When you choose from among this limited set, the recommender takes note of that fact and shows you more of the same, pushing everything to a greige median. All the movies, books and songs you might have liked that were omitted from that initial set are excluded from being recommended in the future. The features of that media that you might have appreciated "decay out of consideration." They are never tested for desirability. The model collapses.
How badly does it collapse? Leek cites Movietweetings' data on which movies people watch: out of a million public movie ratings, half relate to the top 2% of movies in the set. There's 38,000 films in the set, but just 380 titles account for 40% of the ratings. Leek argues (persuasively) that this isn't because recommenders are good at "knowing your taste" – rather, they are good at "narrowing the menu."
Leek relates this to her work on creating LLM "personas" – synthetic populations meant to mimic the tastes and proclivities of real groups of people, that you can interrogate "before you spend money asking actual humans." While this would be useful for many applications, "it fails in exactly the way this whole essay is about."
Leek went to enormous lengths to reproduce the traits that make people interesting to study in aggregate, painstakingly replicating the ways that social connections, psychological outlook and demographic factors predict people's beliefs. The result was a set of LLM personas with "elaborate stories" about how they differed from one another, but whose survey responses about planned actions were homogeneous in a way that real populations are not.
This, Leek writes, is the same force that homogenizes other data-driven predictors. Because she'd ordered her LLM to reproduce the statistically validated relationships between different factors that predict a person's beliefs, each synthetic persona was a homogenized average. It's like the paradox of "The Average Man," where military uniforms sized to the average of all service personnel fit no one, because no one is average:
https://archive.org/details/DTIC_AD0010203
The thing is (as Leek points out) the idea that synthetic personas are a good way to understand the preferences of a real population is not a harmless delusion: it's a product that's being actively sold to governments, campaigning politicians and marketers. It's a self-fulfilling prophecy that drives governance, political campaigns and product design to the same homogeneous median that is making every shopping street in London feel the same.
This matters. As Leek writes, ecologists have long understood the importance of variety for systemic resilience: they call it "the insurance value of biodiversity." A diverse system has reservoirs of species and variation that may not be optimized for how things stand now, but that can move into niches created when things change in ways that lay waste to the previously dominant organisms. As anyone whose favorite banana went extinct can tell you, homogeneity works well, but diversity fails well:
https://en.wikipedia.org/wiki/Gros_Michel
The brittleness of algorithm-induced homogeneity is compounded by the fact that recommenders obscure the true preferences of people. If you watch two Scandinavian crime dramas after Netflix recommends them to you, it will keep showing you more Scandy crime for the next decade – even if there's another kind of programming that you'd vastly prefer (if only you knew about it). This means that decision-makers who choose which shows will get made in the future will keep on funding their safe Danish detectives, to the exclusion of whatever might emerge from the same weird attractor that produced the K-Pop Demon Hunter fortune.
Transpose this failure mode onto states, bank managers and landlords, and we see whole ranges of policies, businesses and activities that never come into existence, despite the popularity, prosperity and joy they might bring us.
But Leek doesn't end with this worrisome note. Instead, she identifies this whole thing – model collapse, placelessness, performativity, even Goodhart's Law – as an expression of one of the best-understood tradeoffs in computer science: "exploration vs exploitation":
Any system learning from feedback has to divide its effort between exploiting what already scores well and exploring options it hasn’t tried, in case they’re better.
Computer scientists have long understood that focusing on exploitation to the exclusion of exploration is a trap that locks you into "the first decent option" so you can never discover the best one.
Which means that this algorithmic homogeneity has a well-understood corrective: "forcing exploration back in." The problem is that markets hate this kind of exploration. A company that lives and dies by how many clicks it gets is never going to sacrifice 20% of its traffic by showing its users weird, untested options that score worse than the median because these weird things have never had a chance to prove that they are desirable.
This is a classic market failure, and, as Leek points out, there are regulatory responses in the UK (the Digital Markets, Competition and Consumers Act) and the EU (the Digital Services Act), both of which require the largest platforms to open up their recommendation systems, but so far, regulators have focused on "online harms" rather than variety (though the DSA does require platforms to offer algorithmic recommendations that are not based on your personal traits).
Leek identifies this willingness of states to set conditions for algorithm design as a means by which "exploration" can be forced back into the system. She's also bullish on interoperability, so that users can leave platforms with bad recommenders, without losing access to their media or social circles. As she writes, "the deepest discipline on a feed that has trapped you is the credible ability to leave it and take your data with you." I couldn't agree more:
She's less hopeful about individual responses. Demanding that you be an "adventurous consumer" is a way of letting systems off the hook. When every street has the same restaurants and every bookshop has the same books and the people in your life are all locked into one of two social media platforms, "choosing wisely" only gets you so far. Shopping isn't politics!
Leek is a superb writer. After reading this piece yesterday, I sent it to half a dozen people and then read everything else in Leek's newsletter archives. Not only is it all brilliant, but I also realized that she'd written one of the most memorable articles about cities and platforms I've read in the last year, "How Google Maps quietly allocates survival across London’s restaurants – and how I built a dashboard to see through it":
I should have added Leek's newsletter to my RSS reader when I read that last December. I've rectified that oversight! What a fantastic thinker, scientist and communicator! If she isn't being relentlessly pestered by editors and literary agents offering her a book deal, then it really does prove that the recommender systems are elevating the bland median over the thoroughly, delightfully spiky outliers.
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
One of my favorite rhetorical and analytical moves is joining things together (showing that two different, seemingly unrelated ideas are aspects of the same phenomenon) and taking them apart (resolving a paradox by demonstrating that what appears to be one, contradictory thing is actually two different things that have been lumped together).
"Taking things apart" is a very useful framework for understanding AI. How do we resolve the (seeming) paradox that some skilled workers report wonderful results from their work with AI, while others are full of dire warnings about the lurking defects in their AI-assisted outputs? Simple: the first group are "centaurs" (humans who are assisted by machines) and the second are "reverse centaurs" (humans who have been pressed into service as peripherals for machines):
What are we to make of the people who've been fired by bosses who replaced them with AI, in light of the fact that AI is demonstrably not able to do their (former) jobs? Again, it's simple if you separate out two distinct phenomena: "AI can do your job" is the first. The second is: "Your boss is a credulous dolt who is infinitely horny for replacing lippy workers with pliable machines, which made him an easy mark for an AI salesman who convinced him to fire you and replace you with an AI that can't do your job":
This is also a useful move for understanding the AI investment bubble. It's not just billionaires who don't think other people are as real as they are and consequently their jobs can be done by chatbots. It's also billionaires who believe that bosses can be sold AI and don't care if the AI is defective, because that's your boss's problem after he buys the AI and fires you. They don't have to believe in AI in order to think it's a good investment: like an investor betting that Joe Rogan can sell millions of dollars' worth of peptides to desperate young men, they are assessing the sales potential, not the merits of the thing for sale:
https://pluralistic.net/2026/08/03/andor/#either
As useful as "taking things apart" is, "putting things together" is also a very important technique for assessing, critiquing and improving AI. In a stellar essay entitled "Temperature Zero for Culture: Why Everything Is Starting to Look the Same" by the data scientist Lauren Leek, we get a top-notch example of "putting things together":
Leek's essay is one of those fabulous, wide-ranging, cross-disciplinary pieces, touching on urban design, music trends, synthetic LLM crowds, Netflix recommendation algorithms, and several other subjects, all seeking to resolve a(nother) (seeming) paradox: how is it that we have so much potential variety, but everything is so manifestly the same?
The answer is complicated and nuanced, but Leek's foundational point is that in a data-driven society, "predictions" are self-fulfilling prophecies. As Leek puts it: "Once prediction shapes the choices in front of us, we lose the ability to tell the difference between what people wanted and what the system made easy to want."
This is a pervasive issue across many domains. Leek says that economists call it "performativity," while machine learning researchers call it "model collapse" and urbanists call it "placelessness."
"Performativity" describes how, once a market has been modeled by economists, that model becomes the foundation for economic policy, which pushes the market to conform to the model:
"Model collapse" describes how machine learning models that are trained on their own predictions become incredibly bland, with all variety disappearing from the system's predictions:
This is hugely consequential: it's why bias proliferates through predictive policing algorithms: train a model with data from racist stop-and-frisks and it will predict that all the weapons and drugs in a city are to be found in Black and brown peoples' pockets. Turn those predictions into recommendations telling cops where to go look for weapons and drugs and they will double down on racist stops, producing even more biased training data, which turns into still more bias in the predictions:
"Placelessness" is the urbanist's name for "when everywhere optimises toward the same template." I think of it as Flinstones Syndrome, where the same background is looped behind Fred and Barney as they drive through Bedrock. In New York City, it's Citibank-bodega-Chipotle-Walgreens; in the Chicago suburbs, it's the strip malls with a Chili's, a gas station, and a big box store.
Leek proposes that these are all expressions of the same underlying phenomenon, a failure mode of data science that takes a world of "granular personal data" and arrives at a world where "personalisation produc[es] more sameness."
To these excellent examples, I'd add another one, from the world of monetary policy: Goodhart's Law, which holds that "When a measure becomes a target, it ceases to be a good measure":
https://en.wikipedia.org/wiki/Goodhart%27s_law
Goodhart's Law captures a wide variety of phenomena. When Google first deployed Pagerank, they showed that by counting the inbound links to all the pages on the web, you could extract a signal about which pages were most important (because there was no reason to link to a page unless you found it noteworthy).
But once Pagerank became the dominant means by which web users found pages, counting links stopped being useful: first, because people used Pagerank to find the best pages and link to them, making it impossible for new pages to get the inbound links needed to supersede incumbent pages; and second, because it's easy for fraudsters to create inbound links for low-quality pages in bulk, once there's a reason to do so.
Counting inbound links was a world-beating retrospective way of predicting which page would best match a searcher's query, but once it shaped the world it sought to analyze, it ceased to be a good prospective way to predict which page would best match your queries.
Leek is a brilliant data scientist and an even better science communicator, with a knack for crisp, readily understood explanations. How can a world of granular, highly varied data turn into a world of homogeneous choices? Simple: start with a set of items ("cuisines, genres, shop types") and a standard algorithm for sorting them. Let users choose from those recommendations. The mode (average) of those choices "gets shown more, so it gets picked more, so the model grows more confident the mode is what people want, and the tails starve." Run this for a few rounds and the evenly distributed catalog of choices "collapses onto one dominant option."
This is intrinsic in the choices we make in designing recommendation algorithms, tilting them towards the likelihood of a successful recommendation. A recommender that wants to succeed every time will make the safest possible recommendations, "so an algorithm that is uncertain about you, and it is always at least a little uncertain, hedges toward the average."
Then she busts out a beautiful, perfect little statistics aphorism: "Personalisation under a standard loss function is regression to the collective mean with extra steps." That is to say, "regression to the mean" (the tendency of varied things to become more standardized) cannot be avoided with the standard personalization algorithm. That algorithm is going to play it safe, showing you things that are broadly palatable, and because your choices are constrained to the average, you will choose average things.
This is how recommendation systems – and other analytical tools that produce predictions that are then turned into action – force so many diverse phenomena (streets, markets, media recommendations) into sameness. The fact that these recommenders are self-fulfilling prophecies means that "they don't have to be right," only "listened to."
This explains the sameness of so many of London's high streets. Leek examines 640 shopping streets, characterizing 18,000 food places spread out across them, flagging all the chain restaurants. Her analysis shows that any two London streets will, on average, share about half of their "food profile."
Obviously, this is most pronounced on streets with chain outlets, and it doesn't take that many chain outlets before a street's sameness shoots up: "A relatively small number of repeated names is enough to make otherwise different streets resemble one another more." So why do streets with chains resemble one another so much? Because the chains use an algorithm (weighting footfall, proximity to train stations, demographics, and competitors) to decide where to put their restaurants. If a street with a Gail's Bakery on it feels like every other street with a Gail's Bakery, that's because Gail's only puts its restaurants in places that have highly similar characteristics, measured to a high degree of accuracy and controlled by a narrow set of tolerances.
In other words, every street that feels like it should have a Gail's will eventually get a Gail's, whereupon that street will feel even more like all the other streets that have a Gail's, because it will share one more common factor with those other streets (a Gail's).
Leek points here to her earlier work on pub closures in the UK. The UK has experienced an epidemic of pub closures, with thousands of pubs disappearing since 2016:
Her research found that the biggest predictor of a pub surviving was its similarity to the median pub; which is to say that the more distinctive a pub was, the more "character" it had, the more likely it was to close. Pubs that are different from the average pub are harder to categorize, which means they're harder for a bank manager to assess for creditworthiness or for a landlord to justify extending a long-term lease to. The algorithms used to allocate capital and real estate are also recommenders, and they also drive variety out of the system.
This same phenomenon acts on culture. In an age of music recommendation algorithms, hit songs are changing; today's songs use a smaller vocabulary of unique words and repeat those words more often:
Vocabulary richness, distinct words relative to length, has fallen by more than a quarter since the early 1960s, while the share of repeated lines has climbed by nearly a third. The modern hit says less and says it more often, because the hook that works gets repeated.
But that's not the whole story! While each song resembles itself more ("saying less more often"), within that constraint, there's far more variety today than before: a given song's (constrained) vocabulary has grown more distinct when compared to all the other songs' vocabularies. Songs repeat the words they use, but the words repeated in songs are getting more different.
For Leek, this is the key to understanding the whole phenomenon and (more importantly) doing something about it. Music recommendation systems optimized for a singable hook, but did not optimize on any of the other variables in songs, so those dimensions acquired a broader range, even as the optmized variable got flatter and narrower.
This means that the tendency of recommenders to "flatten the world" isn't a single blunt outcome: it depends on which dimension we choose to flatten through recommendation, and who chooses to flatten that dimension.
A media recommender optimizes for consumption, showing you a tractable set of things it believes you'll watch, read or listen to. When you choose from among this limited set, the recommender takes note of that fact and shows you more of the same, pushing everything to a greige median. All the movies, books and songs you might have liked that were omitted from that initial set are excluded from being recommended in the future. The features of that media that you might have appreciated "decay out of consideration." They are never tested for desirability. The model collapses.
How badly does it collapse? Leek cites Movietweetings' data on which movies people watch: out of a million public movie ratings, half relate to the top 2% of movies in the set. There's 38,000 films in the set, but just 380 titles account for 40% of the ratings. Leek argues (persuasively) that this isn't because recommenders are good at "knowing your taste" – rather, they are good at "narrowing the menu."
Leek relates this to her work on creating LLM "personas" – synthetic populations meant to mimic the tastes and proclivities of real groups of people, that you can interrogate "before you spend money asking actual humans." While this would be useful for many applications, "it fails in exactly the way this whole essay is about."
Leek went to enormous lengths to reproduce the traits that make people interesting to study in aggregate, painstakingly replicating the ways that social connections, psychological outlook and demographic factors predict people's beliefs. The result was a set of LLM personas with "elaborate stories" about how they differed from one another, but whose survey responses about planned actions were homogeneous in a way that real populations are not.
This, Leek writes, is the same force that homogenizes other data-driven predictors. Because she'd ordered her LLM to reproduce the statistically validated relationships between different factors that predict a person's beliefs, each synthetic persona was a homogenized average. It's like the paradox of "The Average Man," where military uniforms sized to the average of all service personnel fit no one, because no one is average:
https://archive.org/details/DTIC_AD0010203
The thing is (as Leek points out) the idea that synthetic personas are a good way to understand the preferences of a real population is not a harmless delusion: it's a product that's being actively sold to governments, campaigning politicians and marketers. It's a self-fulfilling prophecy that drives governance, political campaigns and product design to the same homogeneous median that is making every shopping street in London feel the same.
This matters. As Leek writes, ecologists have long understood the importance of variety for systemic resilience: they call it "the insurance value of biodiversity." A diverse system has reservoirs of species and variation that may not be optimized for how things stand now, but that can move into niches created when things change in ways that lay waste to the previously dominant organisms. As anyone whose favorite banana went extinct can tell you, homogeneity works well, but diversity fails well:
https://en.wikipedia.org/wiki/Gros_Michel
The brittleness of algorithm-induced homogeneity is compounded by the fact that recommenders obscure the true preferences of people. If you watch two Scandinavian crime dramas after Netflix recommends them to you, it will keep showing you more Scandy crime for the next decade – even if there's another kind of programming that you'd vastly prefer (if only you knew about it). This means that decision-makers who choose which shows will get made in the future will keep on funding their safe Danish detectives, to the exclusion of whatever might emerge from the same weird attractor that produced the K-Pop Demon Hunter fortune.
Transpose this failure mode onto states, bank managers and landlords, and we see whole ranges of policies, businesses and activities that never come into existence, despite the popularity, prosperity and joy they might bring us.
But Leek doesn't end with this worrisome note. Instead, she identifies this whole thing – model collapse, placelessness, performativity, even Goodhart's Law – as an expression of one of the best-understood tradeoffs in computer science: "exploration vs exploitation":
Any system learning from feedback has to divide its effort between exploiting what already scores well and exploring options it hasn’t tried, in case they’re better.
Computer scientists have long understood that focusing on exploitation to the exclusion of exploration is a trap that locks you into "the first decent option" so you can never discover the best one.
Which means that this algorithmic homogeneity has a well-understood corrective: "forcing exploration back in." The problem is that markets hate this kind of exploration. A company that lives and dies by how many clicks it gets is never going to sacrifice 20% of its traffic by showing its users weird, untested options that score worse than the median because these weird things have never had a chance to prove that they are desirable.
This is a classic market failure, and, as Leek points out, there are regulatory responses in the UK (the Digital Markets, Competition and Consumers Act) and the EU (the Digital Services Act), both of which require the largest platforms to open up their recommendation systems, but so far, regulators have focused on "online harms" rather than variety (though the DSA does require platforms to offer algorithmic recommendations that are not based on your personal traits).
Leek identifies this willingness of states to set conditions for algorithm design as a means by which "exploration" can be forced back into the system. She's also bullish on interoperability, so that users can leave platforms with bad recommenders, without losing access to their media or social circles. As she writes, "the deepest discipline on a feed that has trapped you is the credible ability to leave it and take your data with you." I couldn't agree more:
She's less hopeful about individual responses. Demanding that you be an "adventurous consumer" is a way of letting systems off the hook. When every street has the same restaurants and every bookshop has the same books and the people in your life are all locked into one of two social media platforms, "choosing wisely" only gets you so far. Shopping isn't politics!
Leek is a superb writer. After reading this piece yesterday, I sent it to half a dozen people and then read everything else in Leek's newsletter archives. Not only is it all brilliant, but I also realized that she'd written one of the most memorable articles about cities and platforms I've read in the last year, "How Google Maps quietly allocates survival across London’s restaurants – and how I built a dashboard to see through it":
I should have added Leek's newsletter to my RSS reader when I read that last December. I've rectified that oversight! What a fantastic thinker, scientist and communicator! If she isn't being relentlessly pestered by editors and literary agents offering her a book deal, then it really does prove that the recommender systems are elevating the bland median over the thoroughly, delightfully spiky outliers.
Object permanence: Wired v Dutch hackers; NYT v DMCA; Hair gel terrorist threat does not exist; AT&T merger is a screwjob; Smart cities are stupid; RIP Reaganomics.
#25yrsago Awful, stupid Wired report on Dutch hacker camp https://web.archive.org/web/20011007084604/https://www.wired.com/news/culture/0,1284,46033,00.html
#25yrsaog Excellent NYT story about the internal contradictions of the DMCA https://memex.craphound.com/2001/08/13/excellent-nyt-story-about-the/
#20yrsago Our faulty intuition about open systems https://www.ft.com/content/64167124-263d-11db-afa1-0000779e2340
#20yrsago Defending against the last plot won’t save us from the next one https://www.schneier.com/blog/archives/2006/08/terrorism_secur.html
#20yrsago NBC: Hair-gel terrorists posed no risk last week https://web.archive.org/web/20060813194630/http://www.msnbc.msn.com/id/14320452/
#15yrsago AT&T merger leak: it’s all about raising prices and reducing competition https://web.archive.org/web/20110920222524/http://www.broadbandreports.com/shownews/Leaked-ATT-Letter-Demolishes-Case-For-TMobile-Merger-115652
#10yrsago What’s inside a Tiki Bird? https://miehana.blogspot.com/2016/08/fancy-feathers-restoring-tiki-room-birds.html
#5yrsago End of the line for Reaganomics https://pluralistic.net/2021/08/13/post-bork-era/#manne-down
#5yrsago Smart cities are neither, 2021 edition https://pluralistic.net/2021/08/13/post-bork-era/#our-streets
#1yrago Maga's boss class think they are immune to American carnage https://pluralistic.net/2025/08/13/then-they-came-for-me/#boss-politics
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
One of my favorite rhetorical and analytical moves is joining things together (showing that two different, seemingly unrelated ideas are aspects of the same phenomenon) and taking them apart (resolving a paradox by demonstrating that what appears to be one, contradictory thing is actually two different things that have been lumped together).
"Taking things apart" is a very useful framework for understanding AI. How do we resolve the (seeming) paradox that some skilled workers report wonderful results from their work with AI, while others are full of dire warnings about the lurking defects in their AI-assisted outputs? Simple: the first group are "centaurs" (humans who are assisted by machines) and the second are "reverse centaurs" (humans who have been pressed into service as peripherals for machines):
What are we to make of the people who've been fired by bosses who replaced them with AI, in light of the fact that AI is demonstrably not able to do their (former) jobs? Again, it's simple if you separate out two distinct phenomena: "AI can do your job" is the first. The second is: "Your boss is a credulous dolt who is infinitely horny for replacing lippy workers with pliable machines, which made him an easy mark for an AI salesman who convinced him to fire you and replace you with an AI that can't do your job":
This is also a useful move for understanding the AI investment bubble. It's not just billionaires who don't think other people are as real as they are and consequently their jobs can be done by chatbots. It's also billionaires who believe that bosses can be sold AI and don't care if the AI is defective, because that's your boss's problem after he buys the AI and fires you. They don't have to believe in AI in order to think it's a good investment: like an investor betting that Joe Rogan can sell millions of dollars' worth of peptides to desperate young men, they are assessing the sales potential, not the merits of the thing for sale:
https://pluralistic.net/2026/08/03/andor/#either
As useful as "taking things apart" is, "putting things together" is also a very important technique for assessing, critiquing and improving AI. In a stellar essay entitled "Temperature Zero for Culture: Why Everything Is Starting to Look the Same" by the data scientist Lauren Leek, we get a top-notch example of "putting things together":
Leek's essay is one of those fabulous, wide-ranging, cross-disciplinary pieces, touching on urban design, music trends, synthetic LLM crowds, Netflix recommendation algorithms, and several other subjects, all seeking to resolve a(nother) (seeming) paradox: how is it that we have so much potential variety, but everything is so manifestly the same?
The answer is complicated and nuanced, but Leek's foundational point is that in a data-driven society, "predictions" are self-fulfilling prophecies. As Leek puts it: "Once prediction shapes the choices in front of us, we lose the ability to tell the difference between what people wanted and what the system made easy to want."
This is a pervasive issue across many domains. Leek says that economists call it "performativity," while machine learning researchers call it "model collapse" and urbanists call it "placelessness."
"Performativity" describes how, once a market has been modeled by economists, that model becomes the foundation for economic policy, which pushes the market to conform to the model:
"Model collapse" describes how machine learning models that are trained on their own predictions become incredibly bland, with all variety disappearing from the system's predictions:
This is hugely consequential: it's why bias proliferates through predictive policing algorithms: train a model with data from racist stop-and-frisks and it will predict that all the weapons and drugs in a city are to be found in Black and brown peoples' pockets. Turn those predictions into recommendations telling cops where to go look for weapons and drugs and they will double down on racist stops, producing even more biased training data, which turns into still more bias in the predictions:
"Placelessness" is the urbanist's name for "when everywhere optimises toward the same template." I think of it as Flinstones Syndrome, where the same background is looped behind Fred and Barney as they drive through Bedrock. In New York City, it's Citibank-bodega-Chipotle-Walgreens; in the Chicago suburbs, it's the strip malls with a Chili's, a gas station, and a big box store.
Leek proposes that these are all expressions of the same underlying phenomenon, a failure mode of data science that takes a world of "granular personal data" and arrives at a world where "personalisation produc[es] more sameness."
To these excellent examples, I'd add another one, from the world of monetary policy: Goodhart's Law, which holds that "When a measure becomes a target, it ceases to be a good measure":
https://en.wikipedia.org/wiki/Goodhart%27s_law
Goodhart's Law captures a wide variety of phenomena. When Google first deployed Pagerank, they showed that by counting the inbound links to all the pages on the web, you could extract a signal about which pages were most important (because there was no reason to link to a page unless you found it noteworthy).
But once Pagerank became the dominant means by which web users found pages, counting links stopped being useful: first, because people used Pagerank to find the best pages and link to them, making it impossible for new pages to get the inbound links needed to supersede incumbent pages; and second, because it's easy for fraudsters to create inbound links for low-quality pages in bulk, once there's a reason to do so.
Counting inbound links was a world-beating retrospective way of predicting which page would best match a searcher's query, but once it shaped the world it sought to analyze, it ceased to be a good prospective way to predict which page would best match your queries.
Leek is a brilliant data scientist and an even better science communicator, with a knack for crisp, readily understood explanations. How can a world of granular, highly varied data turn into a world of homogeneous choices? Simple: start with a set of items ("cuisines, genres, shop types") and a standard algorithm for sorting them. Let users choose from those recommendations. The mode (average) of those choices "gets shown more, so it gets picked more, so the model grows more confident the mode is what people want, and the tails starve." Run this for a few rounds and the evenly distributed catalog of choices "collapses onto one dominant option."
This is intrinsic in the choices we make in designing recommendation algorithms, tilting them towards the likelihood of a successful recommendation. A recommender that wants to succeed every time will make the safest possible recommendations, "so an algorithm that is uncertain about you, and it is always at least a little uncertain, hedges toward the average."
Then she busts out a beautiful, perfect little statistics aphorism: "Personalisation under a standard loss function is regression to the collective mean with extra steps." That is to say, "regression to the mean" (the tendency of varied things to become more standardized) cannot be avoided with the standard personalization algorithm. That algorithm is going to play it safe, showing you things that are broadly palatable, and because your choices are constrained to the average, you will choose average things.
This is how recommendation systems – and other analytical tools that produce predictions that are then turned into action – force so many diverse phenomena (streets, markets, media recommendations) into sameness. The fact that these recommenders are self-fulfilling prophecies means that "they don't have to be right," only "listened to."
This explains the sameness of so many of London's high streets. Leek examines 640 shopping streets, characterizing 18,000 food places spread out across them, flagging all the chain restaurants. Her analysis shows that any two London streets will, on average, share about half of their "food profile."
Obviously, this is most pronounced on streets with chain outlets, and it doesn't take that many chain outlets before a street's sameness shoots up: "A relatively small number of repeated names is enough to make otherwise different streets resemble one another more." So why do streets with chains resemble one another so much? Because the chains use an algorithm (weighting footfall, proximity to train stations, demographics, and competitors) to decide where to put their restaurants. If a street with a Gail's Bakery on it feels like every other street with a Gail's Bakery, that's because Gail's only puts its restaurants in places that have highly similar characteristics, measured to a high degree of accuracy and controlled by a narrow set of tolerances.
In other words, every street that feels like it should have a Gail's will eventually get a Gail's, whereupon that street will feel even more like all the other streets that have a Gail's, because it will share one more common factor with those other streets (a Gail's).
Leek points here to her earlier work on pub closures in the UK. The UK has experienced an epidemic of pub closures, with thousands of pubs disappearing since 2016:
Her research found that the biggest predictor of a pub surviving was its similarity to the median pub; which is to say that the more distinctive a pub was, the more "character" it had, the more likely it was to close. Pubs that are different from the average pub are harder to categorize, which means they're harder for a bank manager to assess for creditworthiness or for a landlord to justify extending a long-term lease to. The algorithms used to allocate capital and real estate are also recommenders, and they also drive variety out of the system.
This same phenomenon acts on culture. In an age of music recommendation algorithms, hit songs are changing; today's songs use a smaller vocabulary of unique words and repeat those words more often:
Vocabulary richness, distinct words relative to length, has fallen by more than a quarter since the early 1960s, while the share of repeated lines has climbed by nearly a third. The modern hit says less and says it more often, because the hook that works gets repeated.
But that's not the whole story! While each song resembles itself more ("saying less more often"), within that constraint, there's far more variety today than before: a given song's (constrained) vocabulary has grown more distinct when compared to all the other songs' vocabularies. Songs repeat the words they use, but the words repeated in songs are getting more different.
For Leek, this is the key to understanding the whole phenomenon and (more importantly) doing something about it. Music recommendation systems optimized for a singable hook, but did not optimize on any of the other variables in songs, so those dimensions acquired a broader range, even as the optmized variable got flatter and narrower.
This means that the tendency of recommenders to "flatten the world" isn't a single blunt outcome: it depends on which dimension we choose to flatten through recommendation, and who chooses to flatten that dimension.
A media recommender optimizes for consumption, showing you a tractable set of things it believes you'll watch, read or listen to. When you choose from among this limited set, the recommender takes note of that fact and shows you more of the same, pushing everything to a greige median. All the movies, books and songs you might have liked that were omitted from that initial set are excluded from being recommended in the future. The features of that media that you might have appreciated "decay out of consideration." They are never tested for desirability. The model collapses.
How badly does it collapse? Leek cites Movietweetings' data on which movies people watch: out of a million public movie ratings, half relate to the top 2% of movies in the set. There's 38,000 films in the set, but just 380 titles account for 40% of the ratings. Leek argues (persuasively) that this isn't because recommenders are good at "knowing your taste" – rather, they are good at "narrowing the menu."
Leek relates this to her work on creating LLM "personas" – synthetic populations meant to mimic the tastes and proclivities of real groups of people, that you can interrogate "before you spend money asking actual humans." While this would be useful for many applications, "it fails in exactly the way this whole essay is about."
Leek went to enormous lengths to reproduce the traits that make people interesting to study in aggregate, painstakingly replicating the ways that social connections, psychological outlook and demographic factors predict people's beliefs. The result was a set of LLM personas with "elaborate stories" about how they differed from one another, but whose survey responses about planned actions were homogeneous in a way that real populations are not.
This, Leek writes, is the same force that homogenizes other data-driven predictors. Because she'd ordered her LLM to reproduce the statistically validated relationships between different factors that predict a person's beliefs, each synthetic persona was a homogenized average. It's like the paradox of "The Average Man," where military uniforms sized to the average of all service personnel fit no one, because no one is average:
https://archive.org/details/DTIC_AD0010203
The thing is (as Leek points out) the idea that synthetic personas are a good way to understand the preferences of a real population is not a harmless delusion: it's a product that's being actively sold to governments, campaigning politicians and marketers. It's a self-fulfilling prophecy that drives governance, political campaigns and product design to the same homogeneous median that is making every shopping street in London feel the same.
This matters. As Leek writes, ecologists have long understood the importance of variety for systemic resilience: they call it "the insurance value of biodiversity." A diverse system has reservoirs of species and variation that may not be optimized for how things stand now, but that can move into niches created when things change in ways that lay waste to the previously dominant organisms. As anyone whose favorite banana went extinct can tell you, homogeneity works well, but diversity fails well:
https://en.wikipedia.org/wiki/Gros_Michel
The brittleness of algorithm-induced homogeneity is compounded by the fact that recommenders obscure the true preferences of people. If you watch two Scandinavian crime dramas after Netflix recommends them to you, it will keep showing you more Scandy crime for the next decade – even if there's another kind of programming that you'd vastly prefer (if only you knew about it). This means that decision-makers who choose which shows will get made in the future will keep on funding their safe Danish detectives, to the exclusion of whatever might emerge from the same weird attractor that produced the K-Pop Demon Hunter fortune.
Transpose this failure mode onto states, bank managers and landlords, and we see whole ranges of policies, businesses and activities that never come into existence, despite the popularity, prosperity and joy they might bring us.
But Leek doesn't end with this worrisome note. Instead, she identifies this whole thing – model collapse, placelessness, performativity, even Goodhart's Law – as an expression of one of the best-understood tradeoffs in computer science: "exploration vs exploitation":
Any system learning from feedback has to divide its effort between exploiting what already scores well and exploring options it hasn’t tried, in case they’re better.
Computer scientists have long understood that focusing on exploitation to the exclusion of exploration is a trap that locks you into "the first decent option" so you can never discover the best one.
Which means that this algorithmic homogeneity has a well-understood corrective: "forcing exploration back in." The problem is that markets hate this kind of exploration. A company that lives and dies by how many clicks it gets is never going to sacrifice 20% of its traffic by showing its users weird, untested options that score worse than the median because these weird things have never had a chance to prove that they are desirable.
This is a classic market failure, and, as Leek points out, there are regulatory responses in the UK (the Digital Markets, Competition and Consumers Act) and the EU (the Digital Services Act), both of which require the largest platforms to open up their recommendation systems, but so far, regulators have focused on "online harms" rather than variety (though the DSA does require platforms to offer algorithmic recommendations that are not based on your personal traits).
Leek identifies this willingness of states to set conditions for algorithm design as a means by which "exploration" can be forced back into the system. She's also bullish on interoperability, so that users can leave platforms with bad recommenders, without losing access to their media or social circles. As she writes, "the deepest discipline on a feed that has trapped you is the credible ability to leave it and take your data with you." I couldn't agree more:
She's less hopeful about individual responses. Demanding that you be an "adventurous consumer" is a way of letting systems off the hook. When every street has the same restaurants and every bookshop has the same books and the people in your life are all locked into one of two social media platforms, "choosing wisely" only gets you so far. Shopping isn't politics!
Leek is a superb writer. After reading this piece yesterday, I sent it to half a dozen people and then read everything else in Leek's newsletter archives. Not only is it all brilliant, but I also realized that she'd written one of the most memorable articles about cities and platforms I've read in the last year, "How Google Maps quietly allocates survival across London’s restaurants – and how I built a dashboard to see through it":
I should have added Leek's newsletter to my RSS reader when I read that last December. I've rectified that oversight! What a fantastic thinker, scientist and communicator! If she isn't being relentlessly pestered by editors and literary agents offering her a book deal, then it really does prove that the recommender systems are elevating the bland median over the thoroughly, delightfully spiky outliers.
While her evening meal is cooking, Mrs Day settles down on her bed with the evening paper and a spot of sewing. She is working on a balaclava and is accompanied by her cat 'Little One'.
A DAY IN THE LIFE OF A WARTIME HOUSEWIFE: EVERYDAY LIFE IN LONDON, ENGLAND, 1941
If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
In the summer of 2013, two esoteric, technical, incredibly important texts were published within weeks of one another: the first is the Snowden leaks, which revealed a system of global, pervasive digital surveillance; the second was Thomas Piketty's Capital in the 21st Century, a book about the economic inevitability (and political instability) of oligarchy:
If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
In 2013, it wasn't immediately apparent how these two works connected with one another, but in the years since, I've grown increasingly convinced that Snowden and Piketty can only be properly understood as describing two aspects of the same phenomenon.
Piketty's landmark volume was grounded in a detailed analysis of 300 years' (!) worth of global capital flows, painstakingly compiled by a large team of grad students from a massive set of heterogeneous records. The book's conclusion is the statement that "returns to capital exceed the rate of growth over the long term" (abbreviated as "r > g").
This may sound innocuous, but it is explosive. If r > g, then the most wealth will inevitably accumulate in the hands of people who start with the most wealth, irrespective of whether they do anything productive with that money. This means that the alleged heroes of the market system – the entrepreneurs who found and manage the firms that increase public prosperity – are doomed to play second fiddle to the mere plumbers of money, people who "contribute" by accumulating.
The starkest example of this in Capital 21C is Piketty's contrast between L'Oreal heiress Liliane Bettencourt (then the richest woman in the world) and Bill Gates, founder of Microsoft (then the most successful corporation in the world). Piketty compares the growth in the fortunes of Bettencourt and Gates over two periods: first, the period between Microsoft's founding and Gates' retirement as CEO; and second, the period after Gates's retirement from his executive role, when he became a mere investor, no longer an entrepreneur.
During that first period, in which Gates was founding and running the most successful corporation in the world, he accumulated less wealth than did Liliane Bettencourt, who did precisely nothing of value over that period. Bettencourt didn't even manage her investments – that was all handled by some very clever financial planners, lawyers and accountants. In other words: for Bettencourt, doing nothing at all produced more wealth as founding the most successful corporation in the world did for Gates. Bettencourt, a person who owned things, did better than Gates, a person who did things.
And then Gates retired. He stopped doing things and started owning things. He became an investor, whereupon he out-earned both Bettencourt and Gates-the-entrepreneur. Again, the market system allocated fewer rewards to the most successful person in the doing things business than it allocated to that same person once he quit that job and got into the owning things business.
Piketty shows that this holds true across markets and nations and eras: all other things being equal, the market system produces a class of hereditary aristocrats who command the world's capital and direct its deployment, despite never having done anything. The market's most lavish rewards do not go to its most productive participants, but rather, to those participants who have the good fortune to emerge from the luckiest of orifices.
Worse: winning the orifice lottery in no way qualifies you to direct the capital you've inherited. Liliane Bettencourt had no revolutionary new business ideas, invented no miraculous new materials or processes, produced no brilliant art. She merely accumulated, thanks to the professional services of skilled technicians whose job description includes hiring their own successors to ensure that another generation of winners of the Bettencourt orifice lottery could continue to accumulate, commanding more capital and power in society.
Perhaps if these orifice winners were content to allow their bloodless Renfields to allocate their capital while consuming bonbons and attending yacht parties, this could yield a stable politics. But inevitably, people who win the orifice lottery observe that they come from a long line of wealthy people, a line that will continue with their own descendants, and conclude that they have some kind of special, heritable virtue – magic blood – that the system has recognized with their great fortunes and the power those fortunes confer.
That's when things get dangerous: when aristocrats grow bored with their leisure and mobilize their inherited capital to change the way the rest of us live. Billionaire dilettantes are weapons of mass destruction, and their special projects have a wide blast radius and inflict a lot of collateral damage.
Take Bill Gates: his ideological projects have been a catastrophe. A patent maximalist, he funded the lobbyists who successfully blocked South Africa from producing its own AIDS drugs under an IP waiver program, and then deployed them again to stop the Global South from making their own covid vaccines:
Closer to home, Gates's hatred of public institutions led him to allocate millions to dismantling public schools and replacing them with charter schools, particularly for poor and racialized kids, with disastrous results:
Capital's tendency to accumulate in the hands of the already wealthy (r > g) means that these aristocrats end up setting an ever-larger proportion of our societal agenda, despite their manifest unfitness to govern and their absence of any kind of democratic legitimacy.
Piketty argues that inequality is inherently politically destabilizing. A society ruled over fools and monsters who were not voted into power and can't be voted out of power is a doomed society. Eventually – the French Revolution, the World Wars – these societies grow so unstable that they collapse altogether.
This is where Piketty and Snowden converge. When the Snowden leaks broke, there was a lot of talk about the mechanics and the legality of the NSA's global digital surveillance, but precious little consideration was given to the reason for all this surveillance. In 2013, the idea that this spying was about "security" was so obvious as to be self-evident. The questions at the time were whether spying could produce security. We weren't asking why things were so insecure.
In retrospect, the answer is to be found in Piketty. Piketty's Capital includes a long, impassioned plea to both lawmakers and aristocrats to consider redistributive policies (like a wealth tax) as the most affordable way to achieve political stability. Fundamentally, Piketty argues that the cheapest way to stop people from building a guillotine on your lawn is to build hospitals and schools; this is cheaper than paying for guards and prisons to lock up would-be guillotine builders.
Today's AI debates swirl around the question of whether AI can truly make us more productive – that is, if chatbots will allow one person to do the work of two, or three, or four – or 100. But when it comes to surveillance, the digital revolution unquestionably produced a massive productivity dividend.
Consider the spying apparatus of the former East Germany ("the GDR") widely considered the most surveilled society in human history. When the Berlin Wall collapsed, there were about 16m people in the country. Of those East Germans, about 90,000 worked directly for the Stasi (the secret police), aided by another 100-200,000 paid informants:
Call it 200,000 people to spy on 16m. In other words, it took one spy to watch 80 of their neighbors. Contrast this with NSA spying: they accumulated detailed surveillance dossiers on about 6 billion internet users using a staff of no more than 5 million spooks (in 2013, about 5 million Americans were eligible for security clearance). If every single person with security clearance in the USA was working on the NSA's surveillance program, that would mean that by 2013, computers had made it possible for a spy to keep tabs on more than a thousand people.
Orders of magnitude improvements in a mere generation! This is the kind of productivity lift that economists dream of when they fantasize about the dividends from automation.
But why? Why spy?
East Germany spied on its people because the system was so unjust and cruel that its beneficiaries understood that their neighbors forever on the brink of rising up against them. East Germany's leaders were right about that – but if anything, they didn't put enough people onto the spying project. We can tell, because the Berlin Wall fell in 1989!
Of course, the GDR was already paying more than 1.2% of its population to spy on everyone else. It's likely that East Germany's leaders believed that their society simply lacked the fiscal space to hire more spies, even if short-staffing the Stasi risked societal collapse. Now, if Piketty is right, East Germany's leaders could have solved this problem by giving people fewer reasons to want to overthrow the state. They could have taken their hands out of the cookie jar, could have instituted democratic reforms – they could have made a bid for democratic legitimacy and public material comfort. But that would have come at the leaders' own power and wealth, and, lacking the stomach for this sacrifice, they lost everything.
Enter the NSA: the digitization of human civilization has drastically reduced the cost of surveillance, and – again, per Piketty – this vastly increases the amount of inequality the world can sustain before the illegitimacy, incompetence and cruelty of rule by the neoaristocratic winners of the orifice lottery brings the whole thing crashing down.
The Trump years are proof of this. We've reached a high-water mark for rule by illegitimate billionaire dilettantes. The second Trump admin began with DOGE's Bonfire of the Stupidities, where Musk cultists dismantled vast swathes of the American administrative state. Musk didn't just attack foreign aid – though the fact that the world's richest man murdered hundreds of thousands of the world's poorest children for the lulz isn't merely cruel, but also massively destabilizing in a way that will shake the world's politics for generations – but also domestic institutions. It was a DOGE cultist who fed the part of the NIH that tracks cyclosporin outbreaks into the wood-chipper:
Today, tens of thousands of Americans are experiencing the literal enshittification of the American state, and this isn't just a human tragedy (though it is), it's also an economic tragedy, with massive knock-on effects for the businesses that rely on those sickened Americans and for the agricultural sector whose outputs are now being shunned by millions. Whether it's letting Bill Gates decide how your schools will work or letting Elon Musk decide how your public health system runs, the result is political chaos and a societal nudge away from the rule of law and towards guillotines.
Which brings me back to Snowden. The Snowden revelations did spur a global conversation about digital surveillance, with the result that the majority of the world's digital traffic is encrypted today. That's not nothing.
But the American state found new ways to conduct mass-scale, global surveillance, often by collaborating directly with tech giants. Billionaires like Peter Thiel capitalized on Big Tech's conflicted feelings about openly participating in surveillance by founding Palantir, with the express mission of murdering the political opponents of oligarchy:
Over the past decade, the steady march of digital technology, dominated by a cartel of giant global firms who collude with the US government's system of political repression in exchange for tax breaks, antitrust forbearance and fat federal contracts has yielded more mass surveillance productivity gains than the previous 25 years:
The Trump administration is the most unpopular in more than a century. Trump has stolen more money in office than any president in history. Trump presides over spiraling greedflation and collapsing buying power. The Trump administration has also presided over a titanic increase in state-aligned, privatized surveillance. The Trump years are the Flock years:
Trump's authoritarianism is a function of his misrule, and his misrule is enabled by his authoritarianism. The more he steals, the more he destroys with wars of choice, and incoherent tariff policies, and official pronouncements linking autism and vaccinations, the more he needs spy cameras, internet surveillance, vehicle tracking, and facial recognition. Every time Trump talks about a third term in office, or canceling elections, or suppressing the vote, he creates demand for mass surveillance to catch and imprison the people this drives into the streets. The more mass surveillance there is, the safer it is for him to commit unpopular, corrupt acts. It's the world's worst self-licking ice-cream cone.
It's not just Trump, of course. Trump is the vanguard of a movement of orifice lottery winners whose delight in stealing, cheating, maiming and despoiling gives rise to political instability and requires them to divert some of their yacht money to mercenaries:
Take AI: the Trump years are also the AI years. This is the time in which a wildly unpopular technology is being shoved into every part of every app we rely on:
It's an era where corporate bosses can't stop gloating about how many jobs they're planning to destroy and how many paycuts they plan on imposing on the surviving workers:
And – most visibly – it's an era in which people's cities and towns are being despoiled by data centers they don't want, by local governments operating in the most extreme secrecy, who silence and even arrest citizens who demand a democratically legitimate process for deciding whether they will have to give up their power and water and land and peace:
An economist would tell you that there's an equilibrium being sought here: between the cost of bribing a town council to ram through data center approvals, the cost of building a more modest and palatable data center, and the cost of mollifying public critics. The cost of bribing towns to foist a data center on the townsfolk is low, because there are lots of towns that fit the bill, so data center barons can shop around.
But as data center protests grow larger and better organized (oligarchy is destabilizing), the cost of dealing with public opposition is mounting. Which is why the Trump administration is teaming up with its preferred tech and military contractors to engage in detailed surveillance of data center and AI critics:
These corporate spooks aren't just spying on data center critics: they've got a whole portfolio of oligarchy-stabilizing surveillance services, targeting "antifa," immigrants' rights and anti-ICE groups.
They're joined by hardware vendors who offer corporations, the wealthy, and enclaves where both are to be found on literal robocops, the ultimate in cheap guard labor (alas, the robots suck):
Trump and his orifice-winning army are caught in the same trap as the leaders of the GDR. Every gain in guard-labor efficiency creates the space for more of them to stick more of their hands even further into the cookie jar. Every time they do, American society grows more unstable, demanding more guard labor.
As we saw in Minneapolis, guard labor – be it mass surveillance, robocops or ICE chuds – is itself destabilizing. Police states make the people who live in them want to overthrow the state, requiring yet more cops, creating more partisans for tearing the whole thing down.
In theory, the orifice class could decide to stop stealing, cheating and maiming. The problem is that for every plute who realizes that the cheapest way to keep the guillotines off his lawn is to play fair, there are three more who lack the executive function to stop cheating. That means that you might as well keep on cheating, since the instability – and the guard labor bills – are coming no matter what.
In the tale of the "Tragedy of the Commons," a common pasture is grazed to dust by shepherds who each understand that if they don't graze their flock until everything is gone, some other shepherd will do so. The original "Tragedy of the Commons" paper was a racist hoax perpetrated by an academic fraud who wanted to make the case for the expulsion of black and Brown people from America and their mass extermination abroad:
But when it comes to the commons that is "a stable society," the orifice class is caught in an inescapable tragedy, certain of the knowledge that if they don't cheat us, the next American aristo will. Thus the demand for guard labor continues to mount…as does the demand for guillotines.
I had merely assumed that the NSA is spying on me because they can. (Obviously not for the purpose of selling me more crap, which is why social media spies on me.)
