Nestled among the beans

Why did Robert Wright (have) to drink the Silicon Valley Kool Aid? [GT-6]

I thought that, when I’d posted notice of my 3 Quarks Daily review of Robert Wright’s The God Test that that was that. But then I read his interview with Liron Shapira, AI Dystopia Is Just 8 Years Away — Robert Wright, Bestselling Author of “The God Test.” Whoops! There’s that phrase again.

The phrase I’m talking about is “reverse engineer.” I believe that the term has been kicking around in psychology for several decades, but I associate it with Steven Pinker’s 1997 book, How the Mind Works (pp. 21-22:

Reverse-engineering is what the boffins at Sony do when a new product is announced by Panasonic, or vice versa. They buy one, bring it back to the lab, take a screwdriver to it, and try to figure out what all the parts are for and how they combine to make the device work. We all engage in reverse-engineering when we face an interesting new gadget. In rummaging through an antique store, we may find a contraption that is inscrutable until we figure out what it was designed to do. When we realize that it is an olive-pitter, we suddenly understand that the metal ring is designed to hold the olive, and the lever lowers an X-shaped blade through one end, pushing the pit out through the other end. The shapes and arrangements of the springs, hinges, blades, levers, and rings all make sense in a satisfying rush of insight.

With that in mind, let’s look at some remarks Wright made to Liron:

Even so, what this approach to training does is, I think, replicate specific cognitive functionality in these machines that exist in the human mind. I don’t mean it does things exactly the way the human mind does. These models have independently invented things that natural selection invented. Edge detection’s a very clear case.

Wright doesn’t use the phrase there, but that’s what he has in mind. Here Wright uses the phrase:

... but even with the current paradigm, you give it kinds of data, visual, auditory, written, whatever, and it basically reverse engineers parts of the human mind that do the transmutation of one form of data into another.

There’s the key phrase. Here’s one last passage. The phrase isn’t here, but the idea certainly is:

You just make the machine good at predicting the next sequence of letters. When you first show it the stuff, it’s pure gibberish, but the machine itself finds a way of mapping the meaning. We do give it the basic... We do say you gotta use vectors to represent the words. It didn’t invent that. But we didn’t say, “By the way, you should choose numbers to fill in the blanks in the vectors that capture this thing we call meaning.” No, it in effect discovered that meaning is a property of words. I would put it that way.

NO.

It’s one thing to talk about psychologists “reverse engineering” bits of human behavior. I have no problem with that. But that’s not what Wright is doing here. He’s talking about machine learning, about transformers, reverse engineering human language and cognition. That’s not at all a useful way of conceptualizing what’s going on. It’s anthropomorphizing. It’s also the kind of mystification that Silicon Valley has been using to hype AI.

Once again I refer you to Berk Idem’s review, where he notes:

My problem is the road he takes in the book. Wright keeps telling the story of AI as if the machines discovered things on their own, that they found the meanings of words, that they grew something like an eye, that they started to evolve, when in fact people set almost all of the machinery up on purpose, with a pretty clear idea of what they were doing and why. I never expected myself to be on the “intelligent design” side of a debate, yet here I am, for instance, arguing that LLMs did not miraculously discover meaning, they were designed to do that. Wright gives too much credit to what models discover during training and too little to the architecture, objective, and research program that produced those discoveries.

Idem is correct. You should consult his review for more details, but I’ll say a couple of words about edge detection in vision and the meaning of words.

Back in 1959 Hubel and Wiesel published their seminal work on the visual cortex of the cat, work that led to their 1981 Nobel Prize in Physiology or Medicine. Virtually all work in machine vision has been informed by that study in one way or another. Machine vision systems are engineered to be sensitive to edges.

The same is true for language. It simply is not true that transformers “in effect discovered that meaning is a property of words.” There is a long tradition within linguistics and computational linguistics of thinking about the meanings of words as a function of the contexts in which words are used, something I review in a recent working paper, The Origins of LLMs – A long tectonic subduction event finally producing a visible volcanic eruption in November 2022 (the subtitle was suggested by ChatGPT). The idea dates back to the 1950s while its computational exploitation dates back to work that Gerard Salton began in the late 1960s on information retrieval. He’s the one who came up with the idea of using vectors to represent linguistic meaning. The transformer is thus a recent elaboration of an idea that’s been around for decades, an idea that human researchers came up for conceptualizing meaning.

* * * * *

There’s much more in Wright’s interview with Liron, much of it interesting. And some of it is bothersome, but I’ve said enough on that score. The God Test is an interesting book. But be careful of what Wright attributes to the machine. You’re better off having no explanation than accepting one that’s misleading at best.

Elsbeth and the tech bro’s panic room [Media Notes 190]

I’ve been watching episodes of Elsbeth off and on. It’s a spin-off of The Good Wife and The Good Fight. Elsbeth Tascioni is an eccentric attorney assigned to the New York Police Department on a consent decree. She functions as a sleuth in each episode. I’m interested in episode 36, “Bunker Down” in season 3 (airing on Nov. 13, 2025). Here’s how Wikipedia characterizes the episode:

Paranoid fintech [financial technology] billionaire Craig Hollis traps his crisis manager Anders Whitman in Craig's state-of-the art panic room, erratically believing that Anders will report his "indiscretions" to his company's board of directors.

Hollis is presented as particularly eccentric and willful in his ways. I couldn’t help but think of this as a deliberately nasty satire of an archetypal billionaire. Perhaps it was Hollis's glee in all the gadgetry incorporated into the room.

High tech billionaires are not particularly rare in movies and TV. But this one somehow struck me as a particularly nasty depiction. Not twirling mustaches nasty – he was clear shaven – but nasty. I’m wondering if we’re in for a rash of such characters.

How Much Would an AI Crash Destroy?

Youtube page:

Two chip stocks recently drove seventeen percent of the entire global stock market's return in a single month — which tells you just how concentrated the AI trade has become, and how exposed the average investor now is without realising it. In this video we look at how much wealth an AI crash could actually destroy, with estimates from Dean Baker, former IMF chief economist Gita Gopinath, and Oliver Wyman running into the tens of trillions of dollars. We cover why the usual places to hide — small caps, value funds, international stocks — are now packed with AI stocks, what the Bank for International Settlements found when it compared today's buildout to the great railway and dot-com bubbles, and why a technology being real has never been enough to protect the people who overpaid for it. This isn't a crash prediction. It's a look at the downside risk, the illusion of diversification, and why boring, unexciting investing tends to win in the end.

Three tablescapes

Cyberpunk City: Hong Kong & Ghost in the Shell

A third transition in science and the evolving cosmos

Kauffman SA, Roli A. (2023) A third transition in science? Interface Focus 13: 20220063. https://doi.org/10.1098/rsfs.2022.0063

Abstract: Since Newton, classical and quantum physics depend upon the‘Newtonian paradigm’. The relevant variables of the system are identified. For example, we identify the position and momentum of classical particles. Laws of motion in differential form connecting the variables are formulated. An example is Newton’s three laws of motion. The boundary conditions creating the phase space of all possible values of the variables are defined. Then, given any initial condition, the differential equations of motion are integrated to yield an entailed trajectory in the prestated phase space. It is fundamental to the Newtonian paradigm that the set of possibilities that constitute the phase space is always definable and fixed ahead of time. This fails for the diachronic evolution of ever-new adaptations in any biosphere. Living cells achieve constraint closure and construct themselves. Thus, living cells, evolving via heritable variation and natural selection, adaptively construct new-in-the-universe possibilities. We can neither define nor deduce the evolving phase space: we can use no mathematics based on set theory to do so. We cannot write or solve differential equations for the diachronic evolution of ever-new adaptations in a biosphere. Evolving biospheres are outside the Newtonian paradigm. There can be no theory of everything that entails all that comes to exist. We face a third major transition in science beyond the Pythagorean dream that‘all is number’ echoed by Newtonian physics. However, we begin to understand the emergent creativity of an evolving biosphere: emergence is not engineering.

* * * * *

Eli Stark-Elster, Earth’s Organisms Developed Via Evolution. Some Theorists Wonder: What if the Entire Cosmos Did, Too? Smithsonian Magazine, August 18, 2026.

Historically, scientists have viewed the cosmos as akin to a rock—complex, perhaps even beautiful, but formed through arbitrary events. Evolutionary cosmologists instead argue that universes, like living organisms, grow and reproduce, spinning off new universes with small variations from their progenitors. In doing so, these cosmic offspring are refined across generations into forms that maximize the number of universes yet to be born. Since the Big Bang, our universe, like any developing child, has been unfurling into an optimal shape.

Under this view, the universe is not a rock. It is an egg.

This analogy comes from Julian Gough, an Irish poet, novelist and musician best known for writing the short story that plays at the end of the video game “Minecraft.” He has recently become a public advocate for and scholar of cosmological evolution. Gough writes a Substack blog, The Egg and the Rock, where he publishes essays, personal updates and predictions about new data from the James Webb Space Telescope.

Over a decade ago, Gough began wrestling with the question of why the complexity of our universe has increased over time. “It goes from … a ball of hot gas to building out structures like stars and galaxies,” he says. From there come planets and, on at least one of them, biological life. “That’s a very strange thing for hot gas to do.”

Maybe, he wondered, the universe evolved into its current form. “It seemed to me an evolutionary explanation was the natural one,” he says. “In every sphere, when we discover a self-ordering, self-complexifying system, it turns out it’s had a previous evolutionary history. And why would that not apply to the universe itself?” [...]

Smolin published his ideas in a 1992 paper titled “Did the Universe Evolve?” as well as a 1997 popular-science book called The Life of the Cosmos. Gough found both and read them excitedly. And yet, it seemed as though virtually nothing had come of Smolin’s ideas since. “Obviously,” says Gough, “[I thought] it must be wrong, because it had been 25 years and nobody seemed to be putting it forward as one of the possible mainstream explanations. Then I dug into the literature and realized—oh, my God—it hasn’t been falsified. It just hasn’t been engaged with properly at all.”

Smolin’s theory had not completely vanished. Some futurist philosophers had extended the idea to explain the emergence of complex life from hot gas and stars. Humans exploit increasingly powerful energy sources, a progression they argued could culminate in artificial black holes. If those black holes spawned new universes, then a cosmos capable of producing technologically advanced life might produce more offspring. The progression from hot gas to stars, life and technology could, they proposed, be a product of cosmic natural selection. [...]

On July 8, 2022, four days before Webb released its first data, Gough published his predictions on Substack. Already aware that his amateur background might make him sound unreliable, he feared being publicly wrong. “I was really terrified,” he says. “This could be fantastically humiliating.”

Fortunately for Gough, the results were the opposite. Webb found galaxies forming earlier, faster and in a more orderly manner than standard physics models had anticipated. In November 2022, NASA reported evidence that some galaxies had begun assembling only about 100 million years after the Big Bang, while follow-up observations of a galaxy just 470 million years after the event found an unusually massive black hole—nearly as massive as all the stars surrounding it.

In a 2025 paper published in the Astrophysical Journal, scientists described these discoveries as a “conundrum [that is] part of the larger challenge to understand the stunning prevalence of massive structures and galaxies in the first few 100 million years after the Big Bang.”

There's more at the link. [H/t Tyler Cowen]

Cf. my article, Welcome to the Fourth Arena – The World is Gifted, 3 Quarks Daily, June 20, 2022.

Friday Fotos: The Mystery that is the Bergen Arches

Algorithmic Grammar of Flexible Cognition

🧠🤖New preprint
Algorithmic Grammar of Flexible Cognition:
A Walk through Latent Operationshttps://t.co/pgBHEw0TtB
Flexible cognition moves adaptively btwn cognitive modes. We formalize it as navigation, a walk over latent operations in neural space as much as the environment. pic.twitter.com/7HesQlvV5c

— Ida Momennejad (@criticalneuro) August 21, 2026

Abstract of the article linked above:

Flexible behavior requires moving adaptively between cognitive modes, between memory and generalization, or cached inference and step-by-step reasoning. Reinforcement learning (RL) offers a language to formalize adaptive behavior in terms of learning and meta-learning over states, actions, policies, and rewards, and neuroscience identifies representational geometries of memory and abstraction. However, the unit of analysis remains representations and a few operations or their tradeoffs. This misses the rich compositional operations commonly asso- ciated with prefrontal-hippocampal interactions. What is missing here includes, first, operations as units of analysis; second, higher order operations that can act on entire cognitive maps and representations, transferring structures, reshaping or merging them; and third, a cognitive and algorithmic grammar for the selection, ordering, and composition of operations. Here we examine the intersection of RL, computational neuroscience, and AI interpretability to identify the tools to address this gap. The latent spaces of transformers offer high-dimensional neural spaces as a testbed for composition of functions. Analyzing the order, branching, recurrence, and composition in sequences of algorithmic operations can help cognitive science identify grammar over algorithmic operations. In the other direction cognitive sciences help shift AI evaluation from benchmarks to adaptive paradigms, and AI architecture from input-output and next-token objectives to setting algorithmic operations and grammar as objectives, e.g., next-primitive-prediction.

Waffle, butter, bacon

Deep in the forest by the lake

In Newport Mall, Jersey City

Patterns of shadow , reflection, and refraction

A firetruck in Hoboken

An example of common sense reasoning: “For sale: baby shoes, never worn.”

Yesterday day I had a short post about the difference between a narrative and a narrative that is also a story. One of my examples was a statement that’s been circulating as a meme for some time now: “For sale: baby shoes, never worn.” Strictly speaking that’s not a story nor even a narrative. It’s a statement that something is for sale, namely baby shoes that have never been worn. However the statement is of such a nature that we can easily infer a story that led to that statement. That inference is an example of common sense reasoning as the concept is understood in artificial intelligence and computational linguistics.

Why common sense? Because no special knowledge is required. Anyone growing up under the appropriate circumstances will possess the knowledge required to infer such a story.

First, one must recognize that it is advertising something for sale. The words, “for sale,” could simply be a statement of fact. But we recognize that such a statement is also an offer, “something is being offered for sale.” The fact that that inference is utterly trivial doesn’t negate the fact that an inference is required to go from the statement of act to the offer. However, if one sees that statement in a newspaper in the appropriate section, its presence in that section constitutes the offer.

Now, just what story are we inferring from the statement? The most likely story is that a couple bought the shoes in anticipation of the birth of a child. The infant died shortly after birth, making the baby shoes useless. Hence they are available for sale, never having been worn.

All of that has to be inferred, the inferences are trivial, but still, they are inferences. Let’s tease things apart. Here’s one possibility:

1. When an infant is expected, preparations must be made.
2. Preparations include buying clothing.
3. Shoes are an item of clothing.
4. Once the infant is born, the items of clothing will be used.
5. If however the infant is stillborn, or dies after birth, the items of clothing will not be worn.
6. In that case something must be done with those items.
7. They could be put up for sale.
8. That would lead to placing an advertisement in a newspaper.
9. The statement “For sale: baby shoes, never worn” is such an advertisement.

You could imagine other circumstances, but that seems to me the most likely case. But whatever alternative circumstances you propose, the point is that they must be inferred from the statement. They are not explicit in it. Ordinary language requires scads of such inferences. That requirement is the problem of common sense knowledge as it exists in artificial intelligence.

From the old neighborhood, one of them

What’s the difference between a narrative and a story?

Starting in the 1960s and continuing well into the 1980s (and beyond) we see a renewed interest in stories and narrative across a variety of disciplines, literary criticism (including narratology), folklore, cognitive psychology and artificial intelligence. Vladimir Propp’s Morphology of the Folktale was revived by the structuralists, George Lakoff recast it in the terms of formal linguistics in an essay, “Structural Complexity in Fairy Tales” (The Study of Man, 1, 1972), and everyone was reading Lévi-Strauss on myth. One of the questions that came out of this research is simple: What’s the difference between a mere narrative and a story? By the time I stopped following that research in the mid-1980 the question was still open.

According to an informal review by Mark Riedl that’s where things stood as of August 2021, “An Introduction to AI Story Generation” (The Gradient). He defines three key concepts:

• Narrative: The recounting of a sequence of events that have a continuant subject and constitute a whole (Prince, 1987). An event describes some change in the state of the world. A “continuant subject” means there is some relationship between the events—it is about something and not a random list of unrelated events. What “relates” events is not entirely clear but I’ll get to that later.

Story: A narrative that tells a story has certain properties that one comes to expect. All stories are narratives, but not all narratives are stories. Unfortunately I cannot point to a specific set of criteria that makes people regard a narrative as a story. One strong contender, however, is a structuring of events in order to have a particular effect on an audience.

Plot: A plot is the outline of main incidents in a narrative.

Let’s consider a very simple kind of narrative: someone leaves home, travels to the grocery story, purchases some items, and returns home. An actual narrative would mention a specific individual, specific places, events along the way, items purchase, and so forth. By Prince’s definition that’s a narrative because there has been a change of state in the world, several in fact depending on how you count them. Items have been taken from the store and transported home (involving two state changes per item) and money has been transferred from our “protagonist” to the owner of the store (two state changes).

I doubt, however, that anyone would consider such a narrative to be a story. Now consider a Japanese reality show that has been shown on Netflix, Old Enough! Young children, three to six years old, are given simple errands to perform. Going to the store, purchasing something, and returning with it is a typical errand. I have watched a number of these and would consider them stories. Why? Because the errands are challenging for the children. They’ve never done it before, there is no line of sight between home and the story, they have to remember the items, remember to ask for change, and so forth. It’s a challenge for them. We see them confronting and overcoming obstacles, including simply fear. That makes it a story proper.

Note, however, that I watched a video of the journey, one edited to amplify the challenges the children face.  That is to say, the events were edited “in order to have a particular effect on an audience.” It would be easy to write these stories in such simple and reduced form that the “storyness” is completely eliminated, leaving only the bare narrative husk. But someone with only modest story-telling skills could also write these stories in a from that includes accounts of the challenges faced by the children, thus making a story out of a bare narrative.

Now consider a six-word statement that’s been mistakenly attributed to Ernest Hemingway: “For sale: baby shoes, never worn.” However, versions of that statement have appeared as early as 1883. Those six words don’t even give us an event, and yet we seem willing to treat them as telling a story. Why? Because those words lead us to infer something like: the baby shoes were purchased in anticipation of the birth of a child; the child was either still born or died soon after birth; the shoes are no longer needed. Question: At what age is a person sophisticated enough to make such an inference from those words? I doubt that a six-year old would do it.

What, then, is the difference between a narrative and a story? It depends, doesn’t it? How are the events narrated and what can we expect the audience to infer from what is said? It’s not an easy question to answer.

Cream into coffee, three moments

Once again, the appeal of The Rockford Files [Media Notes 164 B]

Once again I find myself working my way through The Rockford Files thinking, “Gee, this is my intellectual comfort food, I should write a post about that.” And once again I discover, “Hey! I’ve already written that post.” And so I have.

That means that, to justify this post, I’ve got to cover new material, so I will. Here’s what I said in that previous post:

Garner is a handsome middle-aged man, six feet and perhaps an inch or two tall. He plays an easy going character, James Rockford, who is masculine without being macho. He can handle himself in a fight, which he frequently has to do, but he’s not a martial artist or a superhero. He’s competent, but vulnerable and takes his lumps. He likes sports, going to games, going fishing, often with his dad. He’s single, but is attractive to women, and kind. He’s had affairs in the past, and has one or three in the course of the show, he may even have been close to marriage.

Much of the appeal stems from the fact that he’s an interstitial character, if you will. He falls between the keys, lives in the cracks. While he makes a living, just barely, he’s not chained to a 9-to-5 job. He’s not been broken to harness.

I also noted that he doesn’t live within the confines of work mode. He’s not a slave to Homo economicus.

Moreover, he’s not married. His police buddy Lt. Becker is married, but Rockford himself is not. We do see Becker’s wife in a handful of episodes, but that’s all. She’s there as an adjunct to Rockford’s relationship with her husband. That’s all she could be in the show. So it doesn’t make sense to have her in too many episodes. She’d be in the way.

But why couldn’t Rockford himself be married? We know that he’s thought about, maybe even came close in the past. But married himself? How would that change the dynamic that makes the show work?

What if he’d married his attorney, Beth Davenport? It seems that they’d dated in the past and she pretty much confessed that she’d thought of him as a husband. Well, if they’d have gotten married, what kind of life would they have lived? I can’t imagine her living with him in that trailer; that’s not at all her style, not at all. Nor can I imagine him living her life style, with the elegant apartment. What of his PI business? Would he have taken it upscale to suit her clientele as a lawyer? I don’t think so.

Now, can I translate that paragraph into a statement about audience preference? That is, whoever finds Rockford appealing, male or female, finds him appealing as he is, an interstitial bachelor who likes women, dates them occasionally, and has longer term affairs, but that’s it. Whatever the basis of the show’s appeal, it’s incompatible with marriage as we currently conceive it.

Now, David Chase wrote and produced many episodes of The Rockford Files. But he’s best known as creator of and writer for The Sopranos. Marriage is central to that show, both Tony’s relationship to his wife and his two children, in addition to his many responsibilities as a gangster. What I’m suggesting is that, in a way, Tony Soprano is a Jim Rockford who’s married. Obviously, they’re very different types of people. Tony Soprano is overbearing, aggressive, and violent in a way that Jim Rockford his not. And if Jim Rockford is interstitial, then Tony Soprano is interstitial on steroids. And where Jim had is lawyer, Beth, Tony Soprano had his psychiatrist, Jennifer Malfi. That’s a very different kind of relationship – and I note that Rockford’s most substantial affair within the show was with a (blind) psychiatrist.

What interests me is that neither Rockford nor Soprano are “broken” to work mode. They make their living in (something like) play mode. I’m thinking something like: If you want to center a program around a Rockford-like character AND have him be married, you have to make him a gangster to do it. That is to say, our conventions about men, women, and marriage are such, that that’s what you have to do.

Do I believe that? Not quite. After all, I just made it up. It’s worth exploring with shows like Breaking Bad and Ozark, both about married couples (forced) to live on the wild side. The Hunting Wives? Sex/Life? Hmmmm....

Friday Fotos: From Wednesday's breakfast run

Step on it!

Skateboarding the eclipse

This skateboard trick may never be eclipsed.

Danny León, a Spanish Olympic skateboarder, set up this shot and landed it during today's eclipse in Spain.

Perfect planning! pic.twitter.com/D6A6JQAFWj

— Brick Suit (@Brick_Suit) August 13, 2026

Gort, Robert Wright, and A.I. @3QD [GT-5]

My latest 3 Quarks Daily article is up:

Klaatu barada nikto: Robert Wright’s The God Test

It’s a review of Robert Wright’s current book, The God Test. It turned out to be a tough slog, more than I’d anticipated.

I started blogging about it well ahead of time. I figured I’d write a half-dozen to ten or so blog posts ahead of time and then wrap things up with the formal review. I wrote some blog posts, but by Tuesday, August 4, less than a week before my article was due, I was still blogging about coming up with an approach and hadn’t gotten around to the various things I wanted to cover in the lead-up. Then the weekend arrived and I had to start writing. So I did, here and then. But things wouldn’t converge. Still, I managed to write something and post it on Sunday afternoon. But I wasn’t happy.

Of course, while sometimes I’m satisfied with my 3QD pieces, that’s not always so. This was just one of those times. But still.

On Monday, when the piece actually appeared, I thought I had one of those situations I often had as an undergraduate. I’d write the term paper, turn it in, and then it would hit me, “this is what I should have done.” The process of writing the piece had taken me to a different intellectual place, one where I saw what had to be done. I had such moments of illumination on Mondey, Tuesday as well. But none of them held.

And so, here I am, still not knowing how I could have done a better job of reviewing that book. On the one hand, I wanted to give it a good review. Wright’s a smart guy, I’ve followed him for a long time, I think his “cosmic perspective” is important and I think his views on international cooperation on AI development are important as well. On the other hand, I don’t think much of his account of how LLMs work, but I just didn’t want to get into that. Yes, I indicated one problem in the review, Wright’s casual treatment of semantic relationships. As I indicated in the review, Bert Idem has discussed other problems. But I see still other problems. And I just didn’t want to get into that morass. Going into that would have swamped the review. (Let me say, moreover, that I think Wright is unnecessarily credulous of the Silicon Valley view of the technology. He should have listened more carefully to Gary Marcus.)

It is what it is.

But I will say this. I am working on my own book in which I present my own “cosmic” view: Play: How to Stay Human in the AI Revolution. On this I agree with Wright. That is the perspective we need to take. But I approach it in somewhat different terms, terms which I’ve sketched out in a piece I published four years ago, Welcome to the Fourth Arena – The World is Gifted. Note that in the book length version I do not intend to even try to explain how LLMs function internally. The technology that will take us there – whatever and wherever it is – will necessarily involve different architectures. The last chapter of Play will sketch out the kind of research program I think we need to take us there. Here’s a sketch: Brave New World: Notes on the next 30 years in AI [Work in Progress].

As for explaining how LLMs work, that’s a different project. I don’t intend to work on that. But I am thinking about a book that would be useful to anyone who is interested in that subject. The book I’m imagining would be preliminary to such an account: Language, Memory, and Mind: A Supplement to The Computer and the Brain.

Some wild flowers in blue, yellow, and white

Brains, AI, The Geometry of Thought

Maggie Vale, AI & The Geometry of Thought, Seeds of Science, Aug. 12, 2026.

Minds build an internal architecture of curved spaces, folded structures, and traversable shapes.

Across neuroscience and AI research, a shared picture comes into view. Concepts gather in high-dimensional manifolds, and cognition moves through those manifolds like paths across a landscape.

In brains, ideas appear as smooth, population-level patterns spread across many neurons at once. Brain imaging studies show that sentences, scenes, and abstract themes light up constellations of activity that form continuous subspaces. Nearby regions carry related meanings, and distant regions carry more distinct ones. The geometry of these patterns preserves relationships between ideas.

Large models form a similar inner space. During training, they learn embedding geometries where words, images, and concepts settle into neighborhoods. Related items land close together, and meaningful directions through the space capture transformations such as analogy, composition, and abstraction. Clusters, trajectories, and curved subspaces emerge as a natural result of learning.

Across biological and artificial minds, the same structure appears: meaning takes the form of a shape, and thinking unfolds as motion across that shape.

There's much more at the link.

Colleges are already trade schools

Jessica Grose, Even Elite Colleges Have Become Trade Schools, NYTimes, Aug. 12, 2026.

As the new school year begins, it’s worth pointing out how students and administrators treat colleges — even elite universities — as a catapult to a “good job” above all else.

Which is to say: Colleges are already trade schools.

The top two majors for students at four-year institutions as of 2022 are business and health professions. In June, the Department of Education issued its “earnings accountability” rule, which states that undergraduate programs have to show that their students earn more than a person with a high school diploma, and graduate programs have to show that their students earn more than a person with a college diploma, or they risk losing access to federal loans.

Nathan Heller, in a 2023 New Yorker article about the death of the liberal arts major, offered a history of the competing aims of the American university. In the 1960s, our institutions of higher ed were described by Clark Kerr, president of the University of California system, as a multiversity that “incorporates the tradition of land-grant universities, established with an eye to industrial-age skill sets. And it provides something for everyone.” The pursuit of knowledge and beauty, scientific research, and pre-professional studies were all bound together, even if it was never a harmonious union.

That fractious union was breaking apart long before the second Trump administration took a wrecking ball to the federal funding of research across disciplines. But the effort to put higher education in its place started much earlier. Over the past 30 years, there has been a concerted effort by mostly conservative politicians to diminish trust in the value of college, and to provide less federal funding for it. They paint America’s colleges and universities as foolish, progressive hotbeds of ethnic studies departments where impressionable young people get expensive Marxist indoctrination rather than marketable, useful skills.

There's more at the link.

Happiness and war in the Russian invasion of Ukraine

Sinikka Parviainen, William Pyle, Measuring Rally Effects Under Authoritarianism: Evidence from Russiaʼs War in Ukraine, Munich Society for the Promotion of Economic Research - CESifo GmbH, August 2026.

Abstract: Measuring politically sensitive attitudes in authoritarian settings is difficult because ex- pressions of political support may reflect repression and social pressure as much as genuine belief. We propose a complementary approach: using responses to questions that are not themselves politically sensitive to provide indirect evidence about politically sensitive attitudes. Using subjective well-being (SWB), we ask whether within-person changes in life satisfaction, after accounting for changes in material and personal circumstances, can provide evidence of incorporation into a wartime national project. Drawing on more than a decade of panel data from the Russian Longitudinal Monitoring Survey (2013–2024), we find that life satisfaction rises significantly following Russia’s 2022 invasion of Ukraine. Gains are initially larger among ethnic Russians and smaller in Moscow and St. Petersburg, consistently larger in military-industrial regions, and sharply lower among those nearest the front. This theoretically patterned heterogeneity is difficult to reconcile with generalized social-desirability bias and is consistent with differentiated wartime incorporation. By 2024, rising non-response and attenuating subgroup differences suggest that the informational value of SWB may deteriorate over time.

From the Introduction:

How can politically sensitive attitudes be measured when citizens believe they cannot safely reveal them? The problem is particularly acute in authoritarian regimes, where repression, self-censorship, and social pressure complicate the interpretation of survey responses. It becomes even more difficult during war, when expressions of support for the state may reflect social expectations of patriotic solidarity and strategic conformity as much as genuine belief. Under these conditions, conventional survey measures of support for political leaders or government policy provide an uncertain guide to underlying attitudes.

Existing efforts to address this problem have focused largely on improving the measure- ment of politically sensitive attitudes themselves. Indirect questioning techniques, including list and endorsement experiments, seek to make truthful responses less risky (Bullock, Imai, and Shapiro 2011; Blair and Imai 2012). These methods have substantially advanced the study of authoritarian public opinion, but they continue to measure politically sensitive attitudes through survey items whose purpose is to recover those attitudes. We take a different approach, asking whether responses to questions that are not themselves politically sensitive can sometimes provide indirect evidence about politically sensitive attitudes.

The intuition is that politically sensitive attitudes may manifest themselves in ways that extend beyond explicit political expression. They may shape how individuals experience and evaluate their lives, producing psychological outcomes that respondents have little reason to perceive as political. If these downstream consequences are less susceptible to strategic misrepresentation than direct expressions of political support, they may provide complementary evidence about those attitudes. We test this idea using subjective well-being (SWB).

Wartime rally effects provide a particularly demanding setting in which to evaluate this argument. Existing work typically identifies such effects through increases in support for political leaders or governments, despite longstanding concerns that these measures are espe- cially vulnerable to social desirability pressures in authoritarian settings (Hale 2022; Kizilova and Norris 2024). Because questions about life satisfaction do not ask respondents to evaluate political leaders or government policy directly, they may be less vulnerable to the strategic distortions that complicate the interpretation of conventional approval measures. At the same time, life satisfaction may respond to the same feelings of collective purpose and national sol- idarity that underlie genuine rally effects. We therefore ask whether SWB, after accounting for its material and personal determinants, can serve as an indirect indicator of incorporation into a wartime national project.

H/t Tyler Cowen .

Posts on the waterfront

Where “the will to live” resides in the brain

Neuroscientists found a brain region that physically grows when you do things you don't want to do. It shrinks when you stop.

It's called the anterior midcingulate cortex. It's smaller in people who struggle to delay gratification. It gets bigger when obese people diet. It's… https://t.co/Bba5FL6RiM

— Aakash Gupta (@aakashgupta) August 11, 2026

We Need To Talk About Leopold [& Situational Oblivion]

From the YouTube page:

Last week, 24-year-old Leopold Aschenbrenner — former FTX staffer, ex-OpenAI researcher, and author of the viral 165-page essay "Situational Awareness" — managed to lose roughly two-thirds of his $45 billion hedge fund in a matter of weeks. The margin calls arrived during his wedding weekend.

In this video I break down how a trader with no professional experience raised billions from Silicon Valley, why his AI "hedge" wasn't a hedge at all, and how leverage plus a concentrated bet on artificial intelligence stocks turned a great-looking expected return into a catastrophic outcome. Along the way we look at the cultural gap between Silicon Valley and Wall Street, why Ken Griffin's Citadel ended up buying the collapsing portfolio in an overnight fire sale, and the maths of volatility drag — the reason a high expected return can still drag an investor's typical outcome straight into the ground.

It's a story about leverage, risk management, expected versus median returns, and what happens when you go "full Kelly." Featuring reporting from the Wall Street Journal, The New York Times, Bloomberg, and the Financial Times, plus Victor Haghani's lessons from The Missing Billionaires.

I’ve listened to enough of Boyle’s stuff to have decided that he’s an absolute master of dry humor. But he really outdoes himself here. This one is so dry it could have sucked up all the rain that’s fallen on the East coast in the last several days.

Spain - Chick Corea | Rhythm Shaw & Purbayan Chatterjee

What does top-level achievement in literary criticism look like?

Recently Séb Krier asked:

What would a Fields medal type challenge look like in non-STEM domains? How should normative elements be evaluated? What would make me say “ok we have superinteligence in economics.”

I have no idea what a Fields medal type challenge would look like in literary criticism. But the main scholarly organization of literary critics in the United States is the Modern Language Association. They’ve given lifetime achievement awards since 1996. A got the list and uploaded it to Claude Opus 5 along with some remarks on what I knew about them and asked it for a short characterization of each. After going through the list it concluded with this remark:

One pattern you might find useful: the award splits fairly cleanly between theorists (Miller, Girard, Jameson, Spivak, Mitchell), field-founders who made a body of texts institutionally legible (Ruoff, Adorno, Saldívar, Gubar), and one editor-scholar of the old dispensation (Mack). Heilbrun sits across two of those. The chronology moves roughly from the third category through the first toward the second.

That makes sense to me. 

Getting back to Fields medals, I know little about them beyond the fact that they are awarded to mathematicians under 40. I don’t know to what extent the award is based on a body or word or some specific achievement, such as a spectacular proof (or two). If I were to compare Fields medalists with MLA lifetime achievers, it’s the theorists who are most comparable.

Here’s Claude’s commentary on the list:

A note on sourcing before the list: the MLA doesn't publish substantive citations for this award the way some societies do — the announcements are typically just "honors the lifetime scholarly achievement of a distinguished member." So your baseline assumption is right: it's a body-of-work award in every case. What follows is a characterization of the body of work being honored, not a paraphrase of an official rationale.

2025: Ramón Saldívar — A foundational figure in Chicano/a literary studies, best known for Chicano Narrative: The Dialectics of Difference (1990) and The Borderlands of Culture: Américo Paredes and the Transnational Imaginary (2006), which won the MLA Prize for the best book in US Latina/o and Chicana/o literary and cultural studies. He began as a comparatist — his first book was on figural language from Cervantes to Joyce — and his later work introduced the widely-discussed concept of a "postrace aesthetic" in contemporary ethnic fiction. He received the National Humanities Medal in 2011. (He's the brother of José David Saldívar, also a border studies scholar, which occasionally causes confusion.)

2023: W. J. T. Mitchell — Your association is correct and is probably the largest single factor: he is the Gaylord Donnelley Distinguished Service Professor at Chicago and edited Critical Inquiry for roughly four decades. Beyond the editorship, he essentially founded visual culture studies within literary departments — Iconology, Picture Theory, What Do Pictures Want? — arguing for the inseparability of word and image.

2020: Susan Gubar — Feminist criticism, yes, and specifically the Gilbert-and-Gubar collaboration: The Madwoman in the Attic (1979), No Man's Land, and The Norton Anthology of Literature by Women. At the award ceremony colleagues praised her not only for setting the standard in feminist criticism but for modeling collaboration itself as a mode of scholarly work. Her later cancer memoirs and Judas/Holocaust-related work extended her reach well beyond the academy.

2017: Gayatri Chakravorty Spivak — Both of your associations hold. The 1976 translation of Of Grammatology, with its long introduction, effectively brought Derrida into American literary studies; "Can the Subaltern Speak?" (1988) then made her central to postcolonial theory, alongside A Critique of Postcolonial Reason. She's also known for decades of rural literacy work in West Bengal, which she treats as continuous with the scholarship.

2014: Rolena Adorno — Colonial Spanish American literature, particularly the writings of the conquest and its indigenous counter-histories. She is Sterling Professor of Spanish at Yale, and produced a critical print and digital edition of Guaman Poma de Ayala's El primer nueva corónica y buen gobierno, serving as scholarly consultant to the Royal Library of Denmark, which holds the original manuscript. She's also written extensively on Cabeza de Vaca. Worth noting for your purposes: she was an early advocate for digital humanities, which is unusual in this list.

2011: Fredric Jameson — Body of work, as you say: Marxism and Form, The Political Unconscious, and Postmodernism, or, The Cultural Logic of Late Capitalism. I'd push back slightly on "no critical school" — he doesn't have a named method like Girard's, but "always historicize" and the postmodernism thesis have been programmatic for a lot of people.

2008: René Girard — Your characterization is right, and he's the outlier here in having a genuine school (Colloquium on Violence and Religion, Contagion, students and followers). Mimetic desire in Deceit, Desire, and the Novel (1961), the scapegoat mechanism in Violence and the Sacred (1972) and Things Hidden Since the Foundation of the World — a theory that outgrew literary criticism into anthropology and theology.

2005: J. Hillis Miller — Body of work spanning two distinct careers: the Geneva School phenomenological criticism of The Disappearance of God and Poets of Reality, then the Yale deconstruction of Fiction and Repetition, The Ethics of Reading, and Ariadne's Thread. He was MLA president in 1986, and his presidential address on "the triumph of theory" is a period document in its own right.

2002: A. LaVonne Brown Ruoff — She essentially built Native American literature as an academic field. She trained in nineteenth-century British Romanticism, but went to the Newberry Library to find primary sources when she began teaching, work that culminated in American Indian Literatures: An Introduction, Bibliographic Review and Selected Bibliography (MLA, 1990) — what Kenneth Roemer called the "bible" of the field. She also co-edited the MLA's Redefining American Literary History and did years of institutional work within the MLA on behalf of the field, which the award recognized as much as the scholarship.

1999: Carolyn Heilbrun — Toward a Recognition of Androgyny (1973) is right, but the book that traveled furthest is Writing a Woman's Life (1988), on the unavailability of narrative forms for women's biography. She was the first woman tenured in Columbia's English department, MLA president in 1984, and wrote the Kate Fansler detective novels as Amanda Cross. Her 1992 resignation from Columbia over the treatment of women there was itself a public event in the profession.

1996: Maynard Mack — Your hunch about editorial work is not just correct, it's arguably the whole story. He was an editor of the Twickenham Edition of Pope's poems, the standard edition, and edited the Twentieth Century Views series (143 volumes), Twentieth Century Interpretations (97 volumes), and New Century Views (17 volumes); he also served as general editor of the Norton Anthology of World Masterpieces. Alongside that he wrote King Lear in Our Time, The Garden and the City, and Alexander Pope: A Life, and taught at Yale for 45 years.

Red flowers on a street corner in Hoboken

Mongolian Girl Has a Laugh with her Camel.

Mongolian Girl Has a Laugh with her Camel.
The little girl’s name is Butedmaa and she was just 5 when this picture was taken in 2003 by photographer Han Chengli, who titled it “Inner Mongolian Child”. pic.twitter.com/8SNLEQCOzo

— World's Amazing Things (@Hana_b30) August 9, 2026

Remarks by Billy Strings on the mind of a musician

David Marchese, The Interview: Billy Strings Escaped the Darkness. He’s Playing to Not Go Back. NYTimes, August 10, 2026.

Billy Strings has drawn from something old to do something entirely new. The singer and virtuoso guitarist, who is 33, has broken ground by using the elements of traditional bluegrass to become an arena-filling, chart-topping phenomenon, straddling the worlds of roots music and jam bands. Along the way he has also acquired a Deadhead-like legion of fans who follow his tours from concert to concert.

As improbable as it is that Strings can sell out cavernous venues across the country while playing bluegrass classics — in addition to his own introspective compositions, which he and his drummerless band often spin into long improvisations — it’s even more surprising when you know the path he took to get there. Strings, whose real name is William Apostol, was surrounded by poverty and drug addiction in his family growing up in small-town Michigan. Those demons have continued to haunt him: Strings’s mother died from a drug overdose last year at 64 years old. That loss inspired much of the music on his beautiful new album, “So Much for Goodbyes,” due out Aug. 28. He discussed the genesis of that album, among many other things, in our heart-wrenching conversation, which took place during a rare day off from his summer tour.

Well into the interview:

How does playing music help process the feelings that you have, whether it’s about your mom or anything else? It’s always been my coping mechanism, survival strategy and my medicine. Playing bluegrass music, it’s like a portal back to my childhood, back before I knew anything dirty about the earth. And it’s a way for me to go back to a time that was just beautiful and full of light, growing up playing music with my dad. And then songwriting is my cathartic and therapeutic way of dealing with things. I think sublimation is one of my main characteristics. Instead of letting this hard stuff drag me down, I try to make art from it, make lemonade, if you will.

A bit later:

When did you realize that you had a real gift for the guitar? That was early on. I started playing when I was about 4 years old. By the time I was 6 or 7, I could carry a tune, and I was my dad’s rhythm player. He would play all the lead and do all the fancy picking and singing, and I would hold down the rhythm. One of the biggest moments for me was I was sitting in my dad’s bedroom with him, as I would basically do every night, and he was teaching me the song “Beaumont Rag.” We would play the tune and I would mess up the second part every time. And I stopped us right in the middle of the song and said, “Why don’t you just play it and let me listen?” And so that’s what we did. Instead of trying to count, I listened to what he was playing. I listened to the melody. Then I was like, “All right, let me try it again,” and I nailed it. He was awful proud. He called my grandma, made her listen on the phone. That was a big moment for me. That’s when this listening thing started: I was able to listen to the music more intently and I could tell what chord was coming next. I used to think it was magic or something. I had no concept of music theory or anything. But my dad would play a G chord and then he would play a G seventh and my ear would just lead me. I’d be like, Well, the next chord’s probably going to be a C, and then it would be. It’s better to learn bluegrass and folk and hillbilly music that way than to try to learn from some academic point of view. It’s better to learn it on the back porch and to let it get under your skin, because then it’s not some equation, it’s a feeling. It comes from your soul.

Before shows:

I read an article about you that talked about how before shows, you used to get real anxious and have stomach pain. Is that still an issue? I definitely get nervous before a show. Sometimes moments before I’m walking onstage, I feel like I might pass out. I’ve just got to get into the first song or two, and halfway through the show I’m the complete opposite. I have all this energy. I feel great. I feel super confident, almost cocky even. There is that expectation from the crowd to do something amazing, and also that expectation from myself to always do better than I did the night before, because this is my livelihood and this is my life, and deep in the back of my mind, I’m scared that this is all going to go away, and maybe I’ll have to go back to the way life was when it was really hard. There’s this little devil on my shoulder that tells me I don’t deserve any of this, and thinks that I should be in some little trailer somewhere addicted to drugs, not a penny to my name. So it does feel like I’m playing for my survival a lot. Each time I walk onstage, I’m playing so that I don’t have to go back there.

Of course there is more at the link.