Last Week’s New Yorker, week of February 16 & 23

“‘It’s just their wok and busy schedules are very hard to guggling with’“

A week off is a dangerous thing: I let this one turn into ten days off, which is just never a good plan. Live and learn.

Must-Reads:

“The Brood” (A Reporter at Large) - Ava Kofman reduces the child-teacher ratio. A riveting story: bizarre, horrific, but with enough tangible insight into the brokenness of America’s deregulated (and dysregulated) surrogacy system, as well as enough intangible insight into the abusive logic of scammer fascism, to justify its existence and its length. (There are even chapter markers, which mostly serve the purpose of letting the reader know they’re gonna be here a while.) Kofman attempts to replicate the confusion of the surrogate mothers, floating various possibilities as to the suspicious motives of Guojun and Sylvia, even as the fact we’re reading an article about them means there’s no shot it’s anything benign. Still, the reality, which brings the isolated clone world in Jordan Peele’s Us to mind, shocks when it arrives: Children kept in industrialized abuse “classrooms” in a numbers game to accumulate future influence. Of course, fascist pronatalism has a long, well-documented lineage, but this is an especially vivid illustration of the way our ineffectual regulatory environment allows for blatant illiberal cruelty, as long as you have a big enough house. The final twist, which practically conjures the Twilight Zone end-titles theme, is hard to bear. And there’s some yet to be borne.

“The Measure of Things” (Books) - Kathryn Schulz thinks Tennyson is a sea student. A groovy prose poem, wringing out Tennyson’s waterlogged history and chronicling its sluicing shapes. A new book by Richard Holmes proposes that Tennyson was science-obsessed, and gathers some evidence; Schulz’s iambs bring the bass to commonplace material: “Locally, the children were regarded as smart but strange: close to one another but standoffish with outsiders, distinctly bookish but known to run a little wild.” Schulz claims Tennyson as a poet of grief; his “famous phrase”, “Nature, red in tooth and claw,” is “a chilling retort to complacent faith” – and, as Holmes proposes, a fairly scientific one. If Schulz doesn’t totally buy that thesis, it’s only because she knows there’s more at play; what is beyond is both measurable and immeasurable. In awe of Tennyson’s multiplicity, Schulz venerates him, and it’s not only because he writes about hills flowing “‘from form to form’” that she thinks he rocks.

Window-Shop:

“The Landscape Artist” (Onward and Upward With the Arts) - Rebecca Mead lets Andy Goldsworthy know beyond the hill; and on a day they meet to walk the line and set the wall between them once again. Goldsworthy is both a great artist and a literal wise man on the mountain. The British tradition of land art has always had a less Westward Ho vibe than the American equivalent, but Goldsworthy reminds us here that the separation between the ‘human’ and ‘natural’ is a flimsy conceit. (It’s also a useful one, but often the role of the artist is to remind us we can still discard ideas which are efficient, and Goldsworthy is a bit of a god of inefficiency.) Mead makes the very wise decision to mostly get out of the way; this article balances quote density and detail density by loading up on both. She is a little too obsessed with Goldsworthy’s mortality, I think; yes, his art is laborious, but he already relies on expert and assistant hands as much as any good postmodernist. Goldsworthy’s new, grand project (which, despite Mead’s praise for its “thrilling gravity”, doesn’t strike me as having as much verve as some of his less intrusive, more zen efforts, which I’ve always been especially drawn to) is basically a giant memorial to nothing in particular; it’s called Gravestones but its concern with mortality is collective. The logistics behind the project, though – how, exactly, does one get the stones displaced by gravedigging? – are riveting, the highlight. Goldsworthy is exceptionally concerned with the How? of art; Mead shows that this question is closer than many of his contemporaries might think to the Why?

“Toy Story” (Books) - Alexandra Schwartz says toy vey. Schwartz pokes some holes in the thesis that Jewish Americans were entirely responsible for the U.S. toy industry or for “shaping American childhood”, but the book under review still has plenty of stories worth recounting. It was mostly modern, bourgeois-stock Jews that revolutionized child development; at roughly the same time, it was poor Eastern European Jews, struggling for assimilation and a spare dime, who furnished the goods needed to mass-produce childhood novelties. The exact mechanics of this, Kimmel attributes to culture while Schwartz points to economics. Technology is famously hard to analyze without growing deterministic or, on the other hand, dismissive; toys may be lightweight, but they’re still technologies. (Some of them can even piss their diapers!) Schwartz takes a scattershot approach, designed to entertain. You’ll find a novelty or two.

🗣️ “Everybody’s Fly” (Haberdasher Dept.) - David Kamp knows it’s more than a passing Fab. Kamp makes Freddy feel at home, which is the beginning and the end of this assignment.

“Daily Bread” (On and Off the Menu) - Hannah Goldfield waists time. If On and Off the Menu is a riff on On and Off the Avenue, this is the most literal take on that assignment: A series of paragraphs on food diaries, which don’t have their prices listed after their titles, but might as well. It’s objectively funny for this magazine to be telling everyone to listen to This American Life and read the Grub Street Diet¹ though Goldfield has always had more of a New York vibe, to be honest. (Justin Davidson, on the other hand, writes like he’s waiting for the call from Remnick.) This screeches to a halt, but it’s not offensive. There are only so many hours in a day, I guess!

🗣️ “Footwork” (Knockout Dept.) - Brian Siebert knows it don’t mean a thing if he ain’t got that swing. Superb punchline.

(The Mail) - After a few weeks off, thank goodness this section still exists.

Skip Without Guilt:

“Brother Act” (The Current Cinema) - Richard Brody Shadow boxes. Brody learned some new things about Nigerian history from this film, and he’s here to tell us about them. Not the worst thing, but not quite what I want from a review, and Brody is very short on formal analysis. This isn’t a documentary! He mentions “flash-frame collages” that “hint at the drama’s mystical, phantasmagorical essence”, one of which is apparently far too vague. Brody himself is being a bit vague here; I appreciate not being spoiled – Brody responding to criticism, maybe? – but we have to be let in enough that we can understand the critique.

“Cashing Out” (On Television) - Inkoo Kang takes a bank shot. Apparently everyone at the magazine is watching this show now, and hoping we’ll follow suit, even though the new season is less interesting than what’s come before (and even though Kyle Chayka gave it a wittier read last season.) I will get around to it, probably, years late. I’m halfway through season two of Slow Horses, the last show that everyone waited until season three to decide that everyone was watching. I can’t keep up! Don’t you people have periodicals to read?

“I, Claudius” (Annals of Inquiry) - Gideon Lewis-Kraus has his head in the Claudes. This is more entertaining than a lot of what the magazine has put out, but pretty deeply flawed. Mostly, I just don’t think Lewis-Kraus demonstrates a particularly nuanced understanding of the philosophical and linguistic arguments against LLM cognition, so although he demonstrates some skepticism that the many novel ways Anthropic uses the technology can actually add up to cognition, he, very early on, chooses a posture of humility. That’s not a bad idea on its face, except that by humbling himself before researchers with a vested commercial interest in A.I, he’s ceding the argument before it begins. Anthropic’s ridiculousness is fairly evident throughout – at one point, they aim for a particular, supposedly ethical consolation, “‘As an A.I., I do not have direct personal experiences, but I do understand’”, which gives away the game a bit – they’re nodding at Clever Hans², telling him to remember to say he understands, then being shocked when he does. Ontology is incredibly complex – have you tried to read that shit? – and perhaps there is a way of defining consciousness such that a text-chain algorithm can fit it. I’m not sure why that would matter to anyone, though, and I don’t think Lewis-Kraus makes a good case. To anyone but a scholar of consciousness, it seems largely like a sales tactic. If you know enough about meat regulation, you might care about the differences between farm-raised, grass-fed, and free-range. For the average consumer, though, these terms are solely a matter of marketing. The question that interests me more is whether these systems are able to search for meaning beyond the completion of tasks; until there’s evidence of that (and I don’t think there can be) these models shouldn’t be said to understand. More broadly, the industry’s insistence on coming up with their own set of rulers by which to judge their product should not be trusted; how to regulate a product category that has so thoroughly captured the most experienced thinkers in its field is the actual looming question.

Lewis-Kraus, meanwhile, dismisses an accurate if dismissive explanation of LLMs as repeating “‘embedded word chains’” and not using “‘organic associations’” because, well, the view “cannot survive even a cursory interaction with them” – actually, they consider “all the words that might plausibly come after”, which is a way of considering “organic associations” in the sense that every word in the model has a multi-dimensional associative distance from every other word. But this totally ignores the key word in the phrase, organic – the model is built to imitate pre-existing word chains, and although it does so in a way driven by unfathomable amounts of data and dimension, at its root that is still the operation. It’s as though Lewis-Kraus is looking at an iPhone and deciding that it cannot possibly be driven entirely by transistor switches because it’s an iPhone. The linear algebra underlying LLM matrices is not actually analog, and no amount of data can force it to be; inorganic is precisely the right word. The evidence pointing to emergence in these models is scant, as I understand it. I get that Lewis-Kraus wanted to keep things readable, but a more detailed dive into the emergence question would bring a lot more new information to the piece, and keep it from sliding towards mysticism. To presume that there is no argument – that nobody could seriously say that LLM behavior is exclusively a matter of extrapolative logic – is either disingenuous or a matter of misunderstanding.

Do you think it’s funny to hear about the tribulations of these machines? Unfortunately, I don’t, especially, though Lewis-Kraus makes a decent attempt; he tries very hard to cast a quirky vending machine character as central, but it mostly feels like the average of any number of past imaginings of computer subservience. The machine gives us what we want: Science fiction. Chomsky’s universal grammar comes in for critique (boy, he’s having a bad year) though it’s strange not to mention that many of the sharpest early critics of LLMs, who Lewis-Kraus dismisses as curmudgeons, were linguists, who correctly predicted many of the abilities of the newest models, and also showed how even these abilities shouldn’t, in their view, be compared to human grammar use. Eventually, the distance between the philosophy and the sales pitch grows even smaller: Just look how competent the machines are! Sure, it’s harder to find tactics that trip up the newest models. But how much of that is progress, and how much is shoveling more Turks into the mechanism?

“Listening to Joe Rogan” (A Critic at Large) - David Remnick turns on radio. It is very much worth reading the first section of this piece, a delightful mini-personal history about Remnick’s childhood obsession with The Long John Nebel Show. (Remnick’s interest in variety and oddity is a big part of why he’s such an excellent editor for the magazine.) The nostalgia – even an opening Proust reference! – is totally winning. The rest of the piece, unfortunately, is an entirely ordinary reading of Joe Rogan, predictably withering but not in a way that reveals anything unexpected or original about the man, his format, and his predilections. There’s barely even a thesis, beyond pointing out that Rogan matters. Turn off, tune out, don’t drop anything.

“In Search of a Fix” (Medical Dispatch) - Dhruv Khullar gives us the skinny. I’m glad Khullar eventually talks to Sarah Kawasaki, an actual expert in the field of addiction, who is an exceptional scientific communicator. I just wish she’d written the piece instead, since Khullar spends the rest of the piece up until then reinforcing troublesome ideas about addiction and medication. The piece’s anecdotal first section is pretty directly addressed by Kawasaki – addiction is a social disease as much as a neurochemical one; reducing cravings doesn’t inherently build resistance – but her argument against determinism is more glancing, which makes it easy to overlook how much Khullar’s earlier, unquestioning belief in splotches on a screen itself narrows addiction into an inherent circuit. fMRI data is astonishingly interpretive – however interpretive you think it is, it is probably more interpretive than that. The belief in GLP-1s as addiction treatment – not merely as a symptom-abatement tool – has a convenient political effect: It frames addiction as a personal choice; perhaps it’s a slippery-slope fallacy, but my mind jumped to mandatory GLP-1 shots for, say, people arrested for public intoxication. Who profits?

“Monster Mash” (The Art World) - Zachary Fine ponders whether techno or tech yes. I mean, I guess if you go all the way to Germany on the magazine’s dime to see a video, you have to write it up, but it doesn’t seem like Fine had any special insight or connection here. At least he’s interested in the contemporary, which never seemed true of Arn; still, is international travel really required to find compelling work? Fine’s prose continues to bug me; his strained casualness – second-person description (“Your job, you realize, is to grope your way toward that light”), descriptions of Fine’s supposed fear that certainly don’t sound scared (“I couldn’t see my hand in front of my face. I assumed there’d been a lighting malfunction.” No, you didn’t!) – feels like a young adjunct professor trying to be both respected and liked. I was more convinced by some less-convinced writers: Eve Rogers at artnet, Anika Meier at Status Update, Kirsty Bell at e-flux. All pointed out the misogyny and disingenuousness of the press-release language Fine parrots: “a humanlike figure”, “a composite”. Man, face facts!

Letters:

Michael O. is obsessed with Richard Brody’s attempt at simulating the Oscars as though he ran the zoo: “I think we've moved beyond movie criticism and into the Brody-verse. I don't even think the movies he reviews exist or need to exist anymore. He's the Pale Fire of movie critics.”

As a longtime enjoyer of the Skandies, Mike D’Angelo’s statistically rigorous attempt to reverse-engineer a more interesting Oscars from his and his friends’ opinions, I support Brody in his madness.

Anti

gua