609 stories
·
2 followers

Finding Peter Putnam

1 Comment

The neighborhood was quiet. There was a chill in the air. The scent of Spanish moss hung from the cypress trees. Plumes of white smoke rose from the burning cane fields and stretched across the skies of Terrebonne Parish. The man swung a long leg over a bicycle frame and pedaled off down the street.

Nautilus Members enjoy an ad-free experience. Log in or Join now .

It was 1987 in Houma, Louisiana, and he was headed to the Department of Transportation, where he was working the night shift, sweeping floors and cleaning toilets. He was just picking up speed when a car came barreling toward him with a drunken swerve. 

A screech shot down the corridor of East Main Street, echoed through the vacant lots, and rang out over the Bayou.

Then silence. 

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

The 60-year-old man lying on the street, as far as anyone knew, was just a janitor hit by a drunk driver. There was no mention of it on the local news, no obituary in the morning paper. His name might have been Anonymous. But it wasn’t. 

How could this genius just vanish into obscurity?

His name was Peter Putnam. He was a physicist who’d hung out with Albert Einstein, John Archibald Wheeler, and Niels Bohr, and two blocks from the crash, in his run-down apartment, where his partner, Claude, was startled by a screech, were thousands of typed pages containing a groundbreaking new theory of the mind.

“Only two or three times in my life have I met thinkers with insights so far reaching, a breadth of vision so great, and a mind so keen as Putnam’s,” Wheeler said in 1991. And Wheeler, who coined the terms “black hole” and “wormhole,” had worked alongside some of the greatest minds in science.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Robert Works Fuller, a physicist and former president of Oberlin College, who worked closely with Putnam in the 1960s, told me in 2012, “Putnam really should be regarded as one of the great philosophers of the 20th century. Yet he’s completely unknown.”

That word—unknown—it came to haunt me as I spent the next 12 years trying to find out why.

The American Philosophical Society Library in Philadelphia, with its marbled floors and chandeliered ceilings, is home to millions of rare books and manuscripts, including John Wheeler’s notebooks. I was there in 2012, fresh off writing a physics book that had left me with nagging questions about the strange relationship between observer and observed. Physics seemed to suggest that observers play some role in the nature of reality, yet who or what an observer is remained a stubborn mystery.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Wheeler, who made key contributions to nuclear physics, general relativity, and quantum gravity, had thought more about the observer’s role in the universe than anyone—if there was a clue to that mystery anywhere, I was convinced it was somewhere in his papers. That’s when I turned over a mylar overhead, the kind people used to lay on projectors, with the titles of two talks, as if given back-to-back at the same unnamed event:

Wheeler: From Reality to Consciousness  

Putnam: From Consciousness to Reality

Putnam, it seemed, had been one of Wheeler’s students, whose opinion Wheeler held in exceptionally high regard. That was odd, because Wheeler’s students were known for becoming physics superstars, earning fame, prestige, and Nobel Prizes: Richard Feynman, Hugh Everett, and Kip Thorne.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Back home, a Google search yielded images of a very muscly, very orange man wearing a very small speedo. This, it turned out, was the wrong Peter Putnam. Eventually, I stumbled on a 1991 article in the Princeton Alumni Weekly newsletter called “Brilliant Enigma.” “Except for the barest outline,” the article read, “Putnam’s life is ‘veiled,’ in the words of Putnam’s lifelong friend and mentor, John Archibald Wheeler.”

THE SCHOLAR: Peter Putnam catches up with the news while studying at the University of Leiden in 1956. Putnam spent a semester in the Netherlands working as an assistant to physicist John Archibald Wheeler. Credit: John Archibald Wheeler, courtesy of Alison Lahnston.

A quick search of old newspaper archives turned up an intriguing article from the Associated Press, published six years after Putnam’s death. “Peter Putnam lived in a remote bayou town in Louisiana, worked as a night watchman on a swing bridge [and] wrote philosophical essays,” the article said. “He also tripled the family fortune to about $40 million by investing successfully in risky stock ventures.”

The questions kept piling up. Forty million dollars? 

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

I searched a while longer for any more information but came up empty-handed. But I couldn’t forget about Peter Putnam. His name played like a song stuck in my head. I decided to track down anyone who might have known him.

The only paper Putnam ever published was co-authored with Robert Fuller, so I flew from my home in Cambridge, Massachusetts, to Berkeley, California, to meet him. Fuller was nearing 80 years old but had an imposing presence and a booming voice. He sat across from me in his sun-drenched living room, seeming thrilled to talk about Putnam yet plagued by some palpable regret.

Putnam had developed a theory of the brain that “ranged over the whole of philosophy, from ethics to methodology to mathematical foundations to metaphysics,” Fuller told me.  He compared Putnam’s work to Alan Turing’s and Kurt Gödel’s. “Turing, Gödel, and Putnam—they’re three peas in a pod,” Fuller said. “But one of them isn’t recognized.”

Fuller led me to Barry Spinello, a filmmaker in Bakersfield, California, who met Putnam at the Apollo Theater in Harlem in 1963. Cannonball Adderley was wailing on the sax. “I turned around and saw this guy doing a ridiculous dance,” Spinello said—half jive, half seizure. Putnam was tall and thin like an overgrown twig, flailing like he might tie himself into a knot. They got to talking, and Spinello found Putnam so fascinating that, 10 years later, he traveled to Louisiana to record a week’s worth of conversations about his work.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

For nights on end, on a cot in Spinello’s studio, I slept to the clicks and hums of an achy reel-to-reel machine, as it fed 35 hours of audio into my digital recorder. I listened to one of the recordings in my headphones on the plane ride back, flinching when Putnam’s voice broke through the static. Suddenly he became a real person, a person with vocal cords. He sounded like Jimmy Stewart with a stutter. “Sometimes I think it would’ve worked with Wheeler,” he was saying, “but it just …” Then he went silent. 

Spinello gave me an email address for Coleman Clarke, who had met Putnam in New York City in the 1960s while Clarke was doing his Ph.D. at Columbia University. “It’s mind blowing to me that you found Putnam in Wheeler’s journals,” Clarke wrote in reply, as if I’d won a scavenger hunt that everyone else had quit playing decades ago.

Illustration by Cornelia Li

Clarke seemed relieved that someone had finally come around asking about Putnam, eager to tell me about this extraordinary man who had slipped through the cracks of history. “He was a genius,” Clarke said. “Every talk with him had this level of significance that was just orders of magnitude higher up than a normal conversation with a normal human being.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

One person led to another. Gary Aston-Jones, head of the Brain Health Institute at Rutgers University, told me he was inspired by Putnam to go into neuroscience after Clarke gave him one of Putnam’s papers.

“Putnam’s nervous system model presaged by decades stuff that’s very cutting edge in neuroscience,” Aston-Jones said, and yet, “in the field of neuroscience, I don’t know anybody that’s ever heard of him.”

Phillips Jones, a physicist who worked alongside Putnam in the early 1960s, told me over the phone, “We got the sense that what Einstein’s general theory was for physics, Peter’s model would be for the mind.”

Even Einstein himself was impressed with Putnam. At 19 years old, Putnam went to Einstein’s house to talk with him about Arthur Stanley Eddington, the British astrophysicist. (Eddington performed the key experiment that proved Einstein’s theory of gravity.) Putnam was obsessed with an allegory by Eddington about a fisherman and wanted to ask Einstein about it. Putnam also wanted Einstein to give a speech promoting world government to a political group he’d organized. Einstein—who was asked by plenty of people to do plenty of things—thought highly enough of Putnam to agree.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

How could this genius, this Einstein of the mind, just vanish into obscurity? When I asked why, if Putnam was so important, no one has ever heard of him, everyone gave me the same answer: because he didn’t publish his work, and even if he had, no one would have understood it. 

“He spoke and wrote in ‘Putnamese,’ ” Fuller said. “If you can find his papers, I think you’ll immediately see what I mean.”

In a January freeze in 2013, I headed to Rochester, New York, to meet Clarke. He was in his late 70s but looked younger—tall and slim with gray hair and a bounce in his step. “I’m just so excited that you’ve found Putnam,” he said warmly. It sounded like, what took you so long? He told me he had some of Putnam’s papers in storage.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

We drove to a nondescript brick building. I followed him down a cold, white hallway until he stopped in front of one of the units. He turned the key and lifted the massive door.  

You have to understand what I was expecting. I thought we were going to his storage unit, that it would be filled with whatever storage units are filled with—old clothes and rusty bikes, mismatched chairs and unused exercise equipment. Then somewhere, beneath a pile of something, in a dusty cardboard box, a few of Putnam’s papers.

That’s not what this was.

As the door rolled up, I caught a glimpse of what lay behind.  

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

“This is all Putnam?” I whispered. 

Clarke nodded. “I’ve never shown anyone before you.” 

Putnam’s remarkable claim was that simply by playing this game, the system will learn.

There were no old clothes. No mismatched chairs. Only filing cabinets. Rows of filing cabinets, all neatly labeled, giving the whole place the appearance of a professional archive. I looked around, stunned. It was the entire library of Putnam’s unpublished writings. His theory, his life. The whole long-lost thing. 

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

When Clarke first heard that Putnam had been killed, he made frantic phone calls to the Putnam family lawyer to find out what was happening with Peter’s papers. The answer seemed to be nothing—they’d been moved from Houma to a Cleveland warehouse and might have been thrown away. Clarke rushed to Ohio, loaded them onto a truck, and drove them to his home in Utah for safe keeping; when he moved to Rochester, the archive moved with him.

Now he dug through drawers, handing me papers and folders until I was holding a stack so large I nearly toppled over. Typed manuscripts at hundreds of pages apiece; binders full of notes and letters; handwritten journals; accordion folders bursting with photos, telegrams, and postcards—we piled as much as we could into the trunk of my rental car and I drove back to my hotel. 

It’s one thing to read through curated papers at a place like the American Philosophical Society, with pages gingerly propped on foam wedges under the watchful eyes of librarians. It’s another to flop down on a white bedspread in a Courtyard Marriott and hold a man’s unprocessed life, alone. You turn it over in your hands, still covered in his pencil marks, smudged with his fingerprints; an envelope singed in the spiral shape of his stove ring, yellowed glue clutching his pet bird’s tattered feather, a letter torn apart seemingly in anger and taped back together in remorse. Suddenly you’re implicated. You’ve disturbed a sleeping thing.

Skimming through the papers I saw that the people I’d spoken to hadn’t been kidding about the Putnamese. “To bring the felt under mathematical categories involves building a type of mathematical framework within which latent colliding heuristics can be exhibited as of a common goal function,” I read, before dropping the paper with a sigh. Each one went on like that for hundreds of pages at a time, on none of which did he apparently bother to stop and explain what the whole thing was really about.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

There was no way I could read it all in a reasonable amount of time, so I spent the next week driving between my hotel and the storage unit. I’d stay up all night, photographing the items page by page, then head back to the storage unit, bleary-eyed in the daylight, to swap it all out for a new batch. I’d already photographed some 10,000 pages of material when Clarke grinned and confessed, “There’s a second storage unit.”

PARTNERS: John “Claude” DeBrew and Peter Putnam in Houma, Louisiana, circa 1978. Putnam met Debrew, an ex-Army major, in a Harlem jazz club. They lived together for 17 years before Putnam died. Courtesy of Coleman Clarke.

Back in my apartment in Cambridge, I began sorting through everything I’d found. Photographs, letters, transcripts, papers—I spread them on the kitchen table like pieces to a jigsaw puzzle. Gradually, Putnam’s life and the scope of his theory came into view.

He developed it over the course of three decades, starting as a teenager in the 1940s. He wrote constantly—in the Navy, when he was sent to the brig for reading poetry on duty; while earning his degrees at Princeton University, and after, teaching physics at the University of Massachusetts at Amherst in the 1950s and Columbia University in New York in the 1960s. He wrote while he was living in Fuller’s office at Barnard College with little more than a cot, a phonograph, and a hot plate, and when he moved into a basement apartment in upper Manhattan, just west of Harlem, his lanky form hunched over a typewriter between the grated window bars.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Putnam spent most of his time alone, Fuller had told me. “Because of this isolation, he developed a way of expressing himself in which he uses words, phrases, concepts, in weird ways, peculiar to himself. The thing would be totally incomprehensible to anyone.”

I took the incomprehensibility as a test. I didn’t know why I was being tested. I only knew I wanted to pass. I was driven in part by the looks on everyone’s faces, a pain that appeared fresh despite the years. “My basic upset is, I feel somehow I failed to get his stuff out there,” Fuller said. Wheeler had felt the same. “I realize I didn’t do my duty by Peter,” he said after Putnam died.

Their regret was now my inheritance, a whisper that grew louder as the years pressed on. I might have walked away if I hadn’t been struck with the same feeling that had taken hold of everyone else: that Putnam was actually onto something. That he was quite possibly a genius. In the beginning, I was chasing Peter Putnam the fantasy, a forgotten janitor who’d discovered the structure of the mind. But the deeper I read, I found myself thinking, Wait, did a forgotten janitor seriously discover the structure of the mind?

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Imagine a fisherman who’s exploring the life of the ocean. He casts his net into the water, scoops up a bunch of fish, inspects his catch and shouts, “A-ha! I have made two great scientific discoveries. First, there are no fish smaller than two inches. Second, all fish have gills.”

The fisherman’s first “discovery” is clearly an error. It’s not that there are no fish smaller than two inches, it’s that the holes in his net are two inches in diameter. But the second discovery seems to be genuine—a fact about the fish, not the net.

This was the Eddington allegory that obsessed Putnam.

When physicists study the world, how can they tell which of their findings are features of the world and which are features of their net? How do we, as observers, disentangle the subjective aspects of our minds from the objective facts of the universe? Eddington suspected that one couldn’t know anything about the fish until one knew the structure of the net.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

That’s what Putnam set out to do: come up with a description of the net, a model of “the structure of thought,” as he put it in a 1948 diary entry.

At the time, scientists were abuzz with a new way of thinking about thinking. Alan Turing had worked out an abstract model of computation, which quickly led not only to the invention of physical computers but also to the idea that perhaps the brain, too, was a kind of Turing machine.

Putnam disagreed. “Man is a species of computer of fundamentally different genus than those she builds,” he wrote. It was a radical claim (not only for the mixed genders): He wasn’t saying that the mind isn’t a computer, he was saying it was an entirely different kind of computer.

A universal Turing machine is a powerful thing, capable of computing anything that can be computed by an algorithm. But Putnam saw that it had its limitations. A Turing machine, by design, performs deductive logic—logic where the answers to a problem are contained in its premises, where the rules of inference are pregiven, and information is never created, only shuffled around. Induction, on the other hand, is the process by which we come up with the premises and rules in the first place. “Could there be some indirect way to model or orient the induction process, as we do deductions?” Putnam asked.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Whenever Putnam made a new friend, his mother warned him, “They’re probably using you for your money.”


Putnam laid out the dynamics of what he called a universal “general purpose heuristic”—which we might call an “induction machine,” or more to the point, a mind—borrowing from the mathematics of game theory, which was thick in the air at Princeton. His induction “game” was simple enough. He imagined a system (immersed in an environment) that could make one mutually exclusive “move” at a time. The system is composed of a massive number of units, each of which can switch between one of two states. They all act in parallel, switching, say, “on” and “off” in response to one another. Putnam imagined that these binary units could condition one another’s behavior, so if one caused another to turn on (or off) in the past, it would become more likely to do so in the future. To play the game, the rule is this: The first chain of binary units, linked together by conditioned reflexes, to form a self-reinforcing loop emits a move on behalf of the system.

Every game needs a goal. In a Turing machine, goals are imposed from the outside. For true induction, the process itself should create its own goals. And there was a key constraint: Putnam realized that the dynamics he had in mind would only work mathematically if the system had just one goal governing all its behavior.

That’s when it hit him: The goal is to repeat. Repetition isn’t a goal that has to be programmed in from the outside; it’s baked into the very nature of things—to exist from one moment to the next is to repeat your existence. “This goal function,” Putnam wrote, “appears pre-encoded in the nature of being itself.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

So, here’s the game. The system starts out in a random mix of “on” and “off” states. Its goal is to repeat that state—to stay the same. But in each turn, a perturbation from the environment moves through the system, flipping states, and the system has to emit the right sequence of moves (by forming the right self-reinforcing loops) to alter the environment in such a way that it will perturb the system back to its original state.

Putnam’s remarkable claim was that simply by playing this game, the system will learn; its sequences of moves will become increasingly less random. It will create rules for how to behave in a given situation, then automatically root out logical contradictions among those rules, resolving them into better ones. And here’s the weird thing: It’s a game that can never be won. The system never exactly repeats. But in trying to, it does something better. It adapts. It innovates. It performs induction.

HAPPY DAYS: Peter Putnam in his Navy uniform shares a laugh with his mother, Mildred Andrews Putnam, at their home on Lake Shore Boulevard in the affluent village of Bratenahl, Ohio, circa 1944. Courtesy of Coleman Clarke.

In paper after paper, Putnam attempted to show how his induction game plays out in the human brain, with motor behaviors serving as the mutually exclusive “moves” and neurons as the parallel binary units that link up into loops to move the body. The point wasn’t to give a realistic picture of how a messy, anatomical brain works any more than an abstract Turing machine describes the workings of an iMac. It was not a biochemical description, but a logical one—a “brain calculus,” Putnam called it.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

As the game is played, perturbations from outside—photons hitting the retina, hunger signals rising from the gut—require the brain to emit the right sequence of movements to return to its prior state. At first it has no idea what to do—each disturbance is a neural impulse moving through the brain in search of a pathway out, and it will take the first loop it can find. That’s why a newborn’s movements start out as random thrashes. But when those movements don’t satisfy the goal, the disturbance builds and spreads through the brain, feeling for new pathways, trying loop after loop, thrash after thrash, until it hits on one that does the trick. 

When a successful move, discovered by sheer accident, quiets a perturbation, it gets wired into the brain as a behavioral rule. Once formed, applying the rule is a matter of deduction: The brain outputs the right move without having to try all the wrong ones first.

But the real magic happens when a contradiction arises, when two previously successful rules, called up in parallel, compete to move the body in mutually exclusive ways. A hungry baby, needing to find its mother’s breast, simultaneously fires up two loops, conditioned in from its history: “when hungry, turn to the left” and “when hungry, turn to the right.” Deductive logic grinds to a halt; the facilitation of either loop, neurally speaking, inhibits the other. Their horns lock. The neural activity has no viable pathway out. The brain can’t follow through with a wired-in plan—it has to create a new one.

How? By bringing in new variables that reshape the original loops into a new pathway, one that doesn’t negate either of the original rules, but clarifies which to use when. As the baby grows hungrier, activity spreads through the brain, searching its history for anything that can break the tie. If it can’t find it in the brain, it will automatically search the environment, thrash by thrash. The mathematics of game theory, Putnam said, guarantee that, since the original rules were in service of one and the same goal, an answer, logically speaking, can always be found.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

“Perhaps I have actually found a place in the world that wants me at last—as I am.” 

In this case, the baby’s brain finds a key variable: When “turn left” worked, the neural signal created by the warmth of the mother’s breast against the baby’s left cheek got wired in with the behavior. When “turn right” worked, the right cheek was warm. That extra bit of sensory signal is enough to tip the scales. The brain has forged a new loop, a more general rule: “When hungry, turn in the direction of the warmer cheek.”

New universals lead to new motor sequences, which allow new interactions with the world, which dredge up new contradictions, which force new resolutions, and so on up the ladder of ever-more intelligent behavior. “This constitutes a theory of the induction process,” Putnam wrote.

In notebooks, in secret, using language only he would understand, Putnam mapped out the dynamics of a system that could perceive, learn, think, and create ideas through induction—a computer that could program itself, then find contradictions among its programs and wrangle them into better programs, building itself out of its history of interactions with the world. Just as Turing had worked out an abstract, universal model of the very possibility of computation, Putnam worked out an abstract, universal model of the very possibility of mind. It was a model, he wrote, that “presents a basic overall pattern [or] character of thought in causal terms for the first time.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Putnam had said you can’t understand another person until you know what fight they’re in, what contradiction they’re working through. I saw before me two stories, equally true: Putnam was a genius who worked out a new logic of the mind. And Putnam was a janitor who died unknown. The only way to resolve a contradiction, he said, is to find the auxiliary variables that forge a pathway to a larger story, one that includes and clarifies both truths. The variables for this contradiction? Putnam’s mother and money.

Putnam grew up with money. He was born in 1927 in Ohio, to John B. Putnam, Sr., charming corporate lawyer, and Mildred Andrews Putnam, fearsome lady-who-lunched. They lived in the village of Bratenahl, a tiny neighborhood outside Cleveland, home to the ultra-rich, in a big, white Victorian house with a round cone-topped turret and the expanse of Lake Erie unfurling from their backyard. Whenever Putnam made a new friend, his mother warned him, “They’re probably using you for your money.”

When Putnam and his older brother, Johnny, were little, their parents told them a story about a boy named Ikey. Ikey’s father had lifted Ikey up and sat him high up on the mantle above the fireplace. Then the father told him, “OK, Ikey. Jump!”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

“He is afraid,” Putnam wrote, “but told his daddy will catch him. He is afraid, but told to have faith, and all will go well.” So Ikey jumps. But his father doesn’t catch him. He steps to the side, lets the kid fall. “When Ikey cries and complains,” Putnam wrote, “he is told never to trust anyone, not even his mother and father.” 

That was the moral of the story that Mildred and John Putnam told to their children. Never trust anyone. Not even us. Johnny cried, but Peter just soaked it in.

TWO FRIENDS: Oval with Points by sculptor Henry Moore, donated by Peter Putnam, on the Princeton University campus. Putnam’s mentor and champion, John Archibald Wheeler, described the work as a “place for two friends to sit side by side.” Photo by Justin Smith.

At 16, Putnam joined the Navy, and it was there, in radar school, that he realized his aptitude for physics. At the same time, he realized his desperation to unravel the mystery of minds. He needed to understand the secret motives his mother warned about, especially now that he was coming to grips with his homosexuality, which left him feeling helplessly set apart.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

In 1944, Putnam received word that his brother, a fighter pilot in the Air Force, had been killed overseas. Peter’s diary entry that day read: “Tuesday—Johnny isn’t.” Johnny had been the Putnam’s golden boy: blond-haired, blue-eyed, confident, athletic. Now there was just Peter: bookish, skinny, painfully shy. The taller of the two, he cultivated a slouch, as if embarrassed by his continued existence. Two years later, still reeling from Johnny’s death, he used his Navy credits to enroll as a physics major at Princeton.

Wheeler took him under his wing, bringing him to Copenhagen to meet Niels Bohr, raving to Bohr about Putnam’s “very great interest in the philosophical aspects of physics.”

After graduation from Princeton, Putnam reluctantly enrolled in Yale Law School. Now that he was the only son, he was expected to become a lawyer like his father and grandfather. Two years later, his father was diagnosed with late-stage leukemia. As he was dying, he told Peter to forget law and use his inheritance to return to his real work. So Peter joined the philosophy department at Harvard University, planning to do a Ph.D. on Eddington. But when his father died in 1951, Mildred, wanting to retain control over the only family she had left, withheld the money. Peter’s decisions would go through her. Peter, determined to make his own money, dropped out of school and took a job at Sanders and Associates, an electronics company in New Hampshire.

From his salary alone, he saved up enough money to quit and return to Princeton to study with Wheeler for his Ph.D. He promptly informed Mildred that he would not accept another dime from her, ever. “I shall not need, and will not accept, any more money from you from here on,” he wrote in a letter.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Cut off from the family money, real friendships suddenly seemed possible. “What a funny delightful sensation it is to be asking for the cheapest rooms, and trying to save money,” he wrote in his diary in 1957. “It makes me smile and smile—as though I had a secret. I can feel the friends I have getting wind of it, and speculating—and the nicest part—is that it makes me feel one of them.” He signed the entry, “Self-righteous Peter.” 

One of those friends was Fuller. They were walking across campus when Fuller casually asked what Putnam was working on. Putnam turned to him and asked, “Do you really want to know?”

He’d never told anyone about his theory, but with money no longer blocking the way, it all came spilling out. “We talked and talked till I was bleary-eyed and dead tired and had to quit,” Fuller said.

He built walls around his work, walls made of words, but he built them too high—they kept everyone out.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Putnam dressed in shabby clothes. He sold the Cadillac convertible Mildred had bought him, used the money to buy a bicycle, and gave the remainder of the proceeds to Princeton. He also gave them all the stock he’d earned at Sanders—some 600 shares, valued around $9,000, which he’d asked for in lieu of raises—on the condition that they wouldn’t sell them until he gave the green light. He did, a decade later, and they sold for more than $1 million.

Putnam asked that the donations be used to buy great works of modern sculpture to be displayed around campus. Clarke told me Putnam’s love of abstract sculpture came from “his thinking about the brain and the centrality of motor pathways”—the sculptural form resolving an artist’s own contradictions, then inviting the viewer to move, to think, in new ways. The collection was to be a memorial to his brother. The donor was to be listed as “Anonymous.”

After a stint teaching physics in Amherst, Putnam followed Fuller to New York City in 1963. Fuller was teaching at Columbia, so Putnam taught a summer seminar in the physics department. His lectures were so heavily laced with philosophy that students from the Union Theological Seminary across the street began showing up. After class they’d go to a nearby café, quoting lines of Putnamese. “Jazz is the mathematization of the soul.” “We know things in the act, not in their essence.” The Seminary hired Putnam, and set him up in the basement apartment on Claremont Ave.

In the day, Putnam taught and wrote; at night he’d walk uptown to Harlem to dance at the jazz clubs, a neural free-for-all to enact his improvisational mind. Most of the time, he was the only white guy there. One night he met a Black ex-Army Major named John DeBrew, who went by the nickname Claude. “Claude sneaked under my defenses as a bird or flower does,” Putnam wrote. “I’ve been most lucky in finding a gentle, affectionate person.” He was open about the relationship with everyone, noting in a letter to Fuller that, when it came to his sexuality, “We should be able to discuss anything, and treat it as we should a problem in mathematics.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

One of Putnam’s students, Kim Hopper, now a medical anthropologist at Columbia, told me that Putnam wrote an article for the Union Seminary Quarterly Review, where he mentioned, in an offhanded way, the “depth and sensitivity of the homosexual community, in which I have been privileged to participate.” “This was at a time when nobody came out,” Hopper said, “especially not in a theological journal!”

“Perhaps I have actually found a place in the world that wants me at last—as I am,” Putnam wrote to his mother. “In any case I am very pleased.”

Mildred was less pleased. She offered Claude $35,000 to leave Peter. Claude didn’t take the bribe, a move that endeared him to Peter for the rest of their lives. Still, Mildred sent a note to Peter, scribbling: “Remember Ikey.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Back in Princeton, Wheeler was coming around to the idea that the observer might be implicated in quantum mechanics, and he knew his best bet for understanding the observer was Putnam. He was hoping that Putnam would return to Princeton so they could work together, uniting a theory of the observer with a theory of the observed. Putnam wanted nothing more. “So many people dream of convincing father images of the value of their work,” he replied.

When Mildred realized what was happening, she jumped in, trying to ensure that Peter would get the job and that he’d owe it all to her. She began dangling donations, offering to build a new physics building at Princeton with Wheeler’s name on it. Wheeler wanted no part of it, but Mildred was a force of nature, a hurricane in pearls. “Hopefully something constructive could be arranged over luncheon,” she wrote to Wheeler, adding that she wanted to keep the arrangement between themselves. “Will you please forget I ever wrote this letter and throw it into the fire, as Peter would never forgive me.”

Putnam pleaded with her to stay out of his relationship with Wheeler, but she continued to allude to secret meetings and quid pro quo donations until he didn’t know who or what to believe.

Unable to trust that Wheeler’s interest was pure, Putnam refused to consider a position at Princeton, or a fellowship at the nearby Institute for Advanced Study. He stuck with teaching at the Seminary.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

It’s clear from Wheeler’s journals his interest in Putnam’s work was genuine and deep. Over and over, he read the few papers that Putnam gave him, writing out notes and questions line by line. “He would throw up his hands in despair,” Wheeler’s daughter, Alison Lahnston, told me, “but he kept at it.”

One morning in 1974, over breakfast in Manhattan, Wheeler took 12 pages of notes as Putnam talked about his work, then submitted a book proposal to W.H. Freeman & Company on Putnam’s behalf. The publishers bit, and were ready to draw up a contract, but Putnam again worried that his mother was behind the offer, and refused to sign.

Just then, a perfect opportunity arose to present Putnam to the public. Wheeler was invited by the Neurosciences Research Program at MIT to speak at their March 1975 meeting on “reality and consciousness.” He insisted he could only do it as half of a pair. From Reality to Consciousness. From Consciousness to Reality.

As the meeting approached, Putnam grew nervous. He demanded to know whether Mildred had been involved behind the scenes. Wheeler assured him that she wasn’t, that the talks were solely his idea. Together, they boarded a plane to Boston.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

I listened to the meeting, recorded on a reel-to-reel, stowed away in the archives at MIT. Here, finally, was Putnam’s chance to explain his ideas to the top neuroscientists of the time. I pressed the headphones tight against my ears.

ONE OF THE GREATS: Peter Putnam (far right) with Robert Fuller, baby Karen, and Fuller’s former wife, Ann Lackritz Fuller, in Amherst, Massachusetts, circa 1961. Fuller, who collaborated with Putnam for 10 years, called Putnam “one of the great philosophers of the 20th century.”  Courtesy of Robert Fuller.

Wheeler had just finished speaking about the observer in quantum mechanics and introduced Putnam with a warning. “Some terms Peter uses, one needs a glossary to translate.” Wheeler placed a transparency on the projector—he’d made an actual glossary of Putnam’s terms. The crowd burst into laughter. I didn’t have to see Putnam’s face to feel it growing hot. When he began to speak, he stuttered. 

“You only perceive signals that are useful for shaping behavior … A game is a special kind of mathematics … But for a game you need a goal function … We’re suggesting that the category repetition is a candidate … You’re searching for rules of choice that allow a repeating or self-reproducing path … There’s a transcendental core to the laws of physics themselves …”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

The crowd grew restless. Wheeler’s talk had gone long, and there wasn’t time for Putnam to finish. The neuroscientists headed out for lunch and the tape cut out.

Things went from bad to worse. Back in New York, Putnam learned he’d lost his job at Union. The President cited “budgetary concerns,” which Putnam took as a veiled attempt to ask for a donation, suspecting that his mother had suggested as much to the administration behind his back.

Wheeler made one last ditch effort to convince Putnam not to give up on academia. I found a handwritten note he wrote in 1975. Not a note, exactly—more like an affidavit.

“I find it utterly impossible to believe that your mother directly or indirectly made any contributions to, or in any way influenced, the action of the Institute for Advanced Study, MIT Neurophysiology, [or] Freeman and Company … I have never been and do not intend to be a party to any arrangement in which relations between you and me, or between you and any institutional setting in which I have any say or knowledge, are dependent in any way whatsoever on any contribution, or any expectation of any contribution, from your mother. There is no lawyer-like or other reservations or loophole in the intent and content of this freely given assurance.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

But even Wheeler couldn’t penetrate Putnam’s defenses. Never trust anyone. It was a rule of behavior that had dug a trench in his neural circuitry, formed a universal, self-reinforcing loop, and no matter how many alternatives competed with it in parallel, it was always strengthened. It always won. He couldn’t risk having people take a cursory interest in his work just to flatter him, to court him for his money. To weed out anyone who wasn’t in it for the right reasons, he refused to provide an easy summary. It was total commitment or nothing at all. So he built walls around his work, walls made of words, but he built them too high—they kept everyone out, and kept Putnam in.

In June 1975, Putnam sat down and wrote a letter to Wheeler:

“It should be obvious that what I’m doing is a lot of nonsense. I didn’t convince any of the big boys at the conference—didn’t even get any excited about any of the points or themes … Clearly all I’ve been doing is hoodwinking a few naive though often top students—after ten years of teaching my crazy course … finally, the right thing has been done.”

Putnam placed his books on a table at Union Seminary for the taking, dumped stacks of manuscripts in the trash, gave his records and turntable to a janitor. Then he opened to a fresh page in his journal and scrawled, “For myself, given my weaknesses, this is the end. I can’t try any more … At least I’ve finished things off. It fits well enough—for a start for someone else … The best I can do—like Rimbaud—is to vanish.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

An oil town built on the swamplands between the Mississippi River and the Gulf of Mexico, Houma, Louisiana, is about an hour’s drive southwest from New Orleans. In 1975, Putnam had signed himself and Claude up for VISTA—Volunteers in Service to America, the domestic branch of the Peace Corps—and VISTA sent them to Houma. They were promised government housing, but when the building manager saw that Claude was Black, their apartment suddenly became unavailable. They tried another housing project. Same story. Finally, they went to Senator Circle, the Black project on the other side of town, but they weren’t welcome there either. Interracial and gay—there was no housing project for that. So Putnam found them a spot in a trailer park. The landlady said she’d pray for them.  

They reported for VISTA duties at the Wayout Clinic, a nonprofit serving the Black community in Terrebonne Parish. The city had scraped together funds to open a new rec center and they’d asked Wayout to run it. Putnam asked the program director why they couldn’t just give the money directly to the Black community and let them run their own rec center instead of having a bunch of white people in charge. The director told Putnam that it was impossible, that there were legal complexities he wouldn’t understand; it would have to go through a nonprofit, it was very complicated. Putnam didn’t mention that he was a physicist or that he’d studied at Yale Law. He just turned around and registered a new nonprofit, the Terrebonne Improvement Association (TIA). He put together an all-Black Board of Directors, then applied for VISTA volunteers of their own.  

The TIA published their own community newsletter, featuring pieces by local Black writers alongside transcripts of speeches by national civil rights leaders. Claude delivered speeches to the TIA, co-written with Putnam behind the scenes. Putnam thought it was important that the community know all the legal tricks the white CEOs and politicians used to keep them down, so Claude spoke about reapportionment and gerrymandering; he urged them to vote in local elections, to make their voices heard on school boards and in town halls. The TIA got two Black representatives elected to the Police Jury. They made plans to fight for better services in minority neighborhoods, for their cut of revenue sharing, for affirmative action all the way up the ladder.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .
Illustration by Cornelia Li

Putnam continued to keep his wealth a secret. Mildred had released the inheritance from his father to him in 1972. Putnam put all the money into a charitable trust and named it the Mildred Andrews Fund so that it wouldn’t bear his name. He again stipulated the money should be used to fund public sculpture, this time in New York City and Cleveland, Ohio. The artworks, he wrote, “shall be so placed as to benefit especially our underprivileged (ghetto) areas and so chosen as to express their life and outlook.” By the mid ’80s, Putnam, through stock investments, had grown the fund to $40 million. “Peter was the most skilled master of finance” he’d ever known, commented the family lawyer. Putnam never touched a penny for himself.

To make ends meet, he took odd jobs repairing radios and shucking oysters. He bought a place where Claude and he could live permanently—a small, one-bedroom apartment with a tiny kitchen and wood paneling on the walls. It was on the main road, next to a vacant lot, but the back door opened out onto the bayou, where they could sit and watch the shrimp boats go by and the moonlight ripple on slow, dark water. “Life is a simple thing,” Claude told Peter. “I want to live my life so people associated with me are happy.”

Eventually Putnam landed a gig as a night watchman and janitor for the Department of Transportation. “It’s clearly the best job I’ve ever had,” he told Spinello on the recordings. “I needed to get other kinds of roots in the community. I think that this position is in some sense symbolically right. Whereas my position as a teacher, you know, symbolically stunk.”

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

The emotion welled up in his voice. “I mean, how could I get in direct relationship to people?” he stuttered. “When there’s a big hunk of money … Even my own prof, who I loved, who I did my thesis under … My mother denied it was going on. She would talk dramatically about committing suicide unless I believed her. I said, ‘I believe you.’ But, you know, how can I protect myself through a thing like that?”

The check from the janitor was the largest single gift the environmental group had ever received.

His voice grew calmer, sadder. “I did kick Wheeler pretty hard in the face for that. I think I was wrong in doing that. The issue is someplace else. I should have been down here sweeping floors.”  

Putnam knew his mother had destroyed his relationship with Wheeler and had prevented him from getting his work out into the world, but he never blamed her. He believed that her tactics, however much they hurt him, were the rules of behavior she needed to survive in a world dominated by men.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

“It is especially hard for any woman to be herself,” he once wrote her, “and believe in herself as she is, surrounded as she is with these ridiculous man-made images of how she is supposed to feel and act.”

It’s easy to say why someone is wrong, Putnam said. The hard part is figuring out why they’re right. And everyone is right. Everyone has some central insight, hard won by the consistency-making mechanism of the brain, built of past experiences, cast as motor predictions, a pattern that repeats, sustains itself in the chaos. Our job is to pan for it like gold, sift it into our own nervous systems, reconcile the resulting contradiction, become something new.

“My life’s work, if there is a one-sentence formula for it, is trying to find some path of reconciliation with you,” Putnam wrote her.

Mildred got sick in 1981, and she moved in with her son and Claude. She could have lived anywhere, a mansion in New Orleans, with a staff, like she had back in Ohio, a chef, a housekeeper, nurses. Instead, Putnam gave her the bedroom, and he and Claude slept on the pullout couch. They made for a strange family, content in the knowledge that they were all there for the right reasons. Putnam and Claude took care of Mildred for three years, until she died in 1984.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

Wheeler continued trying to convince Putnam to publish his work. He drove from Princeton to Houma to visit—noting afterward that Putnam was “living as poor as Job’s turkey.” In 1986, Wheeler wrote to Putnam, ending the letter: “There is so much more I’d like to say, but let me sum it all up in one word, gratitude: gratitude to you for all you’ve meant and done over all these years, gratitude to heaven above that you’re still on this earth, still capable—God willing, and in God’s good time—to publish something great.”

December 7, 1987. Putnam swung his leg over his bicycle, like he’d done so many times before.

The drunken swerve. 

The screech.  

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

The silence.

On a fall afternoon in 2024, I wandered the Princeton campus among the towering sculptures. There was Tony Smith’s abstract Moses, Alexander Calder’s Five Disks: One Empty, Antoine Pevsner’s Construction in the Third and Fourth Dimension, and Jacques Lipchitz’s Song of the Vowels. This was the trail of breadcrumbs Putnam left behind.

I watched as other people—students and professors—strolled right past them, as if the sculptures were invisible. Which was weird, because they’re huge. Louise Nevelson’s Atmosphere and Environment X looms 21 feet tall, a steel screen with geometric forms in cut-out compartments that reminded me of a library, or a secret language, or both. Picasso’s Head of a Woman, with her stark, cubist angles, weighs in at 20,000 pounds, her rosy cheeks rendered in red quartzite, swirling eyes in black granite.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

The Princeton sculptures aren’t Putnam’s only breadcrumbs.

Along Cleveland’s Cuyahoga River is Gene Kangas’ Hart Crane Memorial Sculpture, commissioned by Putnam to memorialize the poet, who, after being assaulted onboard a steamship for being gay, jumped overboard and drowned. At Howard University, there’s A Bridge Above and Beyond by Richard Hunt, symbolizing the connection between Africa and her children in America, dedicated to “Black womanhood” and “single mothers everywhere.”

In New York City’s West Village, in a sliver of greenery known as Christopher Park, across from the Stonewall Inn, where 1969 riots sparked the beginning of the gay rights movement in the United States, is a sculpture of four figures by George Segal. Two men, standing, appear deep in conversation, one’s arm wrapped around the other’s shoulder; two women sit side by side on a bench, one’s hand on the other’s knee. When Putnam commissioned the piece, he stipulated that the work “had to be loving and caring, and show the affection that is the hallmark of gay people … and it had to have equal representation of men and women.” When it was installed, the media called it the “first monument to homosexuals in the United States.”

In the center of the Princeton campus, I came upon Oval with Points by Henry Moore, a massive womblike, hollowed-out thing, its bronze now patinated sea-foam green, with two points reaching in toward the center, almost meeting, but not quite. I sat down in the oval and thought about how many things—people, ideas—hide in plain sight, and how many answers to scientific mysteries might be stashed away, junked, or forgotten.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

I thought of Wheeler, who helped pick out this sculpture—he described it as a “place for two friends to sit side by side”—and of his lifelong fear that Putnam’s work would be lost.

It’s impossible to know what might have happened if Putnam had gotten his ideas out when he was alive. When he first worked out his theory, Turing was here in Princeton visiting John von Neumann as he was building a stored-program electronic computer, which the press referred to as an “electronic brain.” The comparison between brains and Turing machines was immediately embraced by the scientific community and so entrenched during Putnam’s lifetime that his suggestion that the brain is a “computer of a fundamentally different genus” simply couldn’t compute.

The reason an induction machine—a mind—can do more than a universal Turing machine is because it’s always reaching out into the world. Which was exactly what Putnam himself struggled to do.

Putnam turned his writings into a self-contained room where Ikey could hide and no one would find him. The only one who managed to crack open the door was Claude. “He teaches me how to live outside words,” Putnam wrote. Claude lived in their Houma apartment until he died in 2008.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

In Ohio’s Morgan Swamp on the southern shoreline of Lake Erie, among beaver ponds and vernal pools, nestled in a forest of yellow birch and hemlock, are tundra swans and four-toed salamanders, white calla lilies and river otters. In the Animas Mountain range in New Mexico, wild turkeys and long-nosed bats, white-sided jackrabbits and spotted owls, live and breed in 500 square miles of unadulterated ecosystem. On the sandbars and shallows of Nebraska’s Platte River, sandhill cranes swoop down on the floodplain to roost and forage en route to the Arctic. Some stop in the Ohio wetlands on their return, where they dance in the shadows of soaring bald eagles headed to nest in Putnam Marsh.

All of these lands still exist thanks, in large part, to Putnam. His will stipulated that upon his death, his money—all $40 million of it—be given to the Nature Conservancy. When the check from the janitor showed up, it was the largest single gift the environmental group had ever received. Putnam would have been happy to remain anonymous. Only the marsh—which he requested be named for his parents—gives him away.

Today, science is beginning to catch up to Putnam. His ideas about the plasticity of the brain and the importance of neural conditioning have become mainstream. Many cognitive scientists are pursuing a theory known as “embodied mind” that emphasizes the central role of motor behavior in cognition and perception, so central to Putnam’s own theory.

At the same time, as Fuller put it, “there’s stuff in Putnam that no one has thought of yet. There’s precious new material for scientists who are on the cutting edge.” That includes not only those working in cognitive science, but also in artificial intelligence and robotics. AI researchers are eagerly searching for models of general intelligence, wondering how it is that humans learn, or have common sense, or deal with novel situations. How humans, as Putnam explained, can perform induction.

ADVERTISEMENT
Nautilus Members enjoy an ad-free experience. Log in or Join now .

I’d spent more than a decade hunched over inscrutable pages under the weight of so much regret about how Putnam’s story had ended. Now, sitting up in the soft curve of Moore’s sculpture, I traced my fingers along the surface where so many fingers had traced it before; in that one spot, the patina was rubbed clean, and the original bronze shone through. Sunlight glistened off the metal, and it dawned on me that maybe Putnam’s work hadn’t been lost. Maybe it was just waiting for its moment.

The post Finding Peter Putnam appeared first on Nautilus.

Read the whole story
GaryBIshop
9 days ago
reply
Wow! Brilliant!
Share this story
Delete

Where AI Provides Value

1 Comment

If you’ve worried that AI might take your job, deprive you of your livelihood, or maybe even replace your role in society, it probably feels good to see the latest AI tools fail spectacularly. If AI recommends glue as a pizza topping, then you’re safe for another day.

But the fact remains that AI already has definite advantages over even the most skilled humans, and knowing where these advantages arise—and where they don’t—will be key to adapting to the AI-infused workforce.

AI will often not be as effective as a human doing the same job. It won’t always know more or be more accurate. And it definitely won’t always be fairer or more reliable. But it may still be used whenever it has an advantage over humans in one of four dimensions: speed, scale, scope and sophistication. Understanding these dimensions is the key to understanding AI-human replacement.

Speed

First, speed. There are tasks that humans are perfectly good at but are not nearly as fast as AI. One example is restoring or upscaling images: taking pixelated, noisy or blurry images and making a crisper and higher-resolution version. Humans are good at this; given the right digital tools and enough time, they can fill in fine details. But they are too slow to efficiently process large images or videos.

AI models can do the job blazingly fast, a capability with important industrial applications. AI-based software is used to enhance satellite and remote sensing data, to compress video files, to make video games run better with cheaper hardware and less energy, to help robots make the right movements, and to model turbulence to help build better internal combustion engines.

Real-time performance matters in these cases, and the speed of AI is necessary to enable them.

Scale

The second dimension of AI’s advantage over humans is scale. AI will increasingly be used in tasks that humans can do well in one place at a time, but that AI can do in millions of places simultaneously. A familiar example is ad targeting and personalization. Human marketers can collect data and predict what types of people will respond to certain advertisements. This capability is important commercially; advertising is a trillion-dollar market globally.

AI models can do this for every single product, TV show, website and internet user. This is how the modern ad-tech industry works. Real-time bidding markets price the display ads that appear alongside the websites you visit, and advertisers use AI models to decide when they want to pay that price—thousands of times per second.

Scope

Next, scope. AI can be advantageous when it does more things than any one person could, even when a human might do better at any one of those tasks. Generative AI systems such as ChatGPT can engage in conversation on any topic, write an essay espousing any position, create poetry in any style and language, write computer code in any programming language, and more. These models may not be superior to skilled humans at any one of these things, but no single human could outperform top-tier generative models across them all.

It’s the combination of these competencies that generates value. Employers often struggle to find people with talents in disciplines such as software development and data science who also have strong prior knowledge of the employer’s domain. Organizations are likely to continue to rely on human specialists to write the best code and the best persuasive text, but they will increasingly be satisfied with AI when they just need a passable version of either.

Sophistication

Finally, sophistication. AIs can consider more factors in their decisions than humans can, and this can endow them with superhuman performance on specialized tasks. Computers have long been used to keep track of a multiplicity of factors that compound and interact in ways more complex than a human could trace. The 1990s chess-playing computer systems such as Deep Blue succeeded by thinking a dozen or more moves ahead.

Modern AI systems use a radically different approach: Deep learning systems built from many-layered neural networks take account of complex interactions—often many billions—among many factors. Neural networks now power the best chess-playing models and most other AI systems.

Chess is not the only domain where eschewing conventional rules and formal logic in favor of highly sophisticated and inscrutable systems has generated progress. The stunning advance of AlphaFold2, the AI model of structural biology whose creators Demis Hassabis and John Jumper were recognized with the Nobel Prize in chemistry in 2024, is another example.

This breakthrough replaced traditional physics-based systems for predicting how sequences of amino acids would fold into three-dimensional shapes with a 93 million-parameter model, even though it doesn’t account for physical laws. That lack of real-world grounding is not desirable: No one likes the enigmatic nature of these AI systems, and scientists are eager to understand better how they work.

But the sophistication of AI is providing value to scientists, and its use across scientific fields has grown exponentially in recent years.

Context matters

Those are the four dimensions where AI can excel over humans. Accuracy still matters. You wouldn’t want to use an AI that makes graphics look glitchy or targets ads randomly—yet accuracy isn’t the differentiator. The AI doesn’t need superhuman accuracy. It’s enough for AI to be merely good and fast, or adequate and scalable. Increasing scope often comes with an accuracy penalty, because AI can generalize poorly to truly novel tasks. The 4 S’s are sometimes at odds. With a given amount of computing power, you generally have to trade off scale for sophistication.

Even more interestingly, when an AI takes over a human task, the task can change. Sometimes the AI is just doing things differently. Other times, AI starts doing different things. These changes bring new opportunities and new risks.

For example, high-frequency trading isn’t just computers trading stocks faster; it’s a fundamentally different kind of trading that enables entirely new strategies, tactics and associated risks. Likewise, AI has developed more sophisticated strategies for the games of chess and Go. And the scale of AI chatbots has changed the nature of propaganda by allowing artificial voices to overwhelm human speech.

It is this “phase shift,” when changes in degree may transform into changes in kind, where AI’s impacts to society are likely to be most keenly felt. All of this points to the places that AI can have a positive impact. When a system has a bottleneck related to speed, scale, scope or sophistication, or when one of these factors poses a real barrier to being able to accomplish a goal, it makes sense to think about how AI could help.

Equally, when speed, scale, scope and sophistication are not primary barriers, it makes less sense to use AI. This is why AI auto-suggest features for short communications such as text messages can feel so annoying. They offer little speed advantage and no benefit from sophistication, while sacrificing the sincerity of human communication.

Many deployments of customer service chatbots also fail this test, which may explain their unpopularity. Companies invest in them because of their scalability, and yet the bots often become a barrier to support rather than a speedy or sophisticated problem solver.

Where the advantage lies

Keep this in mind when you encounter a new application for AI or consider AI as a replacement for or an augmentation to a human process. Looking for bottlenecks in speed, scale, scope and sophistication provides a framework for understanding where AI provides value, and equally where the unique capabilities of the human species give us an enduring advantage.

This essay was written with Nathan E. Sanders, and originally appeared in The Conversation.

Read the whole story
GaryBIshop
9 days ago
reply
Good thinking.
Share this story
Delete

Android 16 glitch is making some Pixel phones a pain to navigate

1 Comment
Pixel phone gestures and nav controls are going off the rails for some users in latest Android 16 bug.

Read the whole story
GaryBIshop
10 days ago
reply
I have seen this on my 9pro. In the maps app the nav buttons block access to the post button.
jgbishop
10 days ago
NewsBlur suffered from this issue until recently. Looking at the bug on GitHub, it looks like they changed the default positioning behavior to edge-to-edge, causing UI elements to overlap.
Share this story
Delete

Seven replies to the viral Apple reasoning paper – and why they fall short

1 Comment

Also: another paper that seals the deal

The Apple paper on limitations in the “reasoning” of Large Reasoning Models, which raised challenges for the latest scaling hypothesis, has clearly touched a nerve. Tons of media outlets covered it; huge numbers of people on social media are discussing.

My own post here laying out the Apple paper in historical and scientific context was so popular that well over 150,000 people read it, biggest in this newsletter’s history. The Guardian published an adaptation of my post (“When billion-dollar AIs break down over puzzles a child can do, it’s time to rethink the hype”) The editor tells me readers spent a long time reading it, notably longer than usual, as if people really wanted to understand the arguments in detail. (The ACM computer science society is reposting the essay, too, and there is now a French version as well).

Tons of GenAI optimists took cracks at the Apple paper (see below), and it is worth considering their arguments. Overall I have seen roughly seven different efforts at rebuttal, ranging from nitpicking and ad hominem to the genuinely clever. Most (not all) are based on grains of truth, but are any of them actually compelling?

Let’s consider.

  1. Humans have trouble with complex problems and memory demands. True! But incomplete. We have every right to expect machines to do things we can’t. Cars have more stamina, calculators don’t make arithmetical errors. That’s why we invented computers: to do repetitive calculation without errors. And in many cases (including the Tower of Hanoi, which featured prominently in the paper)) we have existing systems that work perfectly without errors. AGI should be a step forward; in many cases LLMs are a step backwards. And note the bait and switch from “we’re going to build AGI that can revolutionize the world” to “give us some credit, our systems make errors and humans do, too”. The real takeaway from the Apple paper is that LLMs can’t be trusted to run algorithms as complexity and distance from the training distribution grows (just as humans shouldn’t serve as calculators). If we want to get to AGI, we will have to better.

  2. The Large Reasoning Models (LRMs) couldn’t possibly solve the problem, because the outputs would require too many output tokens (which is to say the correct answer would be too long for the LRMs to produce). Partial truth, and a clever observation: LRMs (which are enhanced LLMs) have a shortcoming, which is a limit on how long their outputs can be. The correct answer to Tower of Hanoi with 12 moves would be too long for some LRMs to spit out, and the authors should have addressed that. But crucially (i) this objection, clever as it is, doesn’t actually explain the overall pattern of results. The LRMs failed on Tower of Hanoi with 8 discs, where the optimal solution is 255 moves, well within so-called token limits; (ii) well-written symbolic AI systems generally don’t suffer from this problem, and AGI should not either. The length limit on LLM is a bug, and most certainly not a feature. And look, if an LLM can’t reliably execute something as basic as Hanoi, what makes you think it is going to compute military strategy (especially with the fog of war) or molecular biology (with many unknowns) correctly? What the Apple team asked for was way easier than what the real world often demands.

  3. The paper was written by an intern. This one, which infuriated me because is a form of ad hominem argument rather than substance, is misguided and only barely true – and utterly lacking in context. It is true that the first author was an intern at Apple, Parshin Shojaee, but (i) she (not he as some fool I won’t name presumed, without spending two seconds of research) also happens to be a very promising third year Ph.D. student with many paper presentations at leading conferences, (ii) the article, if you actually read it, makes it clear she shared lead responsibility with Iman Mirzadeh, who has a Ph.D.; (iii) the paper actually has six authors, not one, and four have Ph.D.s; for good measure one is Yoshua Bengio’s brother Samy Bengio, well-known and very distinguished in his own right within the machine learning community; (iv) it is a common practice in many scientific fields to put the junior author first, senior author last, as this paper did; thousands of major articles have done the same (and never been criticized for doing so — it’s a true desperation measure); (v) what really matters is the quality of the work. Alfred Sturtevant was an undergraduate when he invented gene maps.

  4. Bigger models might to do better. True, and this is always the case (I have seen one report that o3-pro can do at least one of the problems, at least some of the time; I have not seen a thorough study yet). Bigger models sometimes do better because of genuine improvements in the model, sometimes because of problem-specific training. (From the outside we can never know which). But here’s the thing, we can’t know in advance what model is big enough (if any) for any given problem. And Apple’s result that some pretty large models could succeed at 6 discs, giving the illusion of mastery, but collapse by 8, is ominous. One is left simply having to test everything, all the time, with little guarantees of anything. Some model might be big enough for task T of size S and fail on the next size, or on Task T’ that is slightly different, etc. It all becomes a crapshoot. Not good.

  5. The systems can solve the puzzles with code. True in some cases, and a huge win for neurosymbolic AI, given that they can’t reliably solve the puzzles without code, and given that code is symbolic. Huge vindication for what I have been saying all along: we need AI that integrates both neural networks and symbolic algorithms and representations (such as logic, code, knowledge graphs, etc). But also, we need to do so reliably, and in a general way and we haven’t yet crossed that threshold. (Importantly, the point of the Apple paper goal was to see how LRM’s unaided explore a space of solutions via reasoning and backtracking, not see how well it could use preexisting code retrieved from the web.) An analogy: A student might complain about a math exam requiring integration or differentiation by hand, even though math software can produce the correct answer instantly. The teacher’s goal in assigning the problem, though, isn’t finding the answer to that question (presumably the teacher already know the answer), but to assess the student’s conceptual understanding. Do LLM’s conceptually understand Hanoi? That’s what the Apple team was getting at. (Can LLMs download the right code? Sure. But downloading code without conceptual understanding is of less help in the case of new problems, dynamically changing environments, and so on.)

  6. The paper has only four examples, and at least one of them (Hanoi) isn’t perfect. I doubt any of them are perfect, but the four together provide converging evidence with dozens of other prior papers (including some of my own), and I am confident many more examples will be uncovered. I already found a couple of analogous failures in algorithm application myself, which I will write about in a few days. Tal Linzen at NYU just published yet another example, with “models .. able to do the right thing for simple versions of [a language] problem (small grammars, short strings), but [with] accuracy drop[ping] very quickly as the problem becomes more complex.” Mark my words: in time, we will see scores of papers reinforcing the Apple results.

  7. The paper is not news; we already knew these models generalize poorly. True! (I personally have been trying to tell people this for almost thirty years; Subbarao Rao Kambhampati has been trying his best, too). But then why do we think these models are the royal road to AGI? The real news here, aside from the fact that this was a clever study nailing down an important point, is that people are finally starting to pay attention, to (one of the) two biggest Achilles’ Heels of generative AI, and to appreciate its significance. (Tune in to this newsletter on the weekend to hear about the other.) Hilarious, by the way, is hearing both “it’s wrong” and “we knew it all along”, simultaneously. In at least one case I saw a single person say both, minutes apart!

Bottom line? None of the rejoinders are compelling. If people like Sam Altman are sweating, it’s because they should. The Apple paper is yet another clear sign that scaling is not the answer; for once, people are paying attention.

§

The kicker? A Salesforce paper also just posted, that many people missed:

alt

In the “multi-turn” condition, which presumably would require reasoning and algorithmic precision, performance was only 35%.

Talk about convergence evidence. Taking the SalesForce report together with the Apple paper, it’s clear the current tech is not to be trusted.

Gary Marcus, professor emeritus at NYU, and author of The Algebraic Mind and “Deep learning is hitting a wall”, both of which anticipated the correct results, is thrilled to see people finally realize that scaling is not enough to get us to AGI. Now perhaps we can begin to build better AI.

Marcus on AI is a reader-supported publication. Your support, in the form of a subscription, paid or otherwise, appreciated!

Adblock test (Why?)

Read the whole story
GaryBIshop
11 days ago
reply
Super interesting.
Share this story
Delete

Manufacturing in America

1 Comment

The latest Smarter Every Day video paints a dire picture of the United States' ability to manufacture anything complex. Definitely worth a watch. I definitely want to change the way I purchase products as a result of this, and likely will start avoiding Amazon when possible.

Also, I definitely want one of these scrubbers!

Read the whole story
GaryBIshop
12 days ago
reply
Yes, really scary and also a nice scrubber.
Share this story
Delete

Make Magical-Looking Furniture With Kerf Bend Wizard

1 Comment
Wood bent into a spiral

The intersection between “woodworkers” and “programmers” is not a densely populated part of the Venn diagram, but [Michael Schiebler] is there with his Kerf Bend Wizard to help us make wood twist and bend like magic.

Kerf bending is a fine technique we have covered before: by cutting away material on the inside face of a piece of wood, you create an area weak enough to allow for bending. The question becomes: how much wood do I remove? And where? That’s where Kerf Bend Wizard comes to the rescue.

More after the break…

From spline (user input in black, expected output in pink)…

You feed it a spline– either manually or via DXF–and it feeds you a cut pattern that will satisfy that spline: just enough wood removed in just the right places that the edges of the cut should touch when the bend is achieved. This means less cut time and a stronger piece than eyeballing the kerfs. It works with both a table saw blade or a tapered end mill on a CNC or manual router. You can specify the kerf width of your table saw, or angle of your end mill, along with your desired cut depth.

… to cuts …

The output is DXF, convenient for use with a CNC, and a simple table giving distances from the edge of the piece and which side to cut, which is probably easier for use on the table saw. (Kerf Bend Wizard is happy to handle complex bends that require kerfing both sides of the material, as you can see.)

… to curved wood.

This was [Michael]’s thesis project, for which he hopefully got a good grade. The code is “semi-open” according to [Michael]; there’s a GitHub where you can grab an offline version for your own use, but no open-source license is on offer. Being a broke student and an artist to boot, [Michael] also can’t promise he will be able to keep the web version available without ads or some kind of monetization, so enjoy it while you can!

If CNCs or table saws aren’t your thing, kerf bending has long been used with laser cutters, too.

Our thanks (which, as always, is worth its weight in gold) to [Michael] for the tip. If you’re in the intersection of the Venn diagram with [Michael], we’d love to hear what you’re up to.

Read the whole story
GaryBIshop
12 days ago
reply
Cool computer + woodworking project!
Share this story
Delete
Next Page of Stories