Dear Commons Community,

I have just finished reading The MANIAC by Benjamin Labatut.  The title refers jointly to John von Neumann and the MANIAC computer that he developed. While officially labelled as fiction, all of the characters are real people revealing details of John von Neumann’s life and death and his contributions to computer science, game theory, artificial intelligence, and the development of nuclear weapons.  The first two thirds are all about von Neumann, his genius, his abrasiveness, and his colleagues.  He is on a lifelong quest to prove that “mathematics is the closest we can come to the mind of Hashem (God.)” His death of cancer at the age of 54 is very detailed.

The last third of the book focuses on the development of DeepMind’s AI program AlphaGo and the South Korean Go master Lee Sedol.  von Neumann is never mentioned in these last eighty pages (out of 354 total pages.)  Labatut is attempting to show that DeepMind and its AI programs are beginning to push beyond human intelligence.

A most interesting read. Try it if you are at all interested in von Neumann, the early development of digital computing, and AI.

Below is an excellent review that appeared in The Washington Post.  I give credit to the reviewer, Becca Rothfeld.  This was not an easy book to review.

Tony

——————————————————-

The Washington Post

The MANIAC

Benjamín Labatut’s captivating book explores the frontiers of human invention, including artificial intelligence

Review by Becca Rothfeld

September 21, 2023 at 4:43 p.m.

In 1810, the German Romantic and incurable neurotic Heinrich von Kleist published a short and haunting fable. The narrator is strolling in the park when he sees a dancer watching a puppet show. At first, he is surprised that a living artist is transfixed by mere marionettes, but the dancer is vehement that the wooden troupe has surpassed its human counterpart. Indeed, the dancer claims that the sole flaw is the bumbling puppeteer. When this final “fraction of spirit” is expunged — when the puppets pass “completely into the realm of mechanical forces” — they will at last attain the uncanny perfection unique to inanimate things. The greatest grace of all, the dancer concludes, appears only “in the form of a manikin, or a god.” The former is unconscious; the latter boasts “infinite consciousness.”

Like Kleist’s dancer, the geniuses who fascinate the Chilean author Benjamín Labatut dream of abdicating their personhood to become gods and machines. In his last book, the captivating and unclassifiable “When We Cease to Understand the World” (2021), 20th-century scientists on the verge of making dizzying discoveries chafe against the limits of the human perspective. The astronomer Karl Schwarzschild devises black hole theory and longs to peer into the utter opacity of the abysses he studies, while the physicists Werner Heisenberg and Erwin Schrödinger are racked by mystical visions as they develop the sibylline formulas of quantum mechanics. In order to understand their unnerving new reality, all three scientists must become more — and less — than human.

Labatut’s latest virtuosic effort, at once a historical novel and a philosophical foray, is a thematic sequel, an exploration of what results when we take reason to even further extremes. The resident genius of “The MANIAC” is the polymath and pioneering computer scientist John von Neumann, who displays “a sinister, machinelike intelligence.” When he is lost in thought, it is as if gears are whirring, but it is also as if “the divine reaches down to touch the Earth.”

 

As the book’s many narrators relate, often with a touch of resentment, the mathematician was in every respect “an alien among us.” He could see into other worlds, but he could not tie his shoes. “Spiritually, he was an ignoramus,” a colleague complains. But even if he could not perform basic practical tasks, or comprehend the inconsistency and capriciousness of his irrational species, he routinely executed intellectual feats that would have been the capstones of any other thinker’s career: He helped to invent game theory, laid out the mathematical foundations of quantum physics, predicted how RNA would prove to communicate with DNA when the double helix was discovered a decade later and fantasized about artificial intelligence long before it materialized in its more sophisticated guises.

The MANIAC of Labatut’s title is the “Mathematical Analyzer, Numerical Integrator and Computer,” an early computer designed by von Neumann in the 1950s. But it is also von Neumann himself.

Is “The MANIAC” a work of fiction? Or do we call it fiction because we lack a better word for its creative conquest of fact? Most critics tasked with rendering Labatut recognizable liken him to the melancholic German writer W.G. Sebald, whose gently meandering novels contain long, dreamy meditations on destruction and decay. It is true that both authors toe the wavering line between invention and history, drawing on a wealth of historical and scientific details, but Labatut is that vanishingly uncommon thing: a contemporary writer of thrilling originality. Even more than “When We Cease to Understand the World,” “The MANIAC” is a work of dark, eerie and singular beauty.

It can also be difficult to read. The book is narrated by a cluttered polyphony of characters, among them both of von Neumann’s wives and a number of his teachers and colleagues. But there is a reason for this mad mumble of voices. “A perspective is by nature limited. It offers us one single vision of a landscape,” wrote the mathematician-turned-mystic Alexander Grothendieck, one of the ambivalent heroes of “When We Cease to Understand the World.” “Only when complementary views of the same reality combine are we capable of achieving fuller access to the knowledge of things.” Like von Neumann, “The MANIAC” strives to adopt the impartial standpoint of the universe.

But even the most agile human minds struggle when they reach the furthest outposts of rationality, and “The MANIAC” follows three fraught exchanges between people and icily impersonal forms of intelligence. First comes a physicist so defiantly human that he cannot stand to relinquish the neatly ordered Newtonian universe to the ravages of the quantum revolution; second comes a mathematician who can seem to be more of a machine than a man; finally, we are left with a gleaming computer capable of outsmarting its makers.

Labatut begins with the Austrian physicist Paul Ehrenfest, who could not reconcile himself to the lunacy of the 20th century. In 1933, shortly after Hitler came to power, he fell into a manic depression and shot his disabled son in the head, then turned the gun on himself. He had every reason to fear that the Nazis, who were already gesturing toward their notorious “euthanasia program,” would subject his child to an even crueler end — but in Labatut’s telling, he was also terrified of the chaos that quantum physics had unleashed upon the world. “If one were to believe the novel rules governing the inner realm of the atom,” Labatut writes, “the entire world was no longer as solid and real as it once was.” Ehrenfest could not stand to see the old adamantine certainties crumbling into dust.

Von Neumann, in contrast, was just inhuman enough to embrace the derangements of the giddy new science. Born to a wealthy Jewish family in Hungary, he had systemized the mathematics undergirding Heisenberg and Schrödinger’s quantum innovations by the time he turned 30. But he could not linger to relish his achievement. As Europe veered toward catastrophe, he fled his home country and took refuge alongside other luminaries in exile, among them Albert Einstein, at Princeton’s legendary Institute for Advanced Study.

His critics accused him of adapting all too readily to his adoptive country, with its flair for pugilistic grandstanding. Throughout the Cold War, he rented out his once-in-a-generation brain to IBM, the Rand Corp. and the CIA. Even when he was dreaming up the most fantastic devastations — designing hydrogen bombs at the Manhattan Project in Los Alamos, for instance, or trying to engineer a model that would allow the U.S. military to control the weather (one of his rare failures) — he was chillingly indifferent to the human cost of his cool calculations.

But even this unfeeling colossus would come to harbor reservations about the nascent discipline of computer science. “Technological power … is an ambivalent achievement,” he wrote with surprising lyricism near the end of his life. “It is not a particular perverse destructiveness of one particular invention that creates danger. … For progress there is no cure.” Nor for mortality, and Labatut observes that von Neumann lost the glint of machinery on his deathbed. For the first time in his life, he was “just like any other man.”

The final and most extraordinary segment of “The MANIAC” has two protagonists. The first is the South Korean prodigy Lee Sedol, a master of the enormously complex game of Go — which has so many possible game states that, to model them, computer scientists would need “more space than is available in the known universe.” The second is AlphaGo, a computer program designed by the AI research laboratory DeepMind. When the two faced off in a tournament in 2016, one onlooker scribbled in his notebook in horror: “not a human move.” AlphaGo was also decidedly not a human victor. Sedol lost four of five games, but his opponent did not gloat or boast. Strictly speaking, it did not even care.

Labatut’s characters are all agonizingly human in one respect, no matter how hard they are trying to become machines: They care immensely. They are desperate to get to the bottom of things, all the way to the entrails of reality. In “When We Cease to Understand the World,” a pair of mathematicians go insane in their quest to reach “the heart of the heart,” the fundamental substratum of mathematics, and in “The MANIAC,” Sedol painstakingly dissects each of his games, even the ones that he wins, in a frenzied attempt to understand them exhaustively. Meanwhile, von Neumann is willing to dispense with everything — decorum, ethics, love and, ultimately, sanity — to reach “the leaping point, the heart of the matter.” One of Labatut’s characters confesses that the scientists at Los Alamos were driven not by the desire to win the arms race but by a darker instinct: “the joy of thinking the unthinkable.”

To think the unthinkable is not a matter of rote mechanical calculation. Something closer to revelation is required. Labatut recounts that Georg Cantor, famed for his revolutionary research into the nature of infinity, believed that his insights were a product of “divine intervention,” and another mathematician in “The MANIAC” reflects that “mathematics is the closest we can come to the mind of Hashem.”

It is dangerous for a mortal to approach the mind of God, and it is no surprise that everyone in Labatut’s fiction is always going mad, even the figures mentioned only in passing. Cantor “struggled with increasingly strong episodes of uncontrollable mania, attacks that were followed by deep melancholy”; Kurt Gödel “started seeing ghosts, and became convinced that other mathematicians were intent on killing him.” Even the unflappable von Neumann died of cancer that eventually ate away at his brain, until at last he was raving.

The worst part of these dark tales, though, is not that reason revolts when it is pressed too far but that the world turns out to be disordered, down to its very seams. In “When We Cease to Understand the World,” the demon crouching at the end of Schrödinger and Heisenberg’s neat rows of equations is the insuperable uncertainty of quantum physics; in “The MANIAC,” von Neumann’s efforts to get to “the heart of the heart” of mathematics are thwarted by Gödel’s incompleteness theorems, which demonstrate that there is “no way to unearth the logical foundations that he so desperately wanted to find.” Tumult is not an affront to human reason but its logical culmination. The only way to transcend confusion is to transcend human cognition itself.

When AlphaGo faced off against Sedol, it was apparent that the two of them were playing different games. “AlphaGo did not hesitate and it never thought twice,” Labatut writes. “It was immune to weariness. It knew no self-doubt. It cared not for style or beauty, and it did not waste time with any of the elaborate mind games that all professional players bait each other with.” In other words, it was boring, devoid of the drama and personality with which people invest their contests and competitions. After all, who would care about the outcome of a match between two rival automatons, themselves incapable of caring? Yet more than 200 million people, in Labatut’s telling, tuned into the AlphaGo tournament to watch Sedol, a human master.

They were not disappointed. In the fourth game of the five-game series, he made a move so inventive and so unexpected that it threw AlphaGo into disarray. “It was as if Lee Sedol had just won a victory for our entire species,” Labatut writes. Onlookers were moved to tears.

Some spectators believed he had become a deity. “The hand of God! That is a divine move!” another Go champion shouted. When the team behind AlphaGo plugged Sedol’s move into the algorithm, they suggested that perhaps he had become not a god but a machine: Only “one in ten thousand people” would think to play as he had, “exactly the same probability that AlphaGo had assigned to its own groundbreaking move” in an earlier game. But when asked what he had been thinking, Sedol could only muster an appeal to that most mysterious and least rational of faculties: “sheer inspiration,” as Labatut puts it.

“Go is not a game or a sport, it is an art form,” Labatut writes. But art is limned with meaning by our efforts, our failings and our pathologies. If computers can make something resembling art — and increasingly, of course, they can — it is only an ersatz imitation. Someone should have told the team behind AlphaGo what the narrator of Kleist’s tale should have told the dancer so enamored of lifeless and unerring marionettes: The miracle is not the quality of the dance, but the flawed and fallible human being straining to perform it.