True Enough: “The MANIAC” and “Oppenheimer”
Benjamín Labatut | The MANIAC | Penguin Press | October 2023 | 368 Pages
When We Cease to Understand the World, the first of Benjamín Labatut’s works to be published in English, ends with a discussion of how to assess the age of a lemon tree, to know how long it has left to live. “He told me that there was no way to know, at least not without cutting it down and looking inside its trunk,” Labatut’s narrator reports. “But, really, who would want to do that?” The preceding pages of the “nonfiction novel”—a fiction that draws extensively on real people and events, so much so that it’s unclear where the fiction actually begins—document the lives of scientists and mathematicians like Werner Heisenberg and Alexander Grothendieck who, with their revolutionary theories of the logic (or illogic) undergirding matter and reality, attempt to cut through the world’s trunk and are haunted by what they find inside. Labatut’s latest, The MANIAC, is another nonfiction novel, its source material the life and work of John von Neumann, the twentieth-century mathematician and scientist who lived hatchet in hand and whose unceasing hacks gave us, among many other things, digital computers, game theory, and Mutually Assured Destruction, and led us to our nascent, ominous age of artificial intelligence. But while When We Cease reads like a collection of ghost stories, The MANIAC is more akin to true crime, asking how—and why—this happened.
Formally a triptych, The MANIAC follows the development of artificial intelligence—first as an idea at the beginning of the twentieth century, and then as a practicality at the beginning of the twenty-first—through the lives of three men who faced it. The novel opens with a narrative of the life of physicist Paul Ehrenfest, who is disturbed—to the point of despair—by the developments that have so fundamentally altered his beloved field: “The strange new rationality that was beginning to take shape all around … a profoundly inhuman form of intelligence that was completely indifferent to mankind’s deepest needs.” His story ends when he kills himself and his teenage son. The book’s final section is an account of the 2016 Go match between Lee Sedol and the computer program AlphaGo. The bulk of Labatut’s novel, then, comprises its second section: JOHN or The Mad Dreams of Reason, which traces John von Neumann’s life, from his childhood as Jancsi in Budapest to his death in the United States at age fifty-three.
Each chapter of this second section is narrated in the first person by a different witness to von Neumann’s life (aside from the ominous omniscient interstitials separating some of the chapters). From friends and teachers we hear of von Neumann’s time in university (technically universities plural, where he earns degrees in chemical engineering and mathematics simultaneously) and then of his quest to “find the purest and most basic truths of mathematics.” This is followed by colleagues who describe his ultimate disillusionment with pure theory and his move to the United States, where he changes his name to “John” and puts himself at the service of power and capital. He works for RAND, Princeton, the CIA, the Manhattan Project, anyone who asks him a question, more or less, as long as they also give him the time and money to pursue his own research interests. In the U.S., his obsession with computers and with one of Alan Turing’s final, most outlandish theories, a machine that thinks beyond logic like a human being, leads him to develop the theoretical underpinnings of Mutually Assured Destruction and lay the academic groundwork for artificial general intelligence. The collection of voices that makes up this section is caught between telling the story of a promethean Great Man of Science and undercutting such a tale with the human reality behind it; von Neumann’s life ends as he is sequestered by the U.S. military desperate to scavenge any last theory or calculation from his mind while Johnny himself dies in agony, “just like any other man.” There are humanizing details—young Jancsi is recalled bringing two books into the bathroom, “for fear that he might finish the first one before he was done”—but the novel is foremost concerned with the development of a weaponized intelligence that produces utmost awe in his colleagues, and reverent fear in the people who know him best.
Labatut’s decision to structure von Neumann’s story this way is what provides the novel with its true crime feel; taken together, these short first-person chapters give the impression that we are hearing witness testimony, fractured accounts trying to explain von Neumann and what he wrought. The resultant portrait is of a man hellbent to transcend the limits of human intelligence, and along with it, human accountability. An early observer recalls how von Neumann “would pounce on a subject, strip it down to its bare axioms, and turn whatever he was analyzing into a problem of pure logic”—qualifying this to the effect that von Neumann’s “otherworldly capacity to see into the heart of things … was not merely the key to his particular genius, it was also the explanation for his almost childlike moral blindness.” If there is a question, von Neumann hunts for an answer, whatever it may mean, and he seems destined to solve the most unsolvable mathematical problems until his dreams of a pure natural logic are crushed by Kurt Gödel—the only time he is ever bested intellectually—who reveals the illogic fundamentally inherent in all logic. Von Neumann is “changed” after his run in with Gödel, and fleeing the Nazis in Europe, takes work at the Institute for Advanced Study in the U.S., a country that, according to his second wife, the mathematician Klára Dán, “[does] something to him.” Eugene Wigner, von Neumann’s friend and Nobel Prize-winning physicist, confesses: “From Gödel onward, I was always afraid for him, because once he abandoned his juvenile faith in mathematics he became more practical and effective than before, but also dangerous.”
There is plenty of such foreboding in these remembrances, meant to be damning. We learn that von Neumann not only devised the method of explosion for the first atomic bomb, but calculated the “optimal height” for detonation. “And that,” Richard Feynman reports, “is exactly how high our bombs were when they exploded above … Hiroshima and Nagasaki.” While working on the atomic and then hydrogen bombs, von Neumann becomes infatuated with computers, and we are told that, “the race to build the bomb was accelerated by Johnny’s desire to build his computer, and the push to build the MANIAC was hastened by the nuclear arms race.” The implication being that the MANIAC, the machine von Neumann designs—Mathematical Analyzer, Numerical Integrator, and Computer—was born alongside the “most destructive of human inventions.” Labatut goads us into wondering whether the former may one day usurp the latter, and von Neumann’s work with computers assumes an apocalyptic tone: he becomes convinced that “if our species was to survive the twentieth century, we needed to fill the void left by the departure of the gods, and the one and only candidate that could achieve this strange, esoteric transformation was technology.”
Even the ostensibly humanizing anecdotes do little to help von Neumann’s case. Wigner shares how the Jancsi he knew as a boy “could not understand how he learned to ride a bicycle … how could his body think for itself?” He is both disconnected from his body, floating above in the ether of thought, and trapped by it; in a less esoteric register, Klára tells us that her husband was frequently found “peeking underneath the desks of the secretaries at the institute.” It’s the women in the novel—Klára, his first wife Mariette, the textile artist Lydia Davis—who seem most able to see past the glare of von Neumann’s intelligence. “He was very insecure about his legacy,” Klára tells us (as if any supposed Great Man isn’t). Davis, at an Institute dinner, is horrified to hear von Neumann apply his game theory to nuclear war: “It was reckless men like Jancsi who could not think past mathematics and see the real world inhabited by real human beings, who would be the death of us all.” Von Neumann dismisses her condemnation with his now-famous quote: “I’m thinking about something much more important than bombs, my dear. I’m thinking about computers.” This single-minded callousness is characteristic of Labatut’s von Neumann, and the examples accumulate throughout the narrative. And so, when von Neumann tells Feynman at Los Alamos, “You don’t have to be responsible for the world that you’re in,” it’s clear we are meant to take it as an admission of guilt.
•
“You don’t get to commit sin and then have us feel sorry for you because there are consequences,” Kitty Oppenheimer tells her husband, the scientist and titular protagonist of Christopher Nolan’s latest film. She’s not talking about the atomic bomb in this specific instance, but there is little in Oppenheimer that doesn’t refer back to the question of culpability for the mass death and destruction wrought by the bomb, and in contrast to Labatut’s von Neumann—in the novel, such concerns are left for others—Nolan’s J. Robert Oppenheimer, played by Cillian Murphy, is guilt-ridden and remorseful throughout, at least ostensibly. (Oppenheimer appears in Labatut’s novel only cursorily, but in one of the novel’s New Mexico scenes, Feynman tells us the father of the bomb “felt he had blood on his hands;” in Nolan’s film, Oppenheimer imagines much more than just blood as the result of his pursuit of scientific invention and application.) But like The MANIAC, Nolan’s film also takes the form of a trial, albeit more explicitly than the novel; not a war crimes trial, but a hearing over Oppenheimer’s security clearance, which his past associations with radical politics has jeopardized. The film is structured around this hearing, with the plot of the film recalled through—like the voices recounting von Neumann’s life—witness testimony. We hear from witnesses testifying for and against Oppenheimer’s character, and much from Murphy’s tortured Oppenheimer himself, and so although the film repeatedly emphasizes that it is not a trial—it’s a hearing—it begs a verdict on the man who built the bomb. Is his regret enough? Are visions of nuclear winter penitence enough? Should we admire his principled reversal and refusal to work on the hydrogen bomb in light of the petty yet hawkish motivations of those in power, like Robert Downey Jr.’s Lewis Strauss? (Strauss appears in The MANIAC, too, on more agreeable terms; in the novel, Strauss gives von Neumann’s eulogy, probably because the mathematician proves useful to him, unlike the film’s ambivalent Oppenheimer.) The film, to its credit, seems unsure of Oppenheimer’s redeemability. But our understanding of the case before us is biased by Nolan’s close commitment to seeing the events through Oppenheimer’s eyes. In The MANIAC, we hear from everyone but von Neumann; in Oppenheimer, it’s mostly Oppenheimer on the stand.
The montages of witness testimony both works employ is practical—it’s difficult to dramatize thinking, and so we take others’ word for these scientists’ impressiveness—but also produces a narrative effect; in both the novel and the film, the story takes on momentum until it feels inevitable. We are hurtled from scene to scene, witness to witness, with rarely a moment of decision to suggest that things could’ve turned out any other way. There’s something like destiny in these scientists’ equations, and both The MANIAC and Oppenheimer are rife with prognostications of inevitability. An early chapter in The MANIAC has von Neumann’s brother Nicholas Augustus remembering the first time Jancsi encounters an automatic loom brought home by their father; Jancsi dismantles it to learn how it works to decipher it’s “language,” and Nicholas is unsettled by what the experience does to his brother: “Somehow, upon seeing the loom, he suffered a vague yet intense foreshadowing of the future.” Wigner, too, notes how von Neumann “behaved as if he was looking back at things that had already happened.” Oppenheimer, in an early sex scene from Nolan’s film, before the atom is even split, recites the “I am become death” line from the Bhagavad Gita mid-coitus—already, pre-bomb, he has become death. It’s not just the lives of these scientists that are inevitable, but the work itself as well. Mariette recalls how the young von Neumann was fascinated by childish war games, which he of course always won, because the “outcomes were determined by mathematical calculations.” Feynman, after watching one of the bomb tests, comes right out and says it: “It was as if those awful things had a will of their own, as if they answered to another power, a strange inevitability that makes me shudder if I think about it too much.”
This pretense of inevitability complicates our verdict. We can’t help but notice the others caught at the crime scene. Von Neumann’s work, however revolutionary, requires the efforts of George Cantor, Bertrand Russell, Oskar Morgenstern. Oppenheimer’s Los Alamos is peopled by scientists like Klaus Fuchs and Edward Teller, integral to a successful Trinity Test. And so we have to think: if not von Neumann or Oppenheimer, someone else. In its first half, before it turns to Oppenheimer’s friction with his own government, Nolan’s film is a race against Heisenberg and the Nazis to develop the bomb, the question not who could get there, but who first. In Labatut’s novel, a teacher says of von Neumann’s legacy: “We mustn’t forget that it was not just he who played with fire. His entire generation set loose the hounds of hell.”
In one sense, the story is inevitable—these are fictionalizations, to whatever degree, of a past that has already occurred. But such inevitability is not confined to the past, nor to fiction. Ever-hastening momentum and unavoidable conclusions are also invoked in real-life present-day discussions of the future—just look at the fervor over the latest advancements in artificial intelligence. “I don’t know if we can be trusted with such a weapon,” Oppenheimer says in the film, putting down a rebellion in the ranks at Los Alamos, “but we have no choice.” On ABC news, Sam Altman, CEO of OpenAI, lamented that, “There will be other people who don't put some of the safety limits that we put on,” just ahead of his company’s release of its latest model. “Godfather of AI” Geoffrey Hinton “console[s]” himself with the belief—the fact?—that if he hadn’t done the work, “somebody else would have.” Everyone seems to agree that the technology is outstripping any attempts to regulate it and will continue to do so; ominously, in its overview of potential AI regulations, the New York Times notes that, “Historically, regulation often happens gradually as a technology improves or an industry grows […] Sometimes it happens only after tragedy.”
But potential, or likely, or unavoidable misuse of AI technology also serves, for those doing the work, as justification to continue. Amid gestures at accountability and responsible development are warnings that we—really, they—must proceed lest someone with worse intentions or looser inhibitions get there first. Yet to speak of “moral” obligations and “voluntary” guidelines for AI companies is as incredible as watching Oppenheimer confess that his misgivings began only when it became clear to him that “we” intended to use any weapon he developed. More honest, to my ears, is von Neumann after the Trinity test:
What we are creating now is a monster whose influence is going to change history … But it would be impossible not to see it through. Not only for military reasons, it would also be unethical, from the point-of-view of scientists, not to do what they know is feasible, no matter how terrible consequences it may have.
Reading Labatut’s nonfiction novels is an exercise in figuring out what is true, what isn’t, and how much it matters either way. The final section of The MANIAC is a narrative of the five-game Go tournament played between Lee Sedol—at the time the best Go player on the planet—and DeepMind’s AI program AlphaGo, designed specifically to master Go, a game until then considered out of reach for a computer program. Not subtle with its man versus machine dramatics, it is a scene from the world that von Neumann’s work conjured, “the first glimmer of a true artificial intelligence … a glimpse of the future that is rushing wildly toward us.” There were many details I wanted to confirm the veracity of, some easier—were Elon Musk and Peter Thiel really among the first backers of DeepMind?—than others—was Demis Hassabis, founder of DeepMind, actually driven by his belief that “our monkey brains had taken us as far as they could”?—but none more so than Labatut’s description of the fourth game of the tournament, when Sedol, after losing three straight, none of them particularly close, plays a move so inspired, so human, that AlphaGo becomes “delusional … suddenly losing all sense of position and value.” When AlphaGo resigns the match, Labatut writes: “It was as if Sedol had just won a victory for our entire species.” Reading it, that’s how I felt too; I wanted so badly to believe Sedol’s win was true and not some consolation of fiction. I resisted checking whether Sedol had, in real life, won the fourth game against AlphaGo. I did not want to risk it being untrue. I wanted so badly to have hope that the inevitable future had not arrived quite yet.
But I looked it up, of course, and it was thankfully, blessedly true.
“Winning this one time … it felt like it was enough,” Labatut’s Sedol says after his win, before going on to lose the fifth and final game. “One time was enough.” We’re left with the question of whether or not that’s true—something that, unlike Sedol’s win, is unverifiable. But von Neumann, in the novel, turns pessimistic as he nears death.
In one of his final letters, he writes: “Technological power as such is always an ambivalent achievement … It is not the particularly perverse destructiveness of one specific creation that creates danger. The danger is intrinsic. For progress there is no cure.” Here, again, culpability is offloaded onto the very idea of science itself: unfeeling, inhuman, inevitable. These warnings do not allow for human ingenuity or variety of thought; they cannot imagine another way. But it’s a poor craftsman who blames his tools. In their trial-esque trappings and verdict-begging narratives, stories like The MANIAC and Oppenheimer may provide individual retribution but offer no suggestions for collective restraint; technological progress continues unabated. Labatut’s novel and Nolan’s film raise the question of how much personal remorse matters once what’s done is done—Oppenheimer’s bomb was dropped and von Neumann’s artificial intelligence is either here or on its way, regardless of how either man feels about the ramifications of his work. Oppenheimer’s hallucinatory regret only comes after the fact. But what of before? The witnesses can tell us how we got here, but not how to avoid where we’re going. “Individual conscience lies too close to home, and is archaic,” the naturalist Loren Eiseley writes in The Firmament of Time. “It is better, we subconsciously tell ourselves, to speak of inevitable forces beyond human control.” By better, of course, he means easier.