‘The Maniac’

Benjamín Labatut’s new novel, The Maniac, returns to the author’s theme of what happens when the greatest human minds confront the greatest questions of the universe. The answer: nothing good. In his careful accumulation of detail and painstakingly clear prose, Labatut fills the quest for knowledge with quiet dread. Despite taking place in the bright light of the Trinity detonation, the book reads like a horror novel, haunted by shadows and death.

The Maniac is Labatut’s follow-up to his slim breakout novel, When We Cease to Understand the World, which was shortlisted for the National Book Award and the Booker Prize. In it, the cold truths of the universe break the souls of the great minds who confront them.

The Maniac continues with that same theme, focusing mainly on John von Neumann, considered the smartest man alive by the smartest men alive. Born in Hungary, von Neumann emigrated to America before World War II and helped build the atomic bomb, then pioneered the doctrine of Mutually Assured Destruction in the Cold War. He laid the groundwork for all modern computers and wrote groundbreaking works in quantum mechanics, statistics, and game theory.

Von Neumann regularly went from room to room at the Institute for Advanced Studies and solved his colleagues’ most difficult problems in mere moments. The U.S considered his brain to be a national security asset. When he died, riddled with tumors and suffering from psychotic episodes, the country sequestered him at Walter Reed with the Secretary of Defense to scoop up the last of his ideas and to keep him from spilling any classified secrets.

Labatut is at his best when writing about von Neumann’s non-stop genius and the sheer terror and awe it provokes. He structures this part of the novel as a series of recollections about von Neumann, only letting us see him from outside. But von Neumann is unknowable, even to those closest to him. People called him an alien, an advanced intelligence hiding in human form.

And he scared the hell out of people. He fiddled with uranium to make it deadlier, advocated a first strike on the Soviets to create a Pax Americana, and blithely calculated the most efficient ways to kill people in a thermonuclear war.

But when he’s struck with cancer, Labatut portrays von Neumann as searching for a new god to replace the old ones he’s dispatched. He attempts to create a new form of computerized life in the primitive circuits of the Mathematical Analyzer Numerical Integrator and Automatic Computer Model I, or MANIAC.

As in his earlier novel, Labatut implies there is something rising to meet von Neumann and his fellow scientists. It seems as if there is an intelligence, parasitic and predatory, waiting to be born from MANIAC and all the machines it spawns.

The darkness in Labatut’s picture emerges from von Neumann’s relentless search for something that will surpass us. What is waiting, Labatut hints, is not just inhuman; it is anti-human.

The Maniac ends with a coda that should answer the questions raised by von Neumann’s work, a journalistic digression into a modern-day AI programmed to play the board game Go. Created by DeepMind, a Google subsidiary, AlphaGo utterly destroyed world champion Lee Sedol in 2019. But somehow it fails to match the existential dread of the previous two parts of the novel.

While the story is compelling–the AI reduces Sedol, considered one of the greatest ever to play the game, to a quivering wreck—it doesn’t fit, somehow. AlphaGo should be the apotheosis of von Neumann’s efforts. It’s a machine that inexplicably learns to surpass even the greatest leaps of human intuition and creativity. The construct is unliving proof of that inhuman rationality lurking like a malevolent deity throughout The Maniac.

And yet it’s simply not that frightening to watch a program ruin a single man when Neumann was playing with the lives of literally everyone on Earth. He finally reveals the dark god, and it’s a glorified video game.

Since then, of course, there have been better examples of the inhuman intelligence that we’re building: “deep learning AI’s” and “Large Language Models” (LLMs) like AlphaGo which steal millions of images and books to assemble simulated works of art from the mulch. People who can’t write or draw argue that “creativity privilege” has victimized them for too long, and they use AI to create pictures of SpongeBob piloting a plane into the Twin Towers.

What’s worse is there doesn’t seem to be any real, general intelligence behind the programs. Despite consuming billions of dollars and massive amounts of resources, AI appears, at least in its results, to be nothing more than a series of highly sophisticated decision trees. The programs still need humans to press their buttons. They have yet to create anything a human cannot. And they do not have the capacity to enjoy their own work. A deepfake video of Tom Cruise means nothing without a human audience.

Perhaps this is why AI is also wildly unprofitable, a solution in search of a problem. That seems like the true horror of the machines: we’ve taken the sum total of human achievement and ingenuity just to make the Mona Lisa with bigger tits.

Maybe Labatut will write his next novel about the Google employee who became convinced its AI had achieved sentience, or the man who fell in love with a chatbot. Those seem more suited to his themes and message. But even those stories are living examples of the pathetic fallacy, people assigning human agency to things that are not human.

The most frightening possibility in Labatut’s book is the idea that von Neumann was wrong. Despite the efforts of our greatest minds, we are still alone in the universe, and nothing will replace us, because there is nothing else. There are no gods waiting for us in the future, only canyons filled with the echoes of human voices.

(Penguin, October 3)