I had the chance of reading an advance copy of Benjamin Labatut’s latest novel “The Maniac”, published by @Penguin Press and now available in bookstores. What a book.
Labatut takes us on a roller-coaster of a journey in the history of artificial intelligence, from physics’ existential crisis amidst the quantum revolution to the first stabs at designing self-reproductible automats in the 1950s and ending with the crushing dominance of an AI programme over the world champion of go, one of the most complex games ever invented by humans, less than a decade ago. I recommend “The Maniac” warmly. Equally thriller, philosophical essay, and history book, it is one of the best novels about science and technology I have read recently.
And no, you are not getting a 15% reduction if you mention this blog on the publisher’s buying platform!
The book opens with the tragic story of Paul Ehrenfest (1880-1933). Ehrenfest was venerated among European scientists and regarded as the great inquisitor of physics (he mediated between Niels Bohr and Albert Einstein during their famous disputes about quantum mechanics). By 1931 however a disheartened Ehrenfest confessed feeling “in blind panic” seeing how physics was evolving – like an exhausted dog running after a streetcar taking his master out of sight, he wrote to Bohr. In 1933, terrified by Nazi repression against Jews, Ehrenfest shot Vassily, his ten-year-old son who suffered from Down syndrome, then turned the gun against himself.
The bulk of the book is dedicated to the central figure of John von Neumann (1903-1957), the polymath and pioneer of ground-breaking works in physics, mathematics, computing, and economics. A man greater than nature, whose life Labatut tells through the eyes of people who knew and worked with him – a good number of whom were remarkable scientists themselves.
The pages dedicated to von Neuman’s involvement in the Manhattan Project, as told by Richard Feynman and others, are gripping. While the immensity of the deflagration of the first nuclear bomb in Los Alamos was celebrated with alcohol binges by the scientists involved, the destructive brutality of the first hydrogen bomb a decade later left the witnesses with the feeling that something “unspeakably wrong” had been achieved, Labatut has Feynman say.
John von Neumann’s capacity to turn a comment heard at a conference or someone else’s intuition into new discoveries and theoretical advances seemed boundless. “He wasn’t a man who sat down to think, he was thinking continuously”, Oskar Morgenstern says in the novel. Morgenstern worked with von Neumann on the massive 700-page foundational “Theory of Games and Economic Behavior” (1944), one of the most influential treaties of economic theory of the last Century. The intensity of that collaboration eventually left Oskar Morgenstern completely exhausted, alienated from his family, friends, and colleagues, and ecstatic. “I felt as though I’d touched the holy grail”. For Neumann, however, it was “just one more thing, another achievement in a life that was choke-full of them”.
The number of luminaries and exceptional minds that rubbed shoulders with John von Neumann and that we meet in “The Maniac” is impressive. Of them, only one, however had the upper hand: Kurt Gödel, the developer of the theorems of incompleteness. And he did that twice: first in a conference in 1930 as he shyly expressed what soon after became his first theorem of incompleteness (namely that we can postulate within any consistent formal system an unprovable statement, in other words one that is true but can never be proven within the rules of that system); then a few weeks later when after a particularly intense period of work, von Neumann thought he held a more definitive contribution to theoretical logic than Gödel’s (namely that a complete system could never be consistent). He thought he had outsmarted Gödel … only to learn that Gödel had already come to the same conclusion himself – it became his second theorem of incompleteness – and published the results. Gödel “broke something in him”, Labatut writes. The dream of freeing mathematics from paradoxes and inconsistencies was over, and von Neumann stopped working on theoretical mathematics after that.
The novel’s title, “The Maniac”, refers to an early computer whose powerful new possibilities fascinated von Neumann. It could, however, equally describe the man itself and many others that we meet in the novel. Several of them shared a passion for chess and it is probably not surprising, although bewildering, that the Los Alamos scientists programmed an early AI chess programme for the Maniac and played against the computer when they were not working on the bomb (they had to discard the bishops from the programme to make it simpler).
The understanding of advances in artificial intelligence, measured as a machine’s ability to play games and win against humans, has since flourished. The novel’s third and final part renders in breath-taking details the 2016 five-game match between Lee Sedol, the undisputed human world champion of go, and AlphaGo, the AI programme developed by DeepMind. AlphaGo won four out of the five games and in the process questioned the value of the long-developed body of knowledge of the game carefully passed down through generations of players with beautiful and mysterious proverbs like “Never try to cut bamboo joints” and “Don’t make empty triangles”. Whereas AlphaGo had digested millions of recorded human games to beat Sedol, its developers eventually realised that letting the algorithm play against itself, without the baggage of human experience, made it even stronger.
One closes Benjamin Labatut’s fascinating novel with the impression of understanding the world and dreams of the pioneers of artificial intelligence better. That is welcome as we seek to shape the future of AI to support the betterment of science worldwide.
Photo by Fey Marin on Unsplash
