The MANIAC

Strange but great book. The Maniac is a work of fiction though interwoven as a story about the life of Von Neumann and eventually a picture of the early development of theories of AI alongside a broader biographical look at the man. I am not sure what I was expecting from this when I bought it as it was more just popping up as a recommendation but I haven't read a book quite like this before. For a while I was unsure of whether what I was reading was actual testimony from people in Von Neumann's life but the author notes that the book is a work of fiction. The Maniac is primarily a biography of John Von Neumann, the Hungarian mathematician/polymath involved in the formalizing of quantum mechanics, game theory, an important participant in the Manhattan project, the designer of the modern computer architecture and early student of self-replicating structures. He was an extraordinary mind that the author describes through a series of fictitious interviews with people in his life. It is an effective buy mystic way to tell the story of his life but such a style was highly engaging. The author weaves in the "perspectives" of several renowned mathematicians and physicists as well as his former spouses. The book describes Von Neumann's astounding genius for being able to solve problems with an unparalleled focus, his breath of expertise coupled to a juvenile maturity on many other matters of day to day living. Among the characters detailed in the book one hears from both Feynman and Wigner (a school friend of Von Neumann who travelled to the US as well with him), Oscar Morgenstern (his game theory co-author). The author gives time to his two wive's perspective (not sure where he got all his sources), which highlight how Von Neumann was in many ways completely incompetent but also unbelievably brilliant, both had significant eccentricities themselves. The book morphs from a description of his unparalleled mathematical abilities to his lack of sympathy or empathy for the consequences for his work. In some sense the author implicitly makes the point the authors theory of zero-sum games comes out in his policy beliefs for things like dropping the bomb. Von Neumann was carried by his goals for finding solutions rather than taking a higher perspective on what he thought about the merit of the rules. The book moves on to computation and weaves in the modern story of Deep Mind and alpha go. It goes back and forth from the story of Von Neumann and how his interests shifted to the biological domain. I never read the Computer and the Brain, but I am assuming the author is weaving a story around this later part of Von Neumanns obsession once computing was getting off the ground. Obviously this field has made enormous leaps in the last decade due to neural networks and the scale of compute and so the author highlights how yet again Von Neumann was way ahead of his time. The story of alpha go and Lee Sedol through his experience is a nice conclusion to the book and one is left with a lot to think about and the new world we are now in where our computing solutions are paralleling tasks we thought solely in the domain for humans. Of course since this book has been published we have only gone further and though one comes to respect the brilliance of Von Neumann, one also hesitates on the idea that such minds should lead people given their lack of human perspective on what objectives serve humanity's interest. This issue is much at the forefront of AI today and in some regards this is the biggest takeaway from the book and the development of the H-bomb a rhyming example of a development that serves no benefit for any person, and yet was developed for its potential to increase power rather than welfare. Highly recommended book, creatively written and served a valuable purpose of entertaining the reader while informing us of what brilliance can deliver and what its blind spots can be.

When I saw that Benjamin Labatut had a new novel, I grabbed for it immediately. His previous effort was called “When we cease to understand the World” and i found it to be one of the best books I have read this century. “The MANIAC” is similar in that it is a fictionalized biography, mainly focused on John von Neumann’s life and career. The title is from the computer that von Neumann designed and had built that was housed at the Institute for Advanced Studies at Princeton shortly after World War II. It was one of the world’s first stored program computers, a notion that Von Neumann developed, perhaps inspired by Alan Turing’s original thesis where he fleshed out what is known as a Universal Turing Machine. Stored program computers are also known as being based on a Von Neumann architecture machines. They are, of course, mainly what the world uses today. A signature achievement from a man acknowledged as a mathematical genius by friends and colleagues, a genuine force of nature, a mathematician’s mathematician. Unfortunately, I am not sure Labatut does justice to the shape of Von Neumann’s extraordinary career in applying mathematics to a dazzling array of problems, from quantum mechanics, to Game theory, to the burgeoning science of computing. Despite the title, he also fails to dive very deep into the workings of the MANIAC itself, and the effort to program it, spearheaded by Von Neumann’s relatively unheralded 2nd wife Klara, who labored in his enormous shadow. In the final section of the book, Labatut tries to make a gigantic leap from von Neumann’s very original mathematical conceptualization of self-replicating machines to more recent advances in Machine Learning that underpin the effort of the Google Deep Mind research team to build a computer program called AlphaGo that successfully challenged the best human players of the ancient Chinese game of Go using Deep Neural networks and reinforcement learning. Tying Von Neumann’s work on self-replicating machines to ML using the neural network approach is a little weak IMHO, however. I see more of a direct line from Von Neumann first to the problem of cracking the genetic code, and from there eventually to the world of Artificial Life and some of the amazing folks associated with the Santa Fe Institute. On the other hand, the lineage of the neural network approach actually runs through McCullough and Pitts, and people like Hebb, that was then taken up again by the parallel distributed processing group at UCSD, following Minsky and Papert’s withering critique of the limited computing capabilities of perceptrons. In summary, this is an excellent and thought-provoking book. But it is a notch or two weaker than its outstanding predecessor, which represents a very difficult act to follow. I should also note that some of the primary sources Labatut cites in his Acknowledgment section are definitely worth pursuing if reading this novel scratches an itch. I am thinking of the George Dyson book, the AlphaGo documentary available on YouTube, the Von Neumann biography, etc.

This new theme of fictional biography is too creative. Would definitely suggest to read if you like to read about scientific minds.

It's been a long time since a book brought me to tears. Labatut is one of the greatest storytellers out there. Check out his other 📚

ben scritto, un po' difficile per fare esercizio di inglese

Superbe ouvrage qui nous conduit avec clarté et simplicité à travers le labyrinthe des découvertes et des génies humains, humanistes et a- humanistes, qui ont révélé des lois fondamentales de la physique et ont aussi mené à en faire un usage parfois positif et sauveur ET parfois terrifiant et meurtrier

I love how it combines the history of science with fiction. Truly an amazing and very interesting format. Will definitely be reading more books from the author.