Nothing is more interesting to poker players and less interesting to everyone else than a breathless recounting of who bet how much with a jack and six of clubs in some game years ago. There’s an awful lot of that kind of thing in this book, which celebrates poker players as paradigmatic citizens of a global intellectual community it calls “the River”, which also counts among its inhabitants venture capitalists, crypto traders, fashionable philosophers and mild-mannered statisticians.
One such statistician, Nate Silver himself, came to public prominence as a data-driven analyst of political polls at his website FiveThirtyEight, which predicted the results of US elections in 2008 and 2012 with seemingly uncanny accuracy. But before that he was a poker player, making money especially in the nascent internet-casino business, until Congress banned online poker in 2006. That, he has said, was his political awakening.
People of the River – “Riverians”, if you will – have a rational understanding of risk (poker and blackjack players know all the probabilities) and know when to bet big and when to fold. They are likely to be geeks and libertarians. All this makes them unlike the inhabitants of a rival parallel world, “the Village”, where politicians, regulators, academics and media types live, driven by emotions and hunches, and identifiable by their “distinctly left-of-centre politics”.
The book proceeds as a rambling hagiography of Riverians as they play poker (again and again), but also bet on other casino games in Las Vegas, bet on sport (a huge business in the US especially) and bet on startup companies as venture capitalists, hoping to strike it big with a stake in the next Google. Much of this is entertaining and colourful, especially once we are introduced to the mad zoo of cryptocurrencies, “shitcoins” (currencies of zero intrinsic value, amusingly discussed) and eye-wateringly expensive NFTs (digital pictures of cartoon apes and whatnot). Silver spends several chapters on the story of Sam Bankman-Fried, the FTX crypto enterpreneur now imprisoned for his giant fraud, concluding rather charitably that his subject just had too irrational an approach to risk. “He’s a smart guy,” the author concludes, “but there are a lot of us smart guys.” Modesty is not a notable Riverian characteristic.
In its prolix and flibbertigibbet way, this smart guy’s book now swerves to large gambling-adjacent matters. We are treated to a swift description of philosophical utilitarianism, according to which the right act is the one that generates the greatest happiness for the greatest number, and some standard objections to it, while being introduced, in an interestingly sceptical way, to the modern utilitarian movement of “effective altruism”, of which Bankman-Fried was once a glowing figurehead. There is also a discussion of nuclear deterrence and game theory. “A nuclear war is a lot more likely to kill humanity than a supervolcano,” Silver writes, though this is highly debatable. A supervolcano could erupt at any time based on no human decision-making, and many supervolcano eruptions in history have been more energetic than the simultaneous detonation of the world’s nuclear arsenal would be. Many scientists think the Toba eruption 74,000 years ago, in particular, only narrowly failed to wipe out the human species.
What else might kill absolutely everybody? Why, AI, of course. Depending on whom you listen to, the people pushing AI might be making the biggest gamble in history: risking human extinction in exchange for the possibility of eliminating global poverty. (AI, the handwaving explanation goes, would hugely accelerate economic growth, which would then trickle down inevitably to the global south.) Silver speaks to OpenAI’s CEO Sam Altman, one of those who constantly claims such things while demanding $7tn to train the next generation of his models, as well as to AI “doomers” who think the whole caboodle should be shut down now, with force of arms if necessary.
As elsewhere in the book, Silver declares himself sceptical of any potential role in such matters for regulations and the law. It was lawsuits that after all “destroyed … Napster’s business”, he points out, without noting that Napster’s business was the theft of creative work done by humans – just as is, many argue, the business of current “generative” AI, which “trains” its models on pirated copies of books, or images and music scraped from the internet without permission. The New York Times’s lawsuit against OpenAI for having gobbled its online content is, Silver says, just an example of the way that “the Village and the River are explicitly at war”. This doesn’t perhaps read so excitingly in the light of a renewed threat of actual political assassinations in his own country. Perhaps Silver’s defence of a laissez-faire approach to AI is best explained by the moment in the acknowledgments in which he reveals: “ChatGPT was a significant help in writing this book.”
On the Edge is a book devoted to the art of taking a calculated risk, so it is perhaps understandable that there is never any sense that gambling might be, for some people or for the culture as a whole, a wellspring of toxicity. Isn’t it a problem that it has become impossible to watch any sporting event on commercial media without being bombarded by adverts for betting companies?
A more foundational challenge is the book’s implicit attitude to the meaning of probability itself. Probabilities are very useful for quantifying the expected outcomes of repeated occurrences such as rolls of a die or draws of playing cards. But it makes no sense to ascribe probabilities to complex one-off events except to signal the extent of our uncertainty. For the 2016 presidential election, FiveThirtyEight’s final call was that Trump had only a 29% chance of winning. Silver defends that number here by pointing out that other polls gave Trump an even smaller chance, so his 29% call, he claims was a “great forecast”. But to the degree that it was a “forecast” – a prediction of what would happen, rather than an exercise in misleadingly precise bet‑hedging – it was wrong.
Here, as elsewhere, the book is an exercise in self-justification, but what it fails to justify is the assumption that mathematics can in some way predict the course of human affairs. Silver quotes Bankman-Fried’s claim that “there was a 5% chance he could become president of the United States” as an example of the fraudster’s deranged approach to probability, but it is no more meaningless than saying that there was a 29% chance that Trump could become president or that there is a 98% chance that a superintelligence would kill all humans.
Absent a statistical model of the world that reliably predicts the mental processes and actions of everyone in it, such numbers can be no more than expressions of the strength of our faith in unknown outcomes. At one point, Silver describes himself as “someone who has been around plenty of smart nerds who didn’t know their own limitations”. Has he ever asked himself whether he’s one of them too? Politics is not, in the end, just a game of cards.