(~140k words, 4h read) Before reading, my knowledge of Hugh Everett was limited to basically the following sketch: a young American male who post-WWII suggested taking the Schrodinger wave-equation literally, yielding the infamous Many-Worlds Interpretation, and attacked over it, left academia for Wall Street where he became rich with an optimization algorithm, and in his absence, MWI very gradually gained adherents until it is now a respectable point of view (albeit still counterintuitive), and died at some point; also, some rumor that his daughter shot herself at a casino after losing, in a literal quantum suicide. This turns out to be incorrect and very incomplete: it wasn't Wall Street but the Pentagon, he died quite young, MWI wasn't attacked so much as ignored after being sabotaged, his daughter did commit suicide but it was at home with sleeping pills & had nothing to do with quantum suicide, and he did much more than just MWI & one optimization algorithm.

Byrne starts in media res, with Everett rich and drunk and self-destructing, then jumps back to his parents to start his tale; whether because 'past is prologue' or because of the heritability of personality traits, we get a sense that pathology (substance abuse, emotional problems) ran in the family, and his father survived some scrapes with corruption to finish out a reasonably good life; Everett bade fair to do better as a prodigy, excelling university, and arriving at Princeton & IAS in its golden WWII moment - the war won, von Neumann still alive & at the height of its powers (inventing game theory, modern computers, and steering the Cold War), and academia rushing into its Faustian post-war bargain with the US government and embarking on decade of exponential bloating (which, unsustainable, halted in the '80s or so, and this cauldron of legions of mediocre researchers + government funds + publish-or-perish has contributed to the modern scientific context in which we are awash in bogus results and worthless papers). An exciting time, and a fertile environment. I was surprised to learn that Everett made contributions to game theory, which turns out to later be relevant to one of the main mysteries of MWI (where the subjective or Born probabilities come from), and only then turned to quantum mechanics.

Byrne also covers his future wife, Nancy. He tries to be sympathetic, but it's hard to like or find her interesting at all; her views are shallow and deeply conformist, she comes off as lacking real insight into herself despite all the navel-gazing, lies to herself and others, and to be a lump of flesh going nowhere fast. He wants to paint her as neglected and damaged by her relationship with Everett, and to paint Everett as a loathsome lecher who won't take no for an answer, but it doesn't succeed. I was left with a major question: why would Everett ever want to date her, much less marry her? (Dating her is the real question here since it's clear why he married her: because she got pregnant and refused to abort, and given the straitlaced Pentagon world, he was put between a rock and a hard place. Byrne quotes her as denying this tactic, but that's obvious bullshit, especially given the era.)

After a jump forward to Everett's optimization work, we go back to Princeton and the genesis of MWI: like Columbus and Einstein and some others before him, Everett asked a deceptively simple question - what if we just take it literally? As a nice Schrödinger quote points out, it's odd to accept that the world or objects act like a wave-function up until they are observed and then they collapse into normality but to refuse to accept that 'inside' the wave-function it will also all add up to normality:

"Nearly every result [a quantum theorist] pronounces is about the probability of this or that ... happening - with usually a great many alternatives. The idea that they be not alternatives but all really happen simultaneously seems lunatic to him just impossible. He thinks that if the laws of nature took this form for, let me say, a quarter of an hour, we should find our surroundings rapidly turning into a quagmire, or sort of a featureless jelly or plasma, all contours becoming blurred, we ourselves probably becoming jelly fish. It is strange that he should believe this. For I understand he grants that unobserved nature does behave this way - namely according to the wave equation. The aforesaid alternatives come into play only when we make an observation - which need, of course, not be a scientific observation. Still it would seem that, according to the quantum theorist, nature is prevented from rapid jellification only by our perceiving or observing it. And I wonder that he is not afraid, when he puts a ten-pound note into his drawer in the evening, he might find it dissolved in the morning, because he has not kept watching it."

Pursuing his idea, Everett wrote his thesis, and here we run into the major theme of Byrne's book, one he establishes admirably well: with many quotes from letters and recordings and referee reports, we see Everett's thesis adviser, Wheeler, turn from a courageous physicist, well-regarded for his daring speculations, into a biased coward who bullies Everett into sabotaging & watering down his thesis so as to not give offense to his mentor Niels Bohr.

I'm a little familiar with Bohr's philosophy of science & quantum mechanics from a course I once took on the topic, and I found it entirely without merit (the most unimaginatively instrumentalist 'shut up and calculate' viewpoint was preferable to Bohr's 'complementarity', because at least one was not left with the illusion of knowledge), so to find an excellent case made that it sabotaged the initial presentation of MWI and responsible for a multi-decade drought in one of the best available interpretations... does not leave me with a good impression of Bohr, Wheeler, the power thesis advisors wield, or academic physics in general.

Certainly it is understandable that Everett would leave academia and enter the military-industrial complex where his work was interesting, valuable, valued, and well-remunerated. Everett dived straight into the heart of US nuclear politics, the intersection of nuclear physics with military strategy and game theory and computing and operations research: what levels of bombs would be developed (the Super? and even more exotic weapons?), what military services would get what delivery systems, what would be the effects of nuclear war, what was the best way to run the Cold War? (In the '50s, none of this was set in stone yet.) It's a fascinatingly complicated period, for an overview see:


Turing's Cathedral, Dyson (review)
Radiance (review)
"Old Legends"
Atomic Audit
The Making of the Atomic Bomb & Dark Sun: The Making of the Hydrogen Bomb
Fourth Generation Nuclear Weapons
Shame: Confessionas of the Father of the Neutron Bomb, Cohen (review)
Project Air Force: 1956-1996


Byrne unfortunately is too unsympathetic to cover the period fairly, taking the Dr. Strangelove route: everyone was insane and evil. This biases his coverage badly since he's so opinionated; in discussing the Prisoner's Dilemma, for example, he implies it shows the irrationality of rationality and hence the intellectual bankruptcy of game theory and all related exercises - but this is a confusion of what he would like to be true with what is actually true, because the Prisoner's Dilemma shows up again and again in all sorts of guises in the real world, along with the tragedy of the commons, and you know what? People in real life often do defect unless additional mechanisms are in place (often being put in place as a reaction to all the defecting). One of his footnotes reveals this strikingly:

In other words, rationality is a (sometimes) quantifiable quality. Most human beings would agree that it is not a rational act to cross the street in front of a speeding bus, or to poison the water supply in search of short term profit, or to depend on fossil fuels, etc. But people in power who do obviously irrational things are often compelled to rationalize these actions by falling back on agendized utility values and probability statements. Of course, if you start with an irrational premise, e.g. "nuclear war is a rational option," no amount of utilitarian quantification can, believably, turn it into its opposite. Context is everything.

This is a tissue of nonsense which exposes clearly that Byrne does not deal with the real world, but with a world of ideals in which there are never any hard choices or necessity to make cost-benefit tradeoffs and all that matters is what sounds good. Accordingly, he presents a one-sided picture; a discussion of the Bohm hearings omits any mention of why the US government might be so paranoid and worried about spies (the Venona decrypts come to mind, as do the many high-ranking Soviet spies such as Harry Dexter White) and might target people involved with the Manhattan Project in particular; similarly, he uncritically cites Sakharov claiming the US was responsible for the arms races (which seems like an odd reading of Stalin's character and his fellow researchers, for that matter), and later overestimates of nuclear winter. This bias on the biographer's part makes one wonder to what extent Everett's results about fallout were accurate: it's not like he would tell us if the report was found to be fallacious or since debunked. Still, while irritating and depriving the reader of some key context, the WSEG section seems comprehensive as far as it comes to Everett up until he left the Pentagon to start his own consulting business, and that's what really matters.

The business section is similar, but much less political as they consulted on more civilian topics. What he did is hard to tell: we're held back by Byrne targeting the general audience - I would have liked to know more about the statistical techniques involved, rather than vague descriptions like "QUICK randomly sampled the vast range of probable outcomes to select the most probable results", which could mean a lot of things; I can sort of guess what his 'Bayesian machine' was (sounds like a Kalman filter implemented with MCMC), but I'm completely baffled by the section about '"attribute value" programming' or what sort of database it was. It also sounds like Everett began drinking himself to death at this point (but why? he doesn't come off as so deeply depressed about MWI being ignored that he'd be suicidal in the midst of all his financial success; given Byrne's predilection for psychologizing, it's odd that he seems to let this central mystery pass without much more comment than some speculation that Everett was just hedonistic), and the kids enter their troubled teens (but one would never grow out of it). Somewhat surprisingly, he didn't manage his finances very well, living extravagantly, making deeply questionable investments, and failing to diversify, all in contravention to established financial advice, flaws somewhat surprising in a statistically and economically inclined man. Eventually, he dies.

In the mean time, MWI was gradually being rediscovered and rehabilitated by the likes of Deutsch and novel approaches like a Bayesian justification of Born probabilities developed, leaving off at the present time in which MWI is a respectable position leading to interesting research and believed in by a good-sized minority of physicists; this is interesting, but already familiar to me. I will have to leave it to other readers to judge how good these parts of the book are.

Overall, indispensable to anyone interested in the man, and a good account of a productive yet wasted life.