Can a business model produce a massacre without a single person deciding to cause one? That question sits underneath every chapter of this fifteen-hour audiobook, and Max Fisher spends the running time refusing both of the easy answers. The villains are not cartoon executives twirling mustaches over engagement dashboards, and they are not absolved either. What Fisher builds instead is a mechanism argument: systems optimized for attention discover, without being told, that outrage holds attention better than anything else, and everything downstream follows from that discovery.
Myanmar Is Where the Mechanism Argument Lands
Fisher reported this as a New York Times technology correspondent, and the structure reflects the beat. The book moves between Silicon Valley conference rooms and the places where the consequences landed, which is a harder book to write than either half alone. The elections of populists like Trump and Bolsonaro, the spread of COVID-19 conspiracy theories, and the strife and genocide in countries such as Myanmar are presented as outputs of the same machinery rather than as separate national tragedies with local explanations.
Myanmar is where the argument stops being abstract, and it is the section that justifies the book's existence. Placing a recommendation algorithm and a mass atrocity in the same causal chain is a serious claim, and Fisher does the work rather than gesturing at it. The October 2022 five-star, from Igo Hummaan, puts the reason plainly: the sources are as unimpeachable as you are likely to find, and they include former Silicon Valley venture capitalists, Google and Facebook engineers, and people who held leadership positions inside those same companies.
Sourcing of that kind is the whole game in a book like this, because the thesis is otherwise unfalsifiable dinner-party talk. Fisher has the access, and he uses it to establish what the people building these systems knew and when they knew it. The result reads less like a polemic against technology than like an accident report, which is a considerably more damning register.
Ganim Must Carry Attribution Without Quotation Marks
Recommendation engines are difficult to write about because the reader is inside the thing being described, and the most persuasive response on the page proves the method works. The February 2024 four-star arrives at the argument through a garden composter: look at one on the internet, that reader writes, and you are pulled into ads and information about composters for days afterward, until you ignore them long enough that they finally go away. Fisher's book explains the how and the why behind exactly that experience.
Fifteen hours and fifty-four minutes is generous for a single-thesis book, and the length is spent on cases rather than on repetition. Each region gets its own reporting, its own actors and its own failure. Listening rather than reading also changes the emotional weight of the atrocity material, which arrives in the plain voice of testimony instead of in the middle of a page you can skim past. Some listeners will find that harder to take. That is the honest description, not a warning against it.
Peter Ganim carries a text with an unusual attribution burden. Reported journalism of this kind runs on quoted engineers, quoted executives, quoted survivors and quoted internal documents, and print separates those with quotation marks, block indents and attributive tags the eye takes in at a glance. A single reader has to keep all of it legible by pacing and placement alone, across material that jumps continent to continent within a chapter. That is the practical question to answer with a sample before committing to sixteen hours.
One Five-Star Review Is About a Parcel
Arithmetic on this page does not work, and it is worth spelling out before you weigh the score. The display shows 4.4 beside a tally of one rating, and five written verdicts are visible below it, four graded five and one graded four. A single vote can only be a whole number of stars, so 4.4 cannot come from one rating; the five visible grades average 4.8, so 4.4 does not come from them either. Discard the least relevant of the five and the remaining four still average 4.8. Nothing on the page produces the posted number.
The February 2026 five-star supplies the reason to discard one. That response praises fast delivery, good paper, a solid hardbound back and panels, credits the retailer for shipping speed, adds a parenthetical about bias in the reporting, and concludes that it appears to be a good read so far. It is a review of a parcel and a partly opened printed book, filed on the listing for a recording, and it carries the same weight in the average as the responses from people who finished.
Solutions are the other thing to calibrate. The description promises what we can do about it, and that promise is the weakest part of the delivery, as it usually is in this genre. Fisher is precise about mechanism and considerably vaguer about remedy, so listeners arriving for a policy program will find diagnosis where they expected prescription. The shelf tag compounds the mismatch: this is filed under content creation and social media, a category that suggests marketing advice, and nobody will use this book to grow an audience.
Required Listening for Platform Policy Professionals
Listeners who want to understand the last decade of political rupture through infrastructure rather than through personalities are the core audience. The November 2024 five-star identifies the payload accurately, pointing at the late chapters that set out how the major platforms profited from letting conspiracy theories run. That reader calls the book exhaustively researched, which matches what the reporting actually is.
Skip it if you already believe the conclusion and want the satisfaction of hearing it confirmed, because Fisher's method involves a great deal of company-by-company detail that resists slogans. Skip it also if you need a book that ends with an actionable plan. Anyone working in platform policy, moderation, journalism or education should treat it as required listening, and anyone who has watched a relative disappear into a feed will recognize far too much of it.
Whether you come out of it angrier at the engineers or at the incentives is genuinely left open, and that restraint is what makes the argument stick. Sixteen hours is a long time to spend inside an accident report, and this one earns the time by never letting the machinery become a metaphor.
Does this audiobook require any technical background to follow?
None whatsoever. Fisher explains recommendation systems through their effects rather than their code, so anyone who has ever used a feed already has the background required. The demanding material is the reporting from Myanmar and the internal company decisions behind it, neither of which asks anything technical of a listener.
Does the audiobook offer solutions or only diagnosis?
Mostly diagnosis. The description promises guidance on what we can do about it, and the remedies section is noticeably thinner than the reporting that precedes it. Expect a precise account of how the machinery works and how it was allowed to keep working, with only general suggestions about regulation and design at the end.
How graphic does the material about violence get?
Specific rather than gratuitous. Genocide, mob violence and conspiracy-driven death are documented through survivor accounts and official records, and spoken testimony lands harder than a skimmed page. One reader reports barely being able to set it down, which is fair, but the atrocity chapters are genuinely difficult listening.
