Weapons of Math Destruction

I picked this up expecting a math book and got something closer to an investigation, which was a pleasant surprise given how much of my adult reading skews toward things that require a coffee just to get through a chapter. Cathy O’Neil doesn’t need you to understand the formulas behind the algorithms she’s writing about, she needs you to understand what those algorithms do to actual people once they’re deployed at scale, and she builds that case one concrete example at a time.

The thing that makes this land is the specificity. It’s not a general “algorithms bad” thesis, it’s a working definition of what makes a mathematical model a genuine weapon, opaque, unaccountable, operating at scale on real people’s lives, and then chapter after chapter of exactly that pattern showing up in hiring, credit, policing, insurance. Once you’ve got the framework in your head you start seeing it everywhere on your own, which is either the sign of a good book or a sign I need to close the laptop more often.

If I’ve got a real complaint it’s that a few of the arguments lean on the reader accepting a political read of the data without always spelling out the counterargument first, and I noticed myself wanting a little more of the other side represented before the conclusion landed. That doesn’t undo the value of the actual reporting though, and the reporting is the strongest part of the book by a wide margin.

I’d put this on the shelf next to anything about how the internet actually works behind the scenes, because it’s really a book about power dressed up as a book about data. Should be required reading for anyone who works anywhere near this stuff professionally, and I mean that as a genuine recommendation, not a backhanded one.