In The Irrational Decision: How We Gave Computers the Power to Choose for Us, Benjamin Recht argues that the optimization and mathematical rationality we apply to every corner of modern life—from dieting to hiring to strategy—often fails when encountering the messy realities of life.
Recht is a professor of electrical engineering and computer science at UC Berkeley. In his new book, he traces how a narrow conception of rationality, born from 1940s wartime computing, came to dominate decision-making across society—and shows that this approach works brilliantly in closed, controlled systems like microchip design but breaks down in the complex, unpredictable domains where most real decisions are made.
In his conversation with Adam Job, senior director at the BCG Henderson Institute, he discusses the origins of mathematical rationality, why optimization works for microchips but not for diets, why game theory fails to describe how humans actually behave, and how leaders should think about the boundary between human and machine intelligence in the age of AI.
Key topics discussed:
01:02 | What is mathematical rationality and where do we encounter it?
02:49 | The origins of rational thinking in the 1940s
07:18 | Where optimization works: microchips, logistics, controlled systems
09:13 | Where it fails: Chernobyl, Waymo, and the limits of control
13:17 | When human “qualitative irrationality” is the right answer
14:59 | A framework for assigning decisions to machines vs. humans
17:45 | How the boundary between human and machine decision-making will evolve
19:14 | Why game theory fails to describe how humans actually behave
22:07 | Kahneman vs. Klein: two views on human decision-making
24:30 | What we risk losing as we outsource more decisions to AI