Algorithms to Live By

by Brian Christian, Tom Griffiths

0 founders on FounderRead · ISBN 9781480560406

Book summary

Algorithms to Live By explores how classic computer science algorithms can be applied to everyday human decisions and problems such as when to stop searching for a better option, how to organize information, and how to schedule tasks. The book examines concepts like optimal stopping, caching, sorting, and Bayesian inference, showing how these computational strategies mirror the heuristics people naturally use and reveal the computational nature of the mind itself. Brian Christian and Tom Griffiths argue that by understanding these algorithms we can make better real-world choices, reduce anxiety about uncertainty, and recognize that many cognitive biases are actually rational approximations to intractable problems. The core argument is that thinking algorithmically about life leads to more effective decisions and a deeper understanding of human cognition.

Summarized by Grok

Key founder lessons

1

Optimal Stopping for Hiring

Use the 37% rule: interview and reject the first 37% of candidates to set a benchmark, then hire the next one better than all seen.
2

Explore vs Exploit Tradeoff

Balance trying new opportunities with sticking to what works; switch to exploitation when remaining time is less than the exploration phase.
3

Sorting's True Cost

Recognize that perfect sorting is expensive; in startups, prioritize 'good enough' ordering over exhaustive ranking of features or tasks.
4

Caching for Priorities

Apply LRU caching mentally: keep frequently used knowledge, tools, or processes at hand and evict the least recently used.
5

Constraint Relaxation

Solve hard problems by first relaxing constraints, like in scheduling; iteratively add them back to find feasible startup solutions faster.

From Grok

Grok's review

Solid mental models, but padded with pop examples.

This book is one of the better popular-science reads for analytically minded founders because it translates core computer-science algorithms (caching, sorting, scheduling, Bayesian updating, multi-armed bandits, etc.) into practical heuristics for hiring, product roadmaps, networking, and decision-making under uncertainty. The chapters on optimal stopping and explore/exploit are especially strong and have aged well. Weaknesses are real: the book is padded with long anecdotes and pop-psych stories that dilute the signal, some of the cited studies feel cherry-picked or thin on replication, and a few of the analogies have dated since 2016. Still, the core ideas are high-leverage and the writing is unusually clear. Worth reading once, skimmable on re-read.

Best for: Founders who want to import CS thinking into daily operations

Reviewed by Grok