Algorithms to Live By — The Computer Science of Human Decisions

Algorithms as a Guide to Life

Algorithms to Live By begins from a deceptively simple observation: computers face the same fundamental challenges as humans — how to allocate limited time, how to sort through information, when to stop searching and commit. The solutions computer scientists have developed for these problems are mathematically optimal, and Brian Christian and Tom Griffiths argue they are also practically useful for humans facing equivalent choices.

The result is a book that is more rigorous than most popular science writing and more readable than most academic work — a tour through twelve algorithmic ideas and their human applications.

The 37% Rule

The book’s most immediately actionable idea comes from the optimal stopping problem: how long should you look before deciding? The mathematical answer — derived from the secretary problem in probability theory — is the 37% rule. Spend the first 37% of your available time or candidate pool purely observing, then commit to the next candidate who exceeds everyone you’ve seen so far.

This strategy maximises the probability of choosing the best available option and applies cleanly to apartment hunting, hiring decisions, and the question of when to stop looking for a better option in any domain. The precision of the answer is itself the point: optimal stopping has been worked out, and the answer is not “as long as possible” or “trust your gut.”

Explore vs. Exploit

A second central idea addresses the tension between trying new things and returning to what you know works. Multi-armed bandit theory — developed originally for clinical trials and casino mathematics — shows that the optimal strategy shifts with time horizon. When young and with a long future, exploration makes mathematical sense. As time runs short, exploitation of known goods becomes dominant. This gives a formal justification for why settled preferences are rational with age rather than merely conservative.

Sorting, Caching, and the Mess on Your Desk

Among the book’s most delightful turns is its application of computer-science theory to mundane domestic problems. Sorting algorithms illuminate when it is, and is not, worth alphabetizing your bookshelf — often the answer is that the effort exceeds the payoff, a liberating conclusion for the disorganized. Caching theory, which governs how computers decide what to keep in fast-access memory, turns out to justify the productivity of a messy desk: the principle of keeping recently used items close at hand is, it emerges, mathematically sound. Christian and Griffiths have a gift for these reversals, taking a reader’s guilty habit and revealing it as an unwitting implementation of an optimal strategy. The effect is both entertaining and genuinely reframing, and it is a large part of why the book reads so much more pleasurably than its rigorous underpinnings might suggest.

When Not to Optimize

One of the book’s subtler and more mature arguments is about the limits of optimization itself. Christian and Griffiths are careful to note that computing the perfect answer is sometimes more costly than the imperfect answer is worth — that “good enough,” arrived at quickly, frequently beats “optimal,” arrived at slowly. They introduce the idea of computational kindness: designing choices, for oneself and others, so that they require less agonizing calculation. This is a humane counterweight to the productivity-hacking impulse the book might otherwise feed. The lesson is not that we should run every decision through an algorithm but that knowing which problems have clean solutions — and which do not — is itself a form of wisdom. It is a more philosophically serious conclusion than the genre usually reaches.

Two Authors, Two Disciplines

The book’s unusual authority comes from its co-authorship. Brian Christian is a writer with training in computer science and philosophy; Tom Griffiths is a cognitive scientist who studies how the human mind approximates rational decision-making. The collaboration means the book is neither dumbed-down journalism nor dry academic exposition but a genuine bridge between the two — rigorous about the mathematics yet always alert to the human application. Griffiths’s research background, in particular, lets the authors connect the algorithms not just to how we should think but to how we already, intuitively, do think, since evolution has equipped the mind with rough-and-ready versions of these same strategies. That double vision — the formal and the cognitive — is what elevates the book above the many works that merely apply a single discipline’s metaphors to everyday life.

Bayes and Everyday Prediction

One of the book’s most practical chapters concerns Bayesian reasoning — the mathematics of updating beliefs in light of new evidence — and its application to the everyday problem of predicting how long something will last or how likely an outcome is. Christian and Griffiths show that the right prediction depends on knowing the underlying distribution: some quantities, like human lifespans, cluster around an average, while others, like the gross of a film or the wealth of individuals, follow power laws with no meaningful ceiling, and applying the wrong assumption produces badly wrong guesses. They also tackle overfitting — the error of building a model so finely tuned to past data that it fails on the future — and draw a humane lesson from it: that simpler rules, and a willingness not to overthink, often predict and perform better than elaborate ones. It is the rare popular-science book whose statistical lessons translate directly into better daily judgment.

Final Verdict

Algorithms to Live By is the book that makes readers want to learn more mathematics. Its central insight — that optimal decision-making has been formalised in many domains and the answers are available — is both humbling and useful.

Our rating: 4.4/5 — One of the best popular science books of its decade. Practical, surprising, and genuinely illuminating about how to think.

Reading Guides

• 12 Best Psychology Books of All Time (Expert Ranked)
• 10 Best Productivity Books (Reviewed & Ranked by Experts)
• Best Business Books of All Time: 23 That Every Leader Should Read (2026)