Start free trial
Trending New Genres Authors
EnglishEnglish
EspañolSpanish
简体中文Chinese
繁體中文Chinese (Traditional)
FrançaisFrench
DeutschGerman
日本語Japanese
PortuguêsPortuguese
ItalianoItalian
한국어Korean
РусскийRussian
NederlandsDutch
العربيةArabic
PolskiPolish
हिन्दीHindi
Tiếng ViệtVietnamese
SvenskaSwedish
ΕλληνικάGreek
TürkçeTurkish
ไทยThai
ČeštinaCzech
RomânăRomanian
MagyarHungarian
УкраїнськаUkrainian
IndonesiaIndonesian
DanskDanish
SuomiFinnish
БългарскиBulgarian
עבריתHebrew
NorskNorwegian
HrvatskiCroatian
CatalàCatalan
SlovenčinaSlovak
LietuviųLithuanian
SlovenščinaSlovenian
СрпскиSerbian
EestiEstonian
LatviešuLatvian
فارسیPersian
മലയാളംMalayalam
தமிழ்Tamil
اردوUrdu
Searching...
SoBrief
Algorithms to Live By

Algorithms to Live By

The Computer Science of Human Decisions
Know when to stop searching, what to prioritize, and how to decide with incomplete information.
by Brian Christian 2016 368 pages
4.12
35k+ ratings
Listen
Immersive
V2.3
Amazon Kindle Audible
Try Full Access for 3 Days
Unlock listening & more!
Continue

Key Takeaways

When options vanish, spend the first 37% learning what good looks like.

The search has a shape. In the secretary problem, candidates arrive one at a time, cannot be recalled, and can only be ranked against those already seen. To maximize the chance of choosing the single best, first observe about 37% of the pool without choosing. Then accept the first candidate who exceeds everyone in that initial sample. The same structure appears in apartment hunting: early viewings establish a baseline, while later ones demand a decision.

The odds stay imperfect. Even this optimal method finds the best candidate only about 37% of the time. Its value is not certainty, but a reliable balance between learning too little and searching until every opportunity is gone. The rule changes if candidates can reject offers, return later, or be measured against an objective score.

Analysis

The striking lesson is that a sound process can still produce a disappointing result. That distinction helps counter hindsight bias, the tendency to judge a decision by its outcome rather than by what was knowable at the time. The 37% figure is useful as a model, not a life law: real searches have uneven options, changing priorities, and costs for waiting. The practical move is to decide in advance when information gathering ends, then avoid quietly extending the search whenever commitment feels uncomfortable.

Use long horizons to explore, then spend short horizons on proven favorites.

Discovery needs runway. The explore/exploit tradeoff is the choice between trying unfamiliar options to learn their value and repeatedly choosing what has already worked. A new restaurant may be worth testing early in a long stay, because a good discovery can reward many future visits. Near departure, returning to a known favorite is more sensible: there is little time left to benefit from new information.

The horizon changes behavior. The book connects this pattern to life stages. Children explore widely while adults provide for them; older people often narrow their social circles toward close relationships. Experiments described in the book found that young and older adults changed their social preferences when asked to imagine their time horizon changing. The key variable is not age alone, but how long there is to use what exploration reveals.

Analysis

This framework makes novelty a form of investment rather than a virtue in itself. Exploration has a cost now and a possible return later, so a short horizon naturally favors reliable rewards. The idea also explains why constant novelty can become exhausting: discovery without time to enjoy discoveries is unpaid research. Still, horizons are often uncertain, and people differ in how much they value surprise. A practical balance is to protect some time for deliberate exploration without treating every familiar pleasure as a failure of imagination.

Sort only when repeated searches repay the effort of making order.

Order costs time. Sorting is useful when it makes future searching faster, but it is not automatically worth doing. For a computer, comparison-based sorting has a lower bound of roughly n log n operations; simple methods can take n squared. In ordinary life, the calculation is less dramatic but similar: alphabetizing a bookshelf may take hours to save seconds per search, while an enormous online index is searched constantly and benefits from extensive preparation.

Know what the ranking means. A tournament bracket can efficiently produce a winner, but it cannot reliably rank every team, because defeated teams stop playing. Lewis Carroll made this point about tennis prizes in the nineteenth century. A ranking method is only as useful as its purpose: choose a quick winner, or spend more comparisons to learn the full order.

Analysis

The chapter treats tidiness as an investment decision, not a moral achievement. That reframing is useful wherever people confuse visible order with efficiency. It also reveals that a ranking can hide what it cannot measure: a champion is not necessarily the second-best team, and an algorithmically prominent result is not necessarily the most meaningful one. The caveat is that search costs include more than seconds. Some people value the calm of an orderly space, so the best sorting decision may include psychological benefits the computational model leaves out.

Keep frequently used things nearby, and retire what experience says you rarely need.

Recency predicts reuse. Computers use caches, small, fast storage areas that keep likely-to-be-needed information close at hand. When a cache fills, the Least Recently Used approach removes the item that has gone longest without being accessed. This works because recently used information is often needed again soon. The same logic can guide a closet: place exercise gear near the door if it is used before leaving, and move rarely used belongings to slower storage.

Arrange for actual behavior. Economist Yukio Noguchi kept files in one box and returned each used file to the front. The newest and most recently used documents therefore became easiest to find, without sorting them into elaborate subject categories. The system is not conventional neatness. It is a simple arrangement that adapts as use changes.

Analysis

The useful shift is from asking where an object belongs in an abstract taxonomy to asking where it will be needed next. That makes organization responsive to behavior rather than appearances. Recency, however, is only a proxy for future value. A passport may be rarely used but costly to lose, while an old file may be legally important. The strongest application combines proximity for frequent use with a separate safeguard for items whose importance is high even when their access rate is low.

Choose your scheduling rule only after naming what counts as success.

The metric determines the order. If the goal is to minimize the worst overdue task, schedule by earliest deadline. If the goal is to clear small tasks quickly, do the shortest first. If tasks differ in importance, divide importance by completion time and prioritize the highest ratio. These rules can point in different directions because they optimize different outcomes. A tidy to-do list, fewer late jobs, and maximum total value are not the same goal.

Dependencies can disguise urgency. On NASA's Mars Pathfinder, a low-priority task held a resource needed by a high-priority one, allowing medium-priority work to delay the critical task. Engineers fixed the priority inversion by temporarily raising the blocking task's priority. Human work has similar dependencies: the apparently minor step that unlocks an important project may need attention first. Frequent task switching also carries a real cost.

Analysis

Scheduling theory's most practical warning is that productivity follows the scoreboard. An inbox badge encourages people to reduce the message count, even when the consequential work is elsewhere. Naming the metric exposes that hidden pressure and makes tradeoffs discussable. Yet human priorities cannot always be reduced to a single ratio: care, fairness, and uncertainty resist precise weights. The algorithms are most valuable as diagnostic tools. They reveal what a chosen rule will sacrifice, making it easier to choose deliberately rather than letting a default interface choose on one's behalf.

Predict from the pattern behind the data, not from one observation alone.

Evidence needs a prior. Bayes's Rule updates an initial expectation, called a prior, using evidence. One win does not prove a process always succeeds; its meaning depends on what success rates were plausible beforehand. Laplace's simple estimate for a process with w wins in n attempts is (w + 1) divided by (n + 2), which moderates extreme conclusions from tiny samples.

Different patterns call for different forecasts. The book distinguishes events clustered around a typical scale, such as human lifespans, from power-law events such as movie grosses, where a few outcomes are enormous. For a typical-scale event, forecast toward the average. For a power-law event, the current size helps indicate the likely scale. A memoryless process, such as independent waits between events, calls for a steady estimate rather than assuming a long wait means a result is due.

Analysis

The central contribution is a reminder that small samples can still support useful judgments when experience has supplied a rich background model. That background is often implicit, which makes it powerful and hard to inspect. A forecast about bus arrivals draws on a different pattern from one about a person's lifespan. The danger is that media attention can distort those priors by making rare events feel common. When a judgment feels obvious, it can help to ask what kind of process is being assumed and where that assumption came from.

When measurements are noisy, stop adding factors before you fit the noise.

A perfect fit can mislead. Overfitting occurs when a model matches the data it has seen so closely that it performs poorly on new cases. In a marriage-satisfaction example, a nine-factor curve traced every observed point but predicted implausible swings beyond the study period. A simpler two-factor curve fit less perfectly and made a more credible forecast. Extra detail can capture noise as if it were a real pattern.

Test generalization. Cross-validation holds back some data to see whether a model predicts observations it was not built around. Regularization adds a penalty for complexity, discouraging a model from relying on too many weak factors. The same risk appears outside statistics: a business metric may reward easy cases rather than important ones, and training can teach a narrow routine instead of a transferable skill.

Analysis

This insight explains why additional effort can reduce accuracy when the target is only imperfectly measured. It also clarifies the appeal of simple heuristics: they can be robust precisely because they refuse to chase every fluctuation. Simplicity is not automatically wise, however. A crude model can miss real structure, just as a complex one can mistake noise for structure. The appropriate level of detail depends on the quality of evidence and the cost of error. Independent tests are especially valuable when the metric is merely a stand-in for what people actually care about.

Relax the problem before demanding its impossible perfect answer.

Some searches explode. The traveling salesman problem asks for the shortest route through a set of cities, visiting each once. Trying every route grows factorially, so even ten cities produce millions of possibilities. Computer scientists classify problems by whether they can be solved efficiently; many practical optimization problems have no known efficient method for finding the exact best answer.

Make a simpler version useful. Constraint relaxation temporarily removes a rule. For route planning, allowing repeated visits creates a minimum spanning tree, a fast lower bound on the true route. Continuous relaxation replaces whole-number choices with fractions, which can then be rounded into a usable approximation. Lagrangian relaxation turns a strict constraint into a penalty. These methods do not conjure perfection, but they can bound the answer or produce a close, practical solution.

Analysis

Relaxation is more than lowering standards. It is a disciplined way to learn from a simplified version of reality, then carry that insight back into the harder problem. The method works best when the approximation comes with a clear bound or a test of its quality. Otherwise, an elegant simplification can conceal the very constraint that matters most. In everyday decisions, asking what would change if one obstacle disappeared can clarify priorities, but the answer remains a starting point, not proof that the obstacle is irrelevant.

Use chance strategically when exhaustive reasoning cannot finish.

Sampling can beat enumeration. When mathematician Stanislaw Ulam wondered how often solitaire deals could be won, calculating every possible game was hopeless. Playing many sampled games offered an estimate instead. This approach became the Monte Carlo Method, which uses repeated random samples to approximate answers that are too complex to derive directly. Randomized primality tests likewise check selected values: repeating the test makes the chance of a false result extremely small.

Randomness can escape traps. A local search that repeatedly improves a solution may get stuck on a nearby peak. Simulated annealing begins with greater willingness to accept a worse move, then gradually becomes more selective. That temporary retreat can open a route to a better overall solution. Randomness here is not carelessness; it is a controlled way to explore possibilities that a rigid strategy would never reach.

Analysis

Randomness is valuable when it is designed to reduce a specific bottleneck: impossible enumeration, false certainty, or a search trapped in a local optimum. The distinction between sampling and guessing matters. A random method needs enough well-chosen trials, and its error rate should fit the consequences of being wrong. In low-stakes discovery, broad exploration may be liberating. In medicine or security, even a tiny error probability demands careful validation. Chance is a tool whose usefulness depends on both its calibration and the cost of failure.

Protect responsiveness by batching interruptions and refusing endless queues.

Communication has overhead. Internet messages travel as packets and use acknowledgments to confirm delivery. When transmission repeatedly fails, Exponential Backoff makes each retry wait longer, reducing collisions while preserving the possibility of another attempt. The principle has a human analogue: after repeated cancellations, space invitations farther apart rather than either pursuing someone constantly or ending the relationship abruptly.

A queue can become the problem. Oversized network buffers accept traffic but delay it so long that interactive tasks become sluggish. The authors call this bufferbloat. People create a similar burden when every email, request, and missed show is stored for later attention. Batching routine checks reduces task-switching costs; deliberately dropping low-value requests can be better than keeping everyone waiting. For conversation, brief listener feedback also helps the speaker adjust and continue coherently.

Analysis

The chapter distinguishes capacity from responsiveness. A system can hold more information and still serve people worse if it makes them wait too long or keeps them mentally on call. That applies to institutions as much as devices: a queue that never rejects demand can hide a mismatch between workload and capacity. But dropping messages is not universally kind. Some requests are urgent, and some people have less power to absorb delay. Good communication design makes expectations explicit and gives high-stakes traffic a different path.

Design rules that make cooperation easy and strategic guessing unnecessary.

Bad outcomes can be stable. In the prisoner's dilemma, two partners each do better by betraying the other, whatever the other chooses. Both then receive a worse result than if they had cooperated. The tragedy of the commons scales this problem up: individually reasonable overuse can ruin a shared resource. The answer may be to redesign incentives, not lecture people to behave better. A Vickrey auction, for example, makes honest bids the best strategy by charging the winner the second-highest bid.

Make decisions easier to share. The book's idea of computational kindness is reducing the mental work imposed on others. State a genuine preference instead of asking everyone to infer it, or offer two concrete choices rather than a blank field of possibilities. In a group trip, unspoken assumptions can make everyone endorse a plan nobody actually wants.

Analysis

The connecting insight is that social outcomes depend on the structure of interaction, not simply on individual character. Rules can make cooperation stable, honesty advantageous, or ordinary coordination less exhausting. That perspective shifts attention from blame to design while preserving personal responsibility. It also warns against treating every preference as a game to be optimized: trust, dignity, and unequal bargaining power matter. The most humane structure gives people useful information, makes their choices legible, and avoids forcing them to solve a puzzle whose rules were never explained.

Analysis

Algorithms to Live By is a story-driven tour of computer science as a practical theory of bounded decision-making. Its unifying move is to treat ordinary dilemmas as formal problems with scarce time, space, attention, and information. The benefit is not that life becomes mathematically tidy. It is that familiar anxiety becomes diagnosable: a search has a stopping point, a schedule has an objective, a prediction has assumptions, and a social dilemma may have bad rules rather than bad participants.

The book is strongest when it shows that rationality includes computational cost. The optimal method is not always the one that finds the best answer after unlimited deliberation; it is the method that makes the best tradeoff between accuracy, effort, and delay. This connects naturally to bounded rationality and makes supposedly irrational practices, such as forgetting, sampling, approximation, and simple heuristics, intelligible as adaptations to limited resources.

The limit is equally important. An algorithm optimizes a specified objective; it cannot decide whether the objective is humane, fair, or worth pursuing. The 37% rule assumes a particular search structure. A scheduling rule depends on the chosen metric. A social equilibrium may be stable and still miserable. The book acknowledges this distinction most clearly in its closing account of computational kindness: people design the problems other people must solve.

That makes the book less a recipe collection than a literacy guide. Its models are most useful when treated as questions: What is scarce? What is being measured? What assumptions make this rule optimal? What is the cost of waiting, switching, or being wrong? Such questions preserve the clarity of formal reasoning without pretending that human values can be reduced to a single score.

Summary in 30 Seconds
A search with disappearing options can benefit from spending about 37% of the available time learning before choosing. Scheduling rules only work when the goal is clear, since minimizing lateness differs from maximizing value. Predictions should start with prior expectations and account for small samples. Complex models can mistake noise for signal, so test them on unused data. When exact solutions cost too much, sampling and approximation can produce useful answers.
Contains spoilers

Last updated:

Report Issue

Review Summary

4.12 out of 5
Average of 35k+ ratings from Goodreads and Amazon.

Algorithms to Live By receives mostly positive reviews for applying computer science concepts to everyday decision-making. Readers appreciate its insights into optimal strategies for common problems, though some find the mathematical explanations challenging. The book is praised for its engaging writing style and practical applications, but criticized by a few for oversimplification. Many reviewers highlight the book's unique approach to self-help through algorithmic thinking. Overall, it's recommended for both tech-savvy readers and those interested in improving decision-making processes.

Your rating:
4.54
712 ratings
Want to read the full book?

Glossary

Look-Then-Leap Rule

Observe, then commit

An optimal-stopping strategy for sequential choices when options cannot be recalled and only relative rankings are known. First, pass over a set portion of the options to establish a standard. Then choose the first later option that beats everything seen so far. In the classic secretary problem, the initial observation period is about 37 percent of the pool.

Explore/exploit tradeoff

Learn versus use

The tension between trying options to gain information and choosing an option already known to provide a good result. Exploration can reveal a better choice, while exploitation turns existing knowledge into immediate rewards. The right balance depends strongly on how many future opportunities remain to benefit from what is learned.

Upper Confidence Bound

Favor plausible upside

A strategy for repeated choices under uncertainty. It assigns each option a score based on the highest outcome that remains reasonably plausible given the evidence so far, then selects the option with the highest score. Little-known options receive an exploration bonus, which shrinks as experience makes their likely performance clearer.

Least Recently Used (LRU)

Evict the least recent

A cache-replacement policy that removes the item that has gone longest without being accessed when limited storage fills up. It relies on temporal locality: information used recently is often likely to be needed again soon. The same principle can guide the placement of files or belongings used repeatedly.

Bayes's Rule

Update belief with evidence

A method for revising the probability of a hypothesis after observing evidence. It combines the prior probability of the hypothesis with the likelihood of seeing the evidence if that hypothesis were true. The rule formalizes how existing knowledge and new observations jointly shape a prediction.

Copernican Principle

Assume no special timing

A prediction method for a phenomenon whose total duration is unknown and for which no informative prior knowledge is available. It assumes the observation occurred at an ordinary point in the phenomenon's lifetime. Under the book's formulation, the expected remaining duration is then about as long as the duration already observed.

Overfitting

Mistaking noise for pattern

A failure of prediction in which a model conforms too closely to the particular data used to build it. The model may explain those observations extremely well while performing poorly on new ones, because it has learned random variation rather than a durable relationship. Overly complex models and unreliable measurements increase the risk.

Regularization

Penalize excess complexity

A family of methods that discourages a model from becoming unnecessarily complex. By adding a penalty for using many factors or assigning them excessive influence, regularization can reduce sensitivity to noise and improve predictions on unfamiliar data. It formalizes the tradeoff between closely fitting known observations and generalizing beyond them.

Constraint Relaxation

Temporarily loosen rules

An optimization technique that removes some constraints temporarily to create an easier version of a difficult problem. The solution to the relaxed problem can provide a bound on the real problem or serve as a starting point for a practical solution. Its value depends on how clearly the relaxed result relates back to the original conditions.

Lagrangian Relaxation

Turn rules into penalties

An optimization method that moves selected constraints into the scoring function as penalties. Instead of making a violation impossible, the model assigns it a cost. This can make an intractable problem easier to solve while preserving a way to judge the consequences of violating the original rules.

Monte Carlo Method

Estimate by random samples

A computational method that estimates an answer by repeatedly drawing random samples or simulating possible outcomes. It is useful when exhaustive calculation is infeasible, as in complex probability problems. More samples generally improve the estimate, though the result remains approximate and its uncertainty should be considered.

Exponential Backoff

Lengthen waits after failure

A networking strategy that increases the possible delay before each retry after repeated transmission failures, commonly by doubling the delay range. Random variation helps competing senders avoid colliding again at the same moment. The method reduces congestion without requiring the sender to abandon every future attempt.

Bufferbloat

Excessive delay from queues

A network-performance problem caused by buffers that are so large they accept excessive traffic rather than signaling congestion. The resulting queues create long delays, especially for interactive communication, even when the network appears busy and capable. More storage can therefore reduce responsiveness rather than improve service.

Nash equilibrium

No player benefits by switching

A set of strategies in which each participant's choice is a best response to the choices of the others. No participant can improve their own outcome by changing strategy alone. An equilibrium is stable in this limited sense, but it need not be beneficial or fair for the group.

Computational kindness

Reduce others' decision burden

A design principle for interaction: structure requests, choices, and environments so other people do not have to perform unnecessary mental work. Stating a preference, narrowing a menu of options, or offering concrete meeting times can make coordination easier than leaving others to infer hidden preferences.

FAQ

What's Algorithms to Live By about?

  • Exploration of decision-making: The book explores how algorithms from computer science can be applied to everyday human decisions, such as dating, apartment hunting, and scheduling.
  • Interdisciplinary insights: Authors Brian Christian and Tom Griffiths blend computer science, psychology, and philosophy to provide a comprehensive view of decision-making.
  • Practical applications: It offers strategies for optimizing choices in various scenarios, making complex concepts accessible to a broad audience.

Why should I read Algorithms to Live By?

  • Enhance decision-making skills: The book provides tools and frameworks to improve decision-making in both personal and professional contexts.
  • Engaging storytelling: The authors use relatable anecdotes and examples to make complex concepts engaging and easy to understand.
  • Broader understanding of algorithms: It encourages critical thinking about how algorithms influence various aspects of life, from technology to social interactions.

What are the key takeaways of Algorithms to Live By?

  • Optimal stopping: The "37% rule" suggests sampling 37% of options before making a decision to maximize the chances of selecting the best one.
  • Explore/exploit tradeoff: The book discusses balancing exploration of new options with exploitation of known ones for effective decision-making.
  • Bayesian reasoning: Emphasizes updating beliefs based on new evidence to refine decision-making processes.

What is the "37% rule" in Algorithms to Live By?

  • Optimal stopping strategy: The rule suggests sampling 37% of options before making a decision to maximize the likelihood of selecting the best one.
  • Mathematical foundation: Derived from probability theory, it provides a systematic approach to decision-making in uncertain situations.
  • Real-life applications: Applicable to scenarios like choosing a partner, hiring employees, or selecting a job offer.

How does Algorithms to Live By explain the explore/exploit tradeoff?

  • Balancing act: The tradeoff involves choosing between exploring new options and exploiting known ones, crucial for optimizing decisions.
  • Epsilon-Greedy strategy: Suggests exploring new options a small percentage of the time while primarily exploiting the best-known choice.
  • Practical implications: Helps in making better choices in contexts like investing, job searching, or social interactions.

How does Algorithms to Live By explain Bayesian reasoning?

  • Updating beliefs: Bayes's Rule is used to update the probability of a hypothesis based on new evidence, refining decision-making processes.
  • Practical applications: Applied in situations like medical diagnosis or predicting outcomes, helping make informed decisions.
  • Cognitive implications: Encourages critical thinking and a nuanced approach to decision-making by considering both prior beliefs and new evidence.

What is the Secretary Problem in Algorithms to Live By?

  • Optimal stopping scenario: Involves choosing the best candidate from a pool of applicants interviewed sequentially.
  • 37% rule application: Reject the first 37% of candidates to establish a baseline, then select the next candidate better than all previously seen.
  • Broader implications: Illustrates decision-making challenges under uncertainty and the mathematical principles guiding optimal choices.

What is overfitting and how does it relate to decision-making in Algorithms to Live By?

  • Definition of overfitting: Occurs when a model is too complex, capturing noise rather than the underlying pattern in data.
  • Impact on decisions: Can lead to poor decision-making by relying on overly complex models that don't generalize well.
  • Strategies to avoid overfitting: Use simpler models and regularization techniques to make more robust choices in uncertain environments.

How does Algorithms to Live By address the concept of computational kindness?

  • Definition of computational kindness: Designing systems and algorithms that reduce cognitive load and ease decision-making.
  • Real-world examples: Applied in situations like restaurant seating policies or urban planning to enhance user experience.
  • Broader implications: Encourages creating environments that support better decision-making with less mental effort.

What are some examples of algorithms discussed in Algorithms to Live By?

  • Mergesort: Highlighted for its efficiency in organizing data, demonstrating the importance of sorting in computing and everyday life.
  • Caching algorithms: Discussed for personal organization, emphasizing prioritizing frequently accessed items.
  • Scheduling algorithms: Explored for optimizing time management and task completion in personal and professional contexts.

What does Algorithms to Live By say about scheduling?

  • Single-machine scheduling: Explores managing tasks and deadlines, prioritizing based on due dates and task lengths.
  • Optimal strategies: Presents algorithms like Earliest Due Date and Shortest Processing Time to enhance productivity.
  • Real-world applications: Principles can be applied to everyday life, helping manage time and tasks effectively.

What are the best quotes from Algorithms to Live By and what do they mean?

  • "The heart has its reasons that reason knows not.": Highlights the tension between emotional and rational decision-making.
  • "Life is just too complicated for that.": Reflects the complexity of real-world decision-making, encouraging practical approaches.
  • "Computational kindness isn’t just a principle of behavior; it’s also a design principle.": Advocates for user-friendly systems that facilitate better decision-making.

About the Author

Brian Christian is an accomplished author known for his work on technology and human behavior. His books, including "The Most Human Human" and "Algorithms to Live By," have received critical acclaim and bestseller status. Christian's writing has been featured in prestigious publications and translated into multiple languages. He has appeared on various media platforms and lectured at prominent institutions. With degrees in philosophy, computer science, and poetry, Christian brings a multidisciplinary approach to his work. He has received fellowships and awards for his contributions to science writing and poetry. Currently based in San Francisco, Christian is a Visiting Scholar at UC Berkeley and the Director of Technology at McSweeney's Publishing.

Download PDF

To save this Algorithms to Live By summary for later, download the free PDF. You can print it out, or read offline at your convenience.
Download PDF
File size: 0.28 MB     Pages: 23

Download EPUB

To read this Algorithms to Live By summary on your e-reader device or app, download the free EPUB. The .epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.
Download EPUB
File size: 1.38 MB     Pages: 28
Want to read the full book?
Follow
Listen14 mins
Home
Swipe
Library
Get App
Try Full Access for 3 Days
Listen, bookmark, and more
Compare Features Free Pro
📖 Read Summaries
Read unlimited summaries. Free users get 3 per month
🎧 Listen to Summaries
Listen to unlimited summaries in 40 languages
—
❤️ Unlimited Bookmarks
Free users are limited to 4
—
📜 Unlimited History
Free users are limited to 4
—
📥 Unlimited Downloads
Free users are limited to 1
—
Risk-Free Timeline
Today: Get Instant Access
Listen to full summaries of 26,000+ books. That's 12,000+ hours of audio!
Day 2: Trial Reminder
We'll send you a notification that your trial is ending soon.
Day 3: Your subscription begins
You'll be charged on Nov 24,
cancel anytime before.
Consume 2.8× More Books
2.8× more books Listening Reading
Our users love us
1M+ readers
Trustpilot Rating
TrustPilot
4.6 Excellent
This site is a total game-changer. I've been flying through book summaries like never before. Highly, highly recommend.
— Dave G
Worth my money and time, and really well made. I've never seen this quality of summaries on other websites. Very helpful!
— Em
Highly recommended!! Fantastic service. Perfect for those that want a little more than a teaser but not all the intricate details of a full audio book.
— Greg M
Save 62%
Yearly
$119.88 $44.99/year/yr
$3.75/mo
Monthly
$9.99/mo
Start a 3-Day Free Trial
3 days free, then $44.99/year. Cancel anytime.
Unlock a world of fiction & nonfiction books
26,000+ books for the price of 2 books
Read any book in 10 minutes
Discover new books like Tinder
Request any book if it's not summarized
Read more books than anyone you know
#1 app for book lovers
Lifelike & immersive summaries
30-day money-back guarantee
Download summaries in EPUBs or PDFs
Cancel anytime in a few clicks

Why one summary?

Scanner
Find a barcode to scan

⭐
⭐
⭐
We have a special gift for you
Open
38% off
DISCOUNT FOR YOU
⭐
⭐
⭐
$79.99
$49.99/year
Only $4.16 per month
Continue
2 taps to start, super easy to cancel
Settings
General
Widget
Loading...