Book Review: Algorithms To Live By


    If you spend some serious time staring at my bookcase(s), you wouldn’t find a lot of self-help books. You’d find copies of all the books I’ve written reviews about for the Cycling ’74 newsletter. You’d find a bunch of fiction (lots of Haruki Murakami – who should have won the Nobel Prize again this year, by the way), books that describe odd little bits of the world, and loads of stuff I just plain loved enough to want to own in hardback rather than as an eBook.
    But books whose titles sound like slogans/soundbites or involve talking to myself in the mirror or only wearing certain colors as a means to domination? Nope. I tend to avoid those reductive formulas as far as my reading lists go.
    So you might imagine that I would never have looked a second time at a book with a title like Brian Christian and Tom Griffith's Algorithms To Live By: The Computer Science of Human Decisions, much less read it and then sat down to write about how great it is and how you should read (or listen to) it as soon as possible.
    If I'd gone with my own "avoid self help book algorithm," I'd have missed one of my most interesting reads for the year. Good thing I didn't, and here we are. Since this is a review about processes, there will be two: A brief description, followed by bullet points....
    This is a book that takes something you either understand or might want to understand and proceeds in an unexpected way - rather than describe the myriad of precise methods of calculation used by mathematicians and computers, the book takes on your problems: What do you do or leave undone? When's the best time to act, to stay, or to leave? Does neatness matter? The bewildering complexity of the world is where this book finds its home, and each section of the book itself (Bayes's Rule, Caching, Explore/Exploit, Game Theory, Networking, Optimal Stopping, Overfitting, Randomness, Relaxation, Scheduling, Sorting) is really more about how the machines and the people who taught them came to the formulations we call algorithms a lot later than we did, and that thinking about those ways of thinking will help us to connect, make better guesses, and make our way through the world. The algorithms don't begin on the page or in your patch - they begin in your life. Cue bullet point ranting in 3... 2... 1...
    • First, the book is more fun than formal, and a terrific pleasure to read - it tells you stories that connect you to the world, and then pulls aside the curtain to reveal that you've been thinking about (or thinking with) those algorithm things the whole time. And the stories often come at you in the best kind of sideways - piles of laundry by the bedside, Berkeley library book carts, oppressive to-do list cleanup, apartment hunting, Charles Darwin contemplating marriage, The 37 Percent Solution, Major League Baseball scheduling, and... well, I won't spoil it by giving away more of the game. Page for page, this has the highest BQ (Boggle Quotient) of anything I've read in years.
    • If you're a Max person, you're already hanging out at the edges of the Algorithm Theme Park, even if you don't know it. One could argue that this whole Forum is one big sprawling discussion about how to do X or the best way to do Y or how Max actually does Z behind the curtains. This book is the quickest and clearest explanation of and discussion of the lots of the basic ideas behind those questions I've ever run across, full stop.
    • While lots of Max users take to equations and pseudocode like baby turtles heading to the sea, not everyone does - and lots of those people would very much like to have another way to understand big or helpful ideas that doesn't opt for algebra right out of the gate. Instead, this book concentrates on outcomes and how they're relevant to everyday decisions - it winds up making a very good case for getting a few equations under your belt by showing you why it might matter to you or help you before bringing out the subscripts and superscripts. In my case, it banished my general distaste for randomness by showing me where its judicious use did amazing and elegant things - exactly what I was futzing around with in the Max patch that currently bedeviled me.
    • The whole darned book is worth the price of admission alone for the chapter on sorting algorithms.
    • If, like me, you are one of those people who avoid self-help books, this is one sneaky bit of work - even as you're entertained and amused and enlightened, you wind up getting a truckload of really useful and practical advice about how to expend your effort and use time and space more efficiently.
    • It's worth a listen, too. You're not the first bunch of people I have foisted this book on - and one of my earlier victims has some potentially great advice: In his view, the audio version of the book is a fantastic listening experience of bite-sized sections that (to quote him) "let you finish your tram ride somewhere by being smarter than when you got on." I may have to try it myself when I take another run at the book.
    The final and ultimately nerdiest bit of praise that I can heap on this little book is that I read every single thing in the Notes section at the end, and laughed through a bunch of it, even.
    Give this one a shot, and you owe me a pint if you love it!