Friday, September 8, 2017

Review: Algorithms to Live By: The Computer Science of Human Decisions

Algorithms to Live By: The Computer Science of Human Decisions Algorithms to Live By: The Computer Science of Human Decisions by Brian Christian
My rating: 5 of 5 stars

I got the audio version of this book a year ago. Every time I thought to dive in, I felt a mild quaking in my soul. Gah, this is gonna be so hard, I worried. As a mere mortal without any background in computer science, advanced mathematics, logic, or statistics and risk, I feared my reach exceeded my ability to grasp.

Well, I wasn’t 100% wrong. It was hard. However, I understood and I learned. Yes, I hit replay dozens of times, but I got it. (Of course, after I ran through the audio version twice, I ordered the book because I just had to have it in my library.)

I did not expect the multidisciplinary palette from which the authors created this work. While teaching me about optimization problems in computer science, I came better to understand mean-variance portfolio optimization, game theory, equilibrium strategies, and caching, just to name a very few. This book has great depth. Remarkably, it has even greater range.

When examining the algorithmic dances that computers do nanosecond by nanosecond, we are also examining how we make decisions every day. Should I stay on this jammed expressway? How long should I wait for a table at my favorite eatery? Is it better to do three small laundry loads per week or have one big laundry day? How should I best arrange all of these books on my shelves? If you are like me, you have experienced that frustrating little circle, spinning and spinning, as your computer tries to wrest a result from the digital universe or just from your hard drive. When you are waiting for a taxi or a train, you are experiencing a life-size version of that little spinning circle. When do you chalk it and look for Plan B?

This book describes how computers solve their problems and at the same time it shows us how the problems computers solve are just like the ones we deal with and solve, day in and day out. This isn’t too shocking, since humans set up the computer decision-making trees in the first place. Still, when I am synthesizing many possibilities, or struggling with family schedule optimization problems, I really can’t wait to apply terms like “simulated annealing” and “the price of anarchy”.

At the end of the day, when my family members are all doing the equivalent of sticking a thumb drive in my ear and starting their respective downloads, instead of objecting with: “Wait a minute, one at a time, I have to think!”, it will bring me joy to say, “Don’t trigger a Bufferbloat, guys, no one wants a Tail Drop.”

Most fun fact I learned:

“In contrast to the widely held view that less processing reduces accuracy, the study of heuristics shows that less information, computation, and time can in fact improve accuracy.” My translation? Don’t forget to ask grandma what she thinks, it’s likely to be spot on.

I really loved this book. It is one I will return to often.

View all my reviews

No comments:

Post a Comment