Computing algorithms could help combat the messy compromises of real life, says Oliver Burkeman
I wasnt predisposed to love Algorithms To Live By, a new book by Brian Christian and Tom Griffiths that suggests approaching life decisions like a computer scientist. With the greatest respect to the computer scientists I know, its a job that evokes certain cliches not associated with healthy work-life balance, social skills or high tolerance for sunlight. Open the book at random, and you might find that stereotype confirmed. Did you know that, according to maths, you should marry the first person you meet once you turn 26 whos better than all previous people youve dated? (This assumes you started looking for a spouse at 18 and want to find one by 40.) Of course, nobody could ever bring themselves to live so mathematically, even computer scientists, and yet, by the end of the book, I was convinced. Not because I endorse the idea of living like some hyper-rational Vulcan, but because computing algorithms could be a surprisingly useful way to embrace the messy compromises of real, non-Vulcan life.
Computer science, Christian and Griffiths point out, is all about coping with limitation. We ask computers to do a million complex things, and at lightning speed. But they have limited processing power, so its always a matter of tradeoffs. When is it better to be fast than accurate, or vice versa? When should a computer stop searching for the perfect solution to some puzzle and use a rough-and-ready one instead? Slightly rephrased, these are the central challenges of life. When do you stop searching for a better partner, flat, group of friends, career path or local pub? Youd like to make the best possible choice, but gathering data comes at a price. Spend your whole life auditioning new spouses, friends or jobs, and you wont have spent it well.
The best algorithmic solutions vary according to the scenario. One appealing idea, when youre facing a fork in the road, is to choose the option with the highest upper confidence bound the one that could plausibly perform best in future (even if its also the one with a higher chance of being awful). Theres also the useful method known as constraint relaxation, a technical way of describing self-helpy questions such as, What would you do if you werent afraid, or if money were no object? Examine your predicament, then remove one of the constraints money, time, family disapproval and ask what youd do. The answer may clarify your real-world decision.
But the authors most immediately useful concept may be computational kindness. When making a plan with a friend, it feels polite to say youre flexible about when to meet, or that you dont mind at which restaurant. But refusing to state your wishes imposes a computational cost on the other person: now he or she must make the choice (while guessing at your preferences). For humans, as for computers, deciding makes demands on limited processing power. Dont overtax yours, and dont force your friends to use theirs on your behalf.
oliver.burkeman@theguardian.com
Read more: www.theguardian.com
