Wednesday, December 27, 2023

Book excerpt: First pages of the book

(This is an excerpt from my book, "Algorithms and Misinformation: Why Wisdom of the Crowds Failed the Internet and How to Fix It". The first sentence and first page of a book hook readers in. This book starts with an entertaining tale about algorithms and their importance at the beginning of

The old brick building for Amazon headquarters in 1997 was in a grimy part of downtown Seattle across from a needle exchange and Wigland.

There were only a couple dozen of us, the software engineers. We sat at desks made of unfinished four-by-fours bolted to what should have been a door. Exhausted from work, sometimes we slept on the floor of our tiny offices.

In my office, from the look of the carpet, somebody had spilled coffee many times. A soft blue glow from a screen showing computer code lit my face. I turned to find Jeff Bezos in my doorway on his hands and knees.

He touched his forehead down to the filthy floor. Down and up again, his face disappeared and reappeared as he bowed.

He was chanting: “I am not worthy. I am not worthy.”

What could cause the founder of Amazon, soon to be one of the world’s richest men, to bow down in gratitude to a 24-year-old computer programmer? An algorithm.

Algorithms are computer code, often created in the wee hours by some geek in a dingy room reeking of stale coffee. Algorithms can be enormously helpful. And they can be equally harmful. Either way they choose what billions of people see online every day. Algorithms are power.

What do algorithms do? Imagine you are looking for something good to read. You remember your friend recently read the book Good Omens and liked it. You go to Amazon and search for [good omens]. What happens next?

Your casually dashed-off query immediately puts thousands of computers to work just to serve you. Search and ranker algorithms run their computer code in milliseconds, then the computers talk to each other about what they found for you. Working together, the computers comb through what are billions of potential options, filtering and sorting among them, to surface what you might be seeking.

And look at that! The book Good Omens is the second thing Amazon shows you, just below the recent TV series. That TV series looks fun too. Perhaps you’ll watch that later. For now, you click on the book.

As you look at the Good Omens book, more algorithms spring into action looking for more ways to help you. Perhaps there are similar books you might enjoy? Recommender algorithms follow their instructions, sorting through what millions of other customers found to show you what “customers who liked Good Omens also liked.” Maybe there is something that might be enticing, that gets you to click “buy now”.

And that’s why Jeff Bezos was on my office floor, laughing and bowing.

The percentage of Amazon sales that come through recommender algorithms is much higher than what you’d expect. In fact, it’s astounding. For years, about a third of Amazon’s sales came directly through content suggested by Amazon’s recommender algorithms.

Most of the rest of Amazon’s revenue comes through Amazon’s search and ranking algorithms. In total, nearly all of Amazon’s revenue comes from content suggested, surfaced, found, and recommended by algorithms. At a brick and mortar bookstore, a clerk might help you find the book you are looking for. At “Earth’s Biggest Bookstore”, algorithms find or suggest nearly everything people buy.

How algorithms are optimized, and what they show to people, is worth billions every year. Even small changes can make enormous differences.

Jeff Bezos celebrated that day because what algorithms show to people matters. How algorithms are tuned, improved, and optimized matters. It can change a company’s fortunes.

One of Amazon’s software engineers just found an improvement that made the recommender algorithms much more effective. So there Jeff was, bobbing up and down. Laughing. Celebrating. All because of how important recommender algorithms were to Amazon and its customers.

No comments: