I don’t dislike Google. In fact, Google has tried hard to be transparent about surveillance. It was the first big internet company to start publicly reporting the number of law enforcement requests it received. It has been active in the coalition pushing for reform of the electronic communications privacy law. And Google is appealing the government’s gag order that prevents it from revealing how many requests it receives from the Foreign Intelligence Surveillance Court.
But Google has also repeatedly abused users’ trust. In 2010, Google launched a social networking tool called Buzz that automatically listed people as “followers” of people with whom they frequently e-mailed or chatted on Gmail. Users who clicked on a button “Sweet! Check out Buzz,” were not adequately informed that the identity of their closest Gmail contacts would be made public. Google later agreed to settle charges by the Federal Trade Commission that Buzz was deceptive, and paid $8.5 million to settle a class action lawsuit about Buzz. In 2012, my colleagues and I broke the news that Google was bypassing the privacy settings of the Safari browser used by millions of iPhone and other Apple users by using a special computer code to trick their browsers into allowing Google tracking. Later that year, Google agreed to pay $22.5 million to settle the FTC charges that its Apple circumvention had violated the terms of its Google Buzz settlement. The $22.5 million settlement was, at the time, the FTC’s largest civil penalty of the kind. And in 2013, Google agreed to pay $7 million to settle with attorneys generals from 38 states and the District of Columbia who claimed that Google violated people’s privacy when its Street View cars inadvertently collected personal information from Wi-Fi networks.
I also have too much data stored with Google. My audit revealed that Google had stored all of my searches dating back to 2006 and had identified all 2,192 people that I had e-mailed in that time. [Editor’s note: To find out what Google and other companies knew about her, Angwin conducted a “data audit.” To find her Google data, she primarily explored the information available in her Gmail account settings.] Given the outdated privacy laws, I couldn’t expect the company to keep all that data secret. I needed to go on a Google data diet.
I started by quitting Google search.
My searches are among the most sensitive information about me. If I’m looking into buying a burner phone [a prepaid cell phone bought to try to stymie surveillance], all my searches are about burner phones. If I’m researching an article about facial recognition technology, all my searches are about facial recognition technology. Basically my searches are a fairly accurate prediction of my future actions.
As soon as I switched, I realized how dependent on Google I had become. Without Google’s suggested searches, and Google’s perfect memory of what I usually search for, each search required more work from me. For instance, DuckDuckGo doesn’t know that I live in New York City, so when I mistyped “Naturaly History Museum,” it brought up the Natural History Museum of Los Angeles. For comparison, I checked Google: Sure enough, it corrected my spelling and guessed I was in New York, listing the American Museum of Natural History in Manhattan at the top of my results.
DuckDuckGo’s lack of knowledge about me forced me to be smarter in my searches. For instance, I noticed I had become so lazy that I had been typing URLs—such as CNN.com—into the Google search bar instead of the navigation bar, even though I knew exactly where I was going. So I began typing in the addresses into the correct spot on my web browser.
The next thing I noticed: I had been Googling web pages that I visit very regularly—such as my kids’ schools and my yoga studio schedule—instead of just bookmarking them. And so I began bookmarking them.
In fact, I had gotten so accustomed to letting Google do my work that I found it a bit jarring to have to finish typing an entire word without Google finishing it for me. Without Google’s suggestions, however, I found that I was less distracted to search for things I didn’t need. No more typing in the letter a and having Google suggest “amazon,” and then suddenly remembering I needed to order something from Amazon.com.
With DuckDuckGo, I usually found what I wanted, although sometimes it was strange to be confronted with just three results. I was so conditioned to seeing “millions” of results for everything on Google.
But DuckDuckGo had some black holes. I desperately missed Google Maps and couldn’t find any other online maps that I liked as much. And I missed the Google News section.
Before going to a friend’s dinner party, I searched to remind myself of the promotion he had just landed at Columbia University. There had been some recent news about it, but all my searches on his name alone, Sree Sreenivasan, and his name and Columbia, turned up nothing. Finally, I tried “Sree, Columbia and News” and an article popped up. The news was there. I just had to retrain myself to use DuckDuckGo’s structure for news searches.
It dawned on me that I had tuned myself to Google. I had always thought of Google as a clean sheet of paper—possibly because of its nice white interface—but in fact I had molded my questions to adjust to how Google likes to answer questions.
Now I was tuning myself to a different service, DuckDuckGo, which had different ways of answering questions. It was like a new relationship; I was discovering my new partner’s quirks and foibles. And it was empowering; I was tuning myself to a partner that didn’t have a hidden agenda of tracking me.
I had broken free from Google, and the world was still on its axis. I had mastered another service and could still find the information I needed. The whole experience reminded me of a quote from Marc Andreessen, the co-founder of pioneering web browser Netscape back in 1994. “The spread of computers and the internet will put jobs in two categories,” Andreessen said in a 2012 interview. “People who tell computers what to do, and people who are told by computers what to do.”
Mastering my switch to DuckDuckGo made me feel I had a better chance of being in the category of people who tell computers what to do.
After using DuckDuckGo for a few months, I started to feel a bit uneasy. Who were these guys that I was trusting? And why was their logo a duck with a bow tie? It seemed kind of weird.
For all my dislike of Google’s tracking practices, I had developed an emotional snapshot of Google as a place with all the cheerful arrogance of an Ivy League university. It had principles but few scruples: It took a high-profile brave stand against censorship in China but was making money off my personal data every day.
I was having a hard time getting the same kind of mental image of the principles and scruples behind a cheerful duck in a bow tie.
And so I boarded a train to Philadelphia to meet the people behind the duck. From Philadelphia, I rolled for another twenty minutes through leafy suburbs and past the Bryn Mawr College campus before arriving at my destination, Paoli. It was easy to spot DuckDuckGo’s founder, Gabriel Weinberg, in the parking lot—he was the one whose car had duck stickers on it. Other than his shock of auburn hair, he looked like any other geek with his thick-framed glasses and a hoodie. I jumped in the car and we drove two minutes to his office.
To my surprise, we pulled into the parking lot behind a stone castle with colorful round turrets. “You work in a castle?” I said.
Yep, he did. DuckDuckGo’s offices were on the second floor, the walls decorated with ducks. Weinberg had a polka-dot couch in his office and a play area near his desk for his kids. He told me that his focus for the company wasn’t initially about privacy. He just wanted to build a better search engine. After selling a social networking website called the Names Database for $10 million in 2006, he and his wife moved to Valley Forge, Pennsylvania, so she could be close to her work at the pharmaceutical giant GlaxoSmithKline.
A freshly minted millionaire, Weinberg experimented with a bunch of projects. He made his own TV studio, worked on a social network for golfers, and started a service that sought to use crowdsourcing to find better search results. As he played around with search, he started to get
increasingly annoyed by Google search results that were filled with the equivalent of spam.
So he decided to build a better search engine. “I wanted to go back to the Google old days when the focus was on quality links,” he told me. Privacy came up only after he launched the first version of the website to the technology community, and some users asked about the site’s privacy policies. “Honestly I hadn’t given it one thought at all until then,” Weinberg told me. “So I took a hard look at search privacy. I thought it was pretty creepy what a search engine could have on someone—it’s arguably the most sensitive data you could have on someone on the internet. I decided the coolest course of action would be to take it completely out of my hands—and not store the stuff. After doing that, I realized this is kind of a core thesis for the company.”
By 2011, he had embraced privacy completely. He bought a billboard in San Francisco that said, “Google Tracks You. We Don’t,” and accepted an investment from a venture capital firm, Union Square Ventures, that was betting on the emerging market for privacy tools.
Over take-out sandwiches, a few of his engineers joined us to discuss the challenges of building a search engine from scratch. We talked about the challenges of building better maps and my frustrations with their news search results. Keeping DuckDuckGo privacy-friendly was difficult. The engineers had to build many of their technical tools from scratch. For instance, they had to build their own blogging software because the free blogging software contained tracking technology.
“It’s like you guys are survivalists,” I told Weinberg. “You have to grow your own food and stock your own guns.”
As we talked, I was surprised at how earnestly they approached building a better search engine. Somehow, with the polka-dot couch, the ducks, the castle, and Weinberg’s auburn hair, I had allowed myself to be lulled into thinking that it was more of a hobby than an actual company. But they appeared to be dead serious.
It reminded me of when I was a reporter at the San Francisco Chronicle in the late 1990s. I was dismissive of this newfangled search engine Google. I remember thinking: How could its reliance on machine-based page ranking be better than the hand-curated results on my favorite search engine, AltaVista?
Now I was sitting on a polka-dot couch in suburban Philadelphia wondering how a few guys working in a castle could pose a threat to a search engine that pulls in nearly $30 billion a year. And yet, in the technology industry, some of the best ideas sound crazy at first.
Julia Angwin is an award-winning investigative journalist at ProPublica. Excerpted from Dragnet Nation: A Quest for Privacy, Security, and Freedom in a World of Relentless Surveillance by Julia Angwin, published February 25, 2014, by Times Books, an imprint of Henry Holt and Company, LLC. Copyright © 2014 by Julia Angwin. All rights reserved.