From Lurking: How a Person Became a User, published this month by Farrar, Straus and Giroux.

A user of Google products might be put off by the company’s cheerfulness and believe that its old byword, “Don’t Be Evil,” was always bunk. But its steady dominance over internet infrastructure leaves skeptics with few alternative products. With consumers in this bind, Google released boggling ventures. At Gmail’s launch, users complained about the ads—they’re creepy and it feels like a robot is reading my email! Google Street View appeared, at first, as an obvious invasion of privacy, not to mention an act of hubris with an undercurrent of colonialism. But a person can hardly rail against a ubiquitous technology forever. It wouldn’t be easy to give up searching and, above all, Google is easy. All you have to do is wonder about something.

Not so long ago, the company articulated its fundamental purpose as providing access to all possible online information. Google even assumed responsibility for the absence of information. “We’re trying to build a virtual mirror of the world at all times,” Marissa Mayer, then Google’s vice president of geographic and local services, said at a conference in 2010. The comment was made when internet companies were thought to be quicksilver entities rather than institutions building legacies. “Mirroring the world,” while impossible, was a coherent vision, fitting with the company’s story and execution thus far.

In the Aughts, Google seemed determined to create a digital copy of everything. It was photographing all the streets and scanning all the books in the world, or so they wished you to believe. But one of its projects was an outlier—and in retrospect it signaled where Google was heading. In 2007, when most of the phones in people’s pockets were still dumb ones, the company launched a service called GOOG-411. If a user dialed 1-800-GOOG-411, provided a city, and made a request (“Miami,” “Thai food delivery”), the service would connect them, like a more personalized Yellow Pages. But then, after three years, the service shut down. It turned out that the point of the project had been to collect audio samples of accents and pitches and voices for artificial-intelligence research. “We need to build a great speech-to-text model,” Mayer had said in 2007. “So we need a lot of people talking, saying things so that we can ultimately train off of that.” All the data it indexed and represented would be thrown in the bin when it wasn’t necessary.

The following decade saw the quieter deactivation of services and deletion of some of the same archives Google once boasted about acquiring. Jessamyn West, a librarian and writer in Vermont, told me that part of the issue with the company is that Google has nothing like a support line. Even Comcast lets you call in and ask a human a question. Comcast has customers. Google has users. If a Google user has a question about Google, well, Google wants them to google it. The company’s approach is to give a user tools to find things, which is, as West puts it, the “opposite of what I do. One of the things that is really important to libraries is the concept of institutional memory,” she explained. “It’s not just that you’ve got this building full of information in whatever form it is in, but that you’ve got human beings who understand the corpus of what’s in your buildings or what’s in your collections.” Libraries are designed to serve their communities. Someone’s ability to use the library is a “factor in whether you are doing a good job as a librarian,” West said. “That’s not true with Google. They’re not answerable to people.”

Google had once tried to ingratiate itself to the librarian community. Representatives went to conferences like that of the American Library Association (ALA) with great enthusiasm, eager to partner with various groups, and especially to find librarians who might help scan books. In 2007, Google started a blog called Librarian Center. They hired a “Library Partnership Manager,” who sent out the Google Librarian Newsletter. But by the end of that year, the newsletters were sent less frequently, and they finally came to a stop in 2009. Later, the Librarian Center page was taken offline.

So much had changed in those few years since Google’s ALA debut. In the Aughts, Google was deliberate about identifying itself with book culture. Now there was more knowledge spread digitally over different formats and platforms than ever. An individual might turn to the web before visiting a library to research a subject. With this shift, Google no longer had to associate with libraries—or librarians.

If Google had ever been sincere in its desire to mirror the world, the company’s carelessness and lack of archival standards hindered its execution. In 2014, I was part of a panel discussion organized by the Institute of Contemporary Arts (ICA) in London and hosted remotely over Google Hangouts, as a collaboration between the museum and the Google Art Project. Recently, I went back to the video to verify something another panelist had said. I found the video on the website for the ICA, with the familiar black rectangle. But when I pressed play, there was a notice on a blank screen that read, “This video is unavailable.” I contacted the museum first. The developer, who was new to the position, explained that it was a direct-to-YouTube recording and no separate file existed. I wrote to several contacts at the Google Art Project, but no one could help. It could be that someone switched jobs and had a new email account and now there was no one who could log in to update the settings so that the YouTube video might be made public once again. That’s what comes from rapid growth: Google prioritized scaling up over the maintenance and continuity of its archive and older products.

The hidden implication of “mirroring the world” was that Google could replicate information on its own terms, and with no further commitment to maintaining data; any information erased or lost could be interpreted as something the world itself was missing. The company coasted on user trust garnered from its robust appearance. Why use products other than Google Docs or Gmail, if a startup’s competing offerings are more likely to break down or get hacked? Why bother uploading videos to any service other than YouTube, where it will be stored on Google servers, which are reasonably secure?

You—a user—or a school, or an institution, or any other body smaller than Google, now have habits shaped by Google’s influence. The ICA is a museum, which has standards and practices of archiving, collecting, and preserving objects and information. If Google had never had a hand in the event, the video would probably be available today. One consequence of Google’s dominance is that public institutions have relaxed certain functions and services that they believe Google’s tools provide for free.

I wonder how often Google strips its archives for parts, as it did with GOOG-411, before burying the data. How about the search itself? Are your queries nothing more than raw material to assemble into something else? There are reports that Google will eventually do away with search—do away with googling. The company hopes that you will talk to it like a maid in the kitchen, rather than search it like an archive. It would like to predict what you want to know with the data it has collected from you and about you.

When he was the company’s CEO, Eric Schmidt called multiple search results a “bug.” Google “should be able to give you the right answer just once. We should know what you meant.” When a YouTube video ends and an algorithm selects another, that’s Google’s attempt at a “right answer.” Today, YouTube’s “autoplay” is notorious for pushing users toward men’s rights and conspiracy theory videos, as a consequence of the most common user choices on the site and how the platform’s predictive algorithms are written. The company also has the power to invent what it does not know. Errors on Google Maps have resulted in the renaming of neighborhoods. Fiskhorn in Detroit is now known as “Fishkorn.” Google posted this typo years ago; now some local businesses, published advertising, and other services have codified it.

Google harvests inquisitiveness: something so fundamental to being human. It has so firmly embedded itself in the experience of learning new things that “search”—once a word that signified quest, yearning—is now synonymous with “googling.” Google has monopolized the act of asking a question as it whittles down possible answers and influences to determine which is the “right” one.

More than search or connection, or even artificial intelligence, Google should be remembered for its ceaseless practice of secret deletion and careless disorderliness. In the end, the company’s branding of itself as a fun-loving, ski-bum-in-a-ball-pit workplace is a fitting image: Google is a burnout, a flake. It bails on people.