What is Google Search in 2023? A site that started 25 years ago as a list of blue links has mutated beyond recognition. Today, Google isn’t just an index to help sort through the endless libraries of online information—it’s a reference guide for the physical world too, having mapped most corners of the Earth and cataloged its contents. It is part encyclopedia and part predictive engine, guessing what you might be typing or thinking, serving information based on what others before you typed. It is Moviefone and the stock ticker, a well-trained chatbot, an image repository, a shopping mall.
It is also, like the modern web, creaking under the weight of its enormous size and an infinitely renewable supply of content, both human- and AI-generated. Unlike its streamlined, efficient former self, Google Search is now bloated and overmonetized. It’s harder now to find answers that feel authoritative or uncompromised; a search for healthy toddler snacks is overloaded with sponsored product placement, prompts to engage with “more questions” (How do you fill a hungry toddler? “Meat and Seafood. Bring on the meat!”), and endless, keyword-engorged content. Using Google once felt like magic, and now it’s more like rifling through junk mail, dodging scams and generic mailers.
It’s fitting, then, that as the shine has worn off, Google has also found itself in an unusual moment of vulnerability. The company is on trial and scrambling internally to maintain relevance in the generative-AI era. Over the next several weeks, the Justice Department and a group of states will argue that the company violated antitrust laws by striking deals to protect its monopoly position in the online-search business. It all feels less like a midlife crisis than an existential one.
At the heart of the case against Google is a simple question: Does the company command 90 percent of the U.S. search-engine market because its technology is superior and users genuinely prefer it, or because it paid massive sums to companies like Apple to be used as a default service over competitors such as Bing? In its opening days, the trial has hinged on arguments about the importance of data. The DOJ has argued that having access to user data such as web history and past searches (the kind that are easier to collect when you’re the default browser on millions of phones) is like “oxygen for a search engine”; Google plans to argue that, although user data can improve search quality, “there are diminishing returns to scale.”
That defense is ridiculous. Scale means everything to Silicon Valley and the venture-capital firms that fuel its start-ups—including Google, in its early years. The laws of network effects state that users beget more users, that those users shed data that are helpful in personalizing and streamlining the user experience. Google thrives because it indexes unfathomable amounts of information across the web and processes several billion queries a day—queries that are aided in part by what the company describes as “aggregated and anonymized interaction data” from its users.
The company has long bragged about the scope of its work. “There isn’t anything that has operated at our scale before,” Bill Coughran, Google’s then–senior vice president of engineering, told Forbes in 2011. Google’s mission statement—to “organize the world’s information”—is a project that is defined by scale, one that requires the company to wrap its arms around an ever-evolving internet, process that information, and return the relevant bits to an infinite queue of hungry searchers. To suggest that there are “diminishing returns to scale” is to deny the very model that makes Google Search effective. Internal Google emails submitted as evidence by the DOJ attempt to show that the company’s engineers and executives know this. One email shows Udi Manber, a senior engineer, directly contradicting the argument. “It’s absolutely not true that scale is not important,” Manber wrote. “We make very good use of everything we get.”
And yet, Google’s lawyers might be right about the diminishing returns of its size—just not in the way they intended. In recent years, ex-Googlers have written lengthy blog posts saying that the company has lost its way and is coasting on its own success, which has made it “conservative,” “arrogant,” and lacking a mission and a sense of urgency. Noam Bardin, the longtime CEO of Waze, which was acquired by Google in 2013, has lamented having to navigate Google’s stifling, labyrinthian corporate bureaucracy.
Google’s stagnation has trickled down to its products. Search has grown bloated with advertisements and widgets that prioritize its own services, such as Shopping, Finance, and Maps, over organic results. Although the company constantly updates and tweaks its algorithms, which are powered by natural-language-processing AI tools, there is a creeping sense—among frustrated programmers, searchers, and even journalists—that the site is no longer as useful or intuitive as it once was. One reason for this feeling may be that Google’s algorithms have been successfully gamed by low-quality websites and search-engine-optimization companies that help their clickbaity clients show up in the first page of Google Search results. An SEO expert named Marie Haynes suggested to me last year that Google’s algorithms are getting better but that the website is going through a transitional period—the engine is still working out the kinks. In response to a request for comment, a spokesperson for Google said, in part, “We make thousands of quality improvements every year to ensure we’re delivering the most helpful results, and we’ve developed entirely new ways to search, which is why people around the world continue to use and love Search.”
Still, there are indicators that the product might continue to struggle to deliver high-quality search results on an internet soon to be flooded with AI-generated photos, videos, and text. Just this week, 404 Media reported that the top Google result on a search for the Tiananmen Square “tank man” was, for a period of time, an AI-generated image of the man taking a selfie. Last fall, the former Googler and ex–Yahoo CEO Marissa Mayer offered a charitable explanation as to why Google Search feels worse, arguing that the search engine is merely a reflection of the internet as a whole, which has become more complex and laden with fraudsters and garbage. The explanation lets Google off the hook a bit: The company’s chief purpose is to navigate the sprawling morass of information, after all. Ultimately, it’s a tacit admission that Google has tried and failed to tame the web.
Search’s devolution is a familiar story in an economy that demands untenable growth. The trajectory always looks like this: Invent a world-changing technology, scale it up, monetize it, print money, and take it public. Then, of course, there is the pressure to expand even more to appease investors. In Google’s case, organizing the world’s information meant conquering the web, books, images, inboxes, and geography. But with every success, there is more pressure to scale further, this time into moon-shot territory — self-driving cars and even a project to “cure death.” It becomes easier to acquire companies with the war chest than to build products from scratch, and to ink exclusivity deals in order to keep competitors at bay.
Google Search is arguably the most important innovation of the commercial internet—a modern marvel that has forever changed how billions of people access information. Its success, especially early on, felt like a small miracle. But Google Search is also a tragedy: The democratization of information demanded the invention of a behemoth that could successfully handle the abundance of the web. But chasing scale is a dangerous, unwinnable game. Perhaps the real lesson of Google’s story is that you can try to iterate, but ultimately, no company or product can grow alongside the internet forever without, eventually, being swallowed up.