Search Engine War Continues: David and Goliath, or Google's Fight Against AI Startups

Disclaimer: This is a free translation columns John Herman for New York Magazine, by Technocracy. Subscribe to our channel “Voice of Technocracy”so as not to miss new materials about the AI ​​products market.

In June 2000, Google declared victory in the search wars, announcing, “Google will now be Yahoo's default search engine!” The company's founders, Sergey Brin and Larry Page, added, “Keep Googling.” And they were right.

In nearly 25 years, Google has grown into a $2 trillion company by taking control of the Internet's most valuable resource: the window through which people search for information. Google has become a guardian of that resource.

Over the past two decades, some good alternatives have emerged, but no real competitors have emerged. Microsoft spent over $100 billion on Bing but failed to gain significant market share. Meanwhile, Google has become a household word.

In the early 2000s, Google was better than the directories and search engines that came before it, but in recent years the question of “Is Google really that good?” has become ambiguous. There are no comparable competitors on the market. Who can compare it to?

But now, in 2024, Google has reason to worry. A court has declared the company a monopoly. The search network has become less relevant thanks to social media. And, like Yahoo at its peak, Google is becoming a product built with advertisers in mind, not average users.

And then there’s artificial intelligence. The arrival of ChatGPT, a new kind of service that lets people ask questions, in 2022 sent Google, itself a powerful AI, into a panic. Google redesigned search to make it more like a chatbot, but early results were mixed. Meanwhile, AI companies are making their chatbots more like search engines.

For the first time in years, tech giants and startups are directly competing with Google for control of access to information, and they think they have at least a small chance of winning. And Google seems to think so, too, judging by its risky venture into error-prone generative search.

Recently, users have been able to see AI-generated answers based on online sources in their search results. Google tested the feature for a year and launched it in May. Despite some implementation issues, the company is continuing to develop the concept.

Google is keeping a close eye on competitors like OpenAI, and as Business Insider's Ana Altchek reports, the company appears to be borrowing designs and features from OpenAI and AI startup Perplexity.

Google’s concerns about chatbots as competitors to search are, to some extent, justified. For some types of queries, even early chatbots based on large language models (LLMs) offered a more enjoyable experience than Google. They explained, engaged in conversation, synthesized, and summarized information, though they might occasionally make things up. But for many other queries, they were useless.

In early 2023, ChatGPT could help you with your homework, but it couldn’t tell you which team had just won the Super Bowl — it generated text, not extracted information. It was a fundamentally different system that just had a familiar interface.

Google has always struggled with whether and how to integrate AI-generated answers into search. After all, Google is a company that makes hundreds of billions of dollars a year from traditional search advertising.

But something even more surprising is happening outside of Google: AI-powered chatbots and search tools are becoming more like Google. Perplexity, a fairly popular search tool, has gradually evolved from a chatbot that spits out lots of footnotes into a product that looks a lot like Google.

There's a list of links labeled “sources” here. There's a widget that looks a lot like Google's old “knowledge panels” that predate generative AI. It even has related questions.

But ultimately, Google and the relatively small AI startups have the same approach to solving the users' problem: summarized text and a set of sources where you can read the details. But Google's results from the example are a little unreliable because their answer refers to the 2020 Super Bowl championship, although the Chiefs last won the cup in 2024.

OpenAI is also keeping up with the trend and recently started testing a similar product – SearchGPT, which is a traditional search engine combined with its own chatbot:

This kind of competitive convergence happens all the time in tech — think of how every social media platform turned its app into TikTok, while TikTok absorbed features from traditional ones — but it's fun to see something as old and familiar as Google conceptually merge with a category as new and vaguely defined as LLM-powered chatbots.

For Google, it's a leap into the unknown; for the ultra-hyped AI companies, it's a step from exciting newness to potential profit. But for the young companies, it's a technologically easier step: They already spend a lot of money and time collecting and processing information on the Internet, so they might as well make a search product out of it.

The answer to users' search queries will be summarized text, combined with lists of sources and site excerpts – this is what we can expect from search products in the coming years. Whether these companies can find a solution to the contradiction between the technology that finds content and the technology that generates it is up to them. The companies clearly believe that they should try.

This would not be a new or particularly difficult problem for Google, which has fought off many competent clones before, except for two things. First, the likely result of a recent federal court ruling on the company's search monopoly is that it will have to end deals with companies like Apple, to which it pays billions of dollars in exchange for being the default search engine on iPhones and other devices.

This doesn't guarantee increased competition—Google will likely still be the default search engine for most users, at least in the near future—but it does provide an opportunity for companies building AI products to add search capabilities to them and try to take away Google's user share.

Meanwhile, companies like Apple are showing interest in using chatbots that can answer many basic questions on their own.

Another problem is that none of the screenshots above contain ads. It would be naive to assume that the next generation of search engines will be filled with ads right away. For now, they are using the opportunity to grow their user base, forcing Google to redesign its most popular product away from its core business model.

Here again, there are echoes of the past. In 1998, Danny Sullivan, now a Google employee who occasionally speaks to journalists, in what may have been the first published review of Google Search, described the new product's strengths and questioned its business model:

What about the results? I think many will be pleased, especially for the ever-popular one- and two-word queries. A search for “Bill Clinton” brought the White House to the top spot. A search for “disney” brought disney.com to the top spot, along with its subsites Disney World, Disney Channel, and Walt Disney Pictures. But interesting alternative sites like Werner's Unofficial Disney Park Links also made the list.

Will Google engage in commerce? [Сооснователь компании Ларри Пейдж] does not object to this, but says that there is no hurry.

Compare that to Sam Altman, who said earlier this year that while OpenAI is “a business, and they’ll find something to charge for,” he’s “happy” to be able to offer his core product for free. And while Google in 2024 isn’t quite the Yahoo of the pre-millennium era — it’s both bigger and more in line with its new competitors — it has become a target for others, and their market share is a big piece of the pie that many want a bite out of.

Subscribe to our telegram channel “Voice of Technocracy”Every morning we publish a news digest from the world of IT, and in the evenings we share interesting and useful articles.