Alphabet, the parent company of Google, announced its second-quarter financial results today. In both the press release and in CEO Sundar Pichai’s remarks, AI was mentioned every couple of lines. What wasn’t mentioned? The word “web.” Not even when it came to talking about search, the core product of Google that was built on — and for — the World Wide Web.
In some ways, this wasn’t surprising. As far back as June 1999, when Google held its first-ever press conference to announce a major funding round, then CEO Larry Page was talking about AI. According to Steven Levy’s 2011 book about Google, In The Plex, Page “talked about Google using artificial intelligence and having a million computers someday.”
Some months later, in 2000, Larry Page expanded on that goal, saying that artificial intelligence was “the ultimate version of Google.” He went on to say:
“So if we had the ultimate search engine, it would understand everything on the web. It would understand, you know, exactly what you wanted, and it’ll give you the right thing. And that’s obviously artificial intelligence, you know. It’ll be able to answer any question, basically, because almost everything is on the web, right. And so we’re nowhere near doing that now, however, we can get incrementally closer to that, and that’s basically what we work on.”
Google’s Journey From Web-First to AI-First
It all began so innocently. In the late-1990s, when the company was just getting started, Google’s search engine seemed to be a way for web users to escape the portals that dominated the internet at that time. In January 1999, Page told a German startup magazine:
“People make decisions based on information they find on the Web. So companies that are in-between people and their information are in a very powerful position.” (emphasis mine)
So clearly, Google was created to be a conduit between web users and web publishers. The information people were seeking was “on the Web,” as Page said back then. Google had no intention of standing in the way of people surfing to the source: websites.
In its June 1999 funding announcement, Google described itself as “a start-up dedicated to providing the best search experience on the web.” The announcement added that Google’s mission was “to organize the world’s information, making it universally accessible and useful.” (Indeed, that is still Google’s mission to this day).
While the mission statement has been remarkably consistent over the years, the meaning has undeniably changed after a quarter-century. Back in 1999, it was assumed that making information “universally accessible and useful” meant presenting a list of links on a clean white webpage. But now, 25 years later, it’s obvious that Google sees AI summaries as the primary way to make information “universally accessible and useful.”
AI Summaries: Useful or Not?
How “useful” are AI summaries in 2025? Consider current CEO Pichai’s comments in the Q2 announcements today.
“Search delivered double-digit revenue growth, and our new features, like AI Overviews and AI Mode, are performing well,” he stated in the press release.
But who are Google’s AI features “performing well” for, other than Alphabet’s shareholders? I’d argue it’s often a poor experience for web users, because of the increased chances of false information (a.k.a. AI hallucinations) and getting far less context than they would by clicking through to the original sources on the web. So, perhaps useful to users — but also flawed.
An example of a Google AI Overview, summarizing one of my recent articles on The New Stack.
“We see AI powering an expansion in how people are searching for and accessing information, unlocking completely new kinds of questions you can ask Google,” Pichai added in his remarks.
That wording (“…new kinds of questions”) glosses over the fact that Google’s goal with AI features is to provide answers without users having to click out to the open web. So, AI summaries are not useful to web publishers.
Wait a minute, aren’t attribution links present in Google’s AI summaries? Yes, but as PEW reported this week, those links are less likely to be clicked compared to search results with no AI. 404 Media followed up by noting that sometimes the attribution links aren’t even linked to the original source. 404 Media cited its recently reported story of AI content on Spotify, which was then aggregated by other sites — and those sites received the credit on a Google AI summary.
Transcending the Web?
The really worrying question for me, as a longtime, passionate fan of the web, is this: was Google’s ambition from the very start to transcend the open web? In other words, the open web becomes simply a source of information for its AI — kind of like the human batteries in The Matrix — but otherwise it’s kept in the background. Out of sight, out of mind.
Of course, Larry Page didn’t say anything like that in his 2000 quote about AI. But he did say that an AI will “be able to answer any question” because the AI will get its answers from the web.
The cognitive dissonance comes because Google has always promoted itself as an advocate for the open web. When the company launched its Chrome browser in September 2008, the announcement post stated:
“So why are we launching Google Chrome? Because we believe we can add value for users and, at the same time, help drive innovation on the web.”
And what is a browser for, fundamentally? That’s right: to browse web pages and web applications. To click around the open web! Later in the post, Chrome was described as “a tool to run the important stuff — the pages, sites and applications that make up the web.” The value for users of Chrome, according to this post, was the same as for Google search:
“Like the classic Google homepage, Google Chrome is clean and fast. It gets out of your way and gets you where you want to go.”
How did we go from Google getting “out of your way” and helping you browse information on the web, to Google wanting to keep all the web’s information to itself via its AI summaries?
Why go out onto the open web when you can find what you need in our AI summaries seems to be the new message. If that is indeed the intention for the AI-powered “ultimate version of Google,” then the fatal flaw is obvious: what happens when the original sources of the web start to fade away and ultimately disappear? I guess the AI will have to feed on itself at that point — which doesn’t seem very smart to me.
Originally published at The New Stack: https://thenewstack.io/is-ai-the-ultimate-version-of-google-as-larry-page-wanted/