IE6, AI, and the future of browsing the Web
How Google's retreat from web standards signals a new era of browser stagnation—and what comes next
Chrome is the next IE6.
The signs are there. A few years back, Google cut funding to longstanding community contractors working on standards with the W3C. Community members I spoke with at CSS Day 2023 in Amsterdam found the move confusing, given that Google had supported this vital work for so many years. And maybe it was important—to the Web. But what Google was saying with its actions was, “This work is not important for Google right now.”
And with good reason. For the first time ever, Google searches are dropping. And despite rhetoric that “users won’t pay to use the Web” they are paying to use AI services competing with the Web like Perplexity. This happened once before when mobile apps hit the scene, and Google rose to defend the Web (and its advertising income) by launching Chrome, AMP, and a host of other initiatives to keep the web competitive with walled garden app stores.
But this time is different. AI just broke the Web that Google built.
Chrome’s IE6 era
While Google’s Chrome Platform team has been shipping new standards for web developers, Chrome, the browser itself, hasn’t made any significant improvements to its user experience in years. (By comparison, Firefox has shipped more UI updates in the same 5 years, including ones like sidebar navigation for tabs.) Perhaps Google sees no need to “fix” what is already working for ~70% of the Internet.
A similar situation occurred with Microsoft’s Internet Explorer 6 (IE6). After finally achieving a 90% browser market share, Microsoft lost interest and diverted its resources to sell more Windows and Office licenses.
The next largest browser is Safari, with around 15% of the browser market share, thanks to its iron grip on iOS. If you see Google/lobbyists pushing for Apple to allow other browser engines onto its devices in courts, it is likely more about gobbling up that high-value slice of market share than giving users the power of choice. Chrome is, and always has been, a vehicle for ad revenue and data collection for Google’s money-making duo: Search and Ads. In protecting these interests, Google has gone so far as to backpedal on deprecating third-party cookies, which underpin data collection and ad targeting.
Although Chrome receives the bulk of browser use, and thus Google has little motivation to innovate and compete, the writing on the wall is that Chrome’s days are numbered. User behavior suggests people are finding LLMs more convenient for finding answers than Googling, where they must leap over hurdles of ads, dive several pages into the search results, and then pogo in and out of websites to find answers to some of their most banal questions. And Google knows this. On May 7th, Eddy Cue, Apple’s senior VP of services, shared that searches on Safari dipped for the first time last month, which he attributed to people using AI like Perplexity. Google I/O touted a vision of an Agentic and AI-driven future brought to you by Gemini.
Google is shifting its attention
We can watch externally and see how Alphabet is reorganizing Google to accommodate this upcoming modality. We can see them testing AI-driven search results in Google. They reorganized their devices and Chrome teams a few weeks back, cutting industry-favorite Chrome team members. What do these actions say?
Google is saying, “I’m not interested in continuing to pour the same resources into Chrome anymore. I’m not interested in stewarding the Web as we know it.”
Consider also that the FTC is currently deciding if Google and Chrome should be split up (which could see Chrome bought by Perplexity or even Yahoo). Google executives would have to be blind not to see where the ball is going, that ChatGPT and its ilk threaten its core business model of serving ads to people searching for answers. This may explain why, at Google I/O this year, new AI features were shipping at the OS level, not the browser level. You could interpret all of this as Google pulling back resources from a product that is about to be outmoded, sold off, or, at the very least, doesn’t require further innovation to remain the top browser.
The Web’s post-ads era
If you build websites, you might be offended—“How dare Google stop investing in the free and open web?”
My friend, Google was always investing in keeping the Web competitive with apps so its ad revenue wouldn’t decline, to the point where it has been creating a “shadow Internet” in the form of AMP. It has never been altruistic, though it has wisely employed altruists.* And arguably, there’s nothing inherently “evil” about supporting an ecosystem of content with ads. The downside, however, is that the APIs, protocols, and standards Google fostered grew the Web only in one direction: ad-based revenue. The Agentic Web is poised to disrupt the ads-supported ecosystem because LLMs can deliver facts directly, and AI Agents can pierce through ads and data-collecting tech to harvest content on behalf of users, no human eyes required.
The question Google—and you—should be asking is, “How do we adapt to a world where people have a magical answer box embedded in their machine and don’t need to visit websites much anymore?”
This series of posts seeks to address the new challenges facing the Web, its users, and its creators with the introduction of “magical answer boxes,” generated content, and agentic systems. I shall endeavor to share neither optimistic nor pessimistic views but only to drive at the truth of what is happening and what we can do to bring about a thriving ecosystem of creators and consumers. I seek to address the following:
Money Economic models, payments, and protocols for a post-ads economy
Identity How we distinguish between ourselves, each other, and our “extended selves,” both personally and programatically.
Reach Optimization for reaching audiences using agents and LLMs
Security Methods to defend servers and information from unwanted agents and bots
Evolution: The creation of new browsers and “browser alternatives”
The Web is always a work in progress, and now it is changing more rapidly than ever before. Let’s navigate this shift together.
Homework/further reading
From Penny Press to Protocols: The Structural Shift AI Forces on the Internet picks up where Stratechery left off, drawing parallels with the telegraph’s disruption of the Penny Press and proposing new protocols for content consumption.
Everything Is Chrome is a deep dive into Chrome’s history, positing that Google has been trying to control the Web through Chrome’s marketplace power.
Originally posted as a thread on BlueSky, enough people asked for a link to the article version that it seemed like the perfect launch post for this promised series. Thank you to everyone who pre-signed up for this first missive and for encouraging me to share my thoughts and findings on how AI is reshaping our relationship with the Web and each other.
Special thanks to the following folks for their proofreading and feedback: Jack Herrington, Paul Asjes, Vale, Bob Monsour.
About the author
I’ve had the privilege of contributing to projects at Microsoft Edge, Meta’s React team, and AWS Amplify, in addition to working with startups focused on AI, auth, and design systems. As an Invited Expert at the W3C, I helped shape the Web Animation Standard and collaborated with Mozilla on development tools and documentation initiatives. Through these experiences, I’ve learned that while I may have been doing this work for quite some time, the real magic happens in the community. I believe deeply that we—developers, creators, and users—are the true architects of the Web. While standards bodies and large companies influence its direction, the Web ultimately belongs to all of us. Every small contribution we make helps pave the way forward.
The opinions I share in these posts are my own and do not reflect those of my clients or employers, past or present. They reflect only publicly available information.
Great first post! I look forward to reading more.
Love your logo. FYI, it's difficult to read the logo on a black background, such as in my dark-mode email program.