This week, OpenAI introduced ChatGPT Atlas, an AI-driven web browser, which got me thinking: Is it finally time to move on from Safari?
That announcement was top of mind as Max Zeff, Sean O’Kane, and I explored the current browser market—including some lesser-known options—on the newest episode of the Equity podcast. Still, none of us seem eager to make a major change just yet.
Sean pointed out that many companies have attempted to challenge the dominant browsers, but most have failed because browsers alone don’t generate enough revenue. That’s likely less of a concern for OpenAI, given its enormous and growing funding.
Meanwhile, Max has experimented with Atlas and other browsers that claim AI agents will handle tasks for you. He described the benefit as a “minor efficiency improvement” at best. Often, you just watch the AI agent “navigate a site”—but is that really what everyday users want? There are also notable security concerns.
Below is a shortened and edited excerpt from our discussion.
Anthony: I’m still using Safari, but when it comes to search engines—which are closely linked to browsers—I’ve been trying out alternatives to Google because I’m tired of all the generative AI content crowding my search results.
There’s also the bigger question: If AI browsers become mainstream, what does that mean for the open web? You can still visit websites, but it’s not far-fetched to think that websites will become less relevant as more of our browsing is managed by AI tools and chatbots.
Max: This is a major topic of conversation: What will the agentic web actually look like? It’s a fascinating issue. Many have tried to develop solutions for a future they believe is inevitable.
It reminds me of earlier tech trends, where the real question is, “What’s the actual user experience? What’s the real value for someone using these tools?”
Right now, it’s not especially compelling. I’ve tested ChatGPT Atlas and Comet, and honestly, the best you can say is that they offer a slight boost in efficiency. You might get things done a bit faster.
But most of the time, when I’ve used these tools, I end up watching the AI slowly click through a website, doing something I’d never actually do myself. For example, having it find a recipe and add all the ingredients to Instacart—I’ve never done that. Tech influencers always use that example in demos, but I doubt most people actually do it.
There’s a huge disconnect here: The tech industry keeps saying, “We’re building all these agentic web tools,” but it’s unclear why the average person would want to use them. I’m not convinced.
Sean: I haven’t tried any of those AI browsers, mostly because I’m pretty old-school about searching and browsing—I spend a lot of time looking for documents, which means digging through different parts of web pages I know well, and doing lots of Boolean searches on Google. Maybe I’ll give these new tools a shot if Google ever gets rid of Boolean search, which sometimes feels inevitable, but that hasn’t happened yet.
What stands out to me about these AI browsers is that other companies have tried to break into the browser market and always lose, because it’s nearly impossible to make a browser profitable. Some have tried charging upfront, which works for a while, but it’s not sustainable against giants like Safari, Chrome, or Firefox.
What’s different now is that some of these companies have almost unlimited resources, so they can keep going as long as they want without worrying about immediate profits. OpenAI, for example, doesn’t need to make money from this right away—they can just let it develop over time.
