who are ai browsers for?

OpenAI's ChatGPT Atlas and the Future of Web Browsing

This week saw the unveiling of ChatGPT Atlas, an AI-driven web browser from OpenAI, prompting consideration of whether it’s time to move away from established browsers like Safari.

The topic was central to a discussion on the latest Equity podcast episode, featuring Max Zeff, Sean O’Kane, and myself. Despite the innovation, a significant shift in browser preference doesn’t appear imminent for any of us.

The Challenge of Browser Monetization

Sean pointed out that numerous companies have attempted to challenge dominant browsers, but have consistently failed due to the difficulty of generating revenue directly from the browser itself.

However, this obstacle is less likely to affect OpenAI, given its substantial and growing financial resources.

Early Impressions of AI-Powered Browsing

Max has personally tested ChatGPT Atlas alongside other AI-integrated browsers, reporting only a “slight efficiency gain” in the best-case scenario.

He observed that users often find themselves observing the AI agent navigating websites, raising the question of whether this is a genuinely desired functionality. Furthermore, substantial security concerns are present.

Below is an excerpt from our conversation, condensed for brevity and clarity.

Exploring the Impact on the Open Web

Anthony: I continue to use Safari, but I’ve begun experimenting with search engines beyond Google, driven by a desire to avoid the increasing prevalence of GenAI content in my search results.

A key question arises: If AI browsers gain widespread adoption, what implications will this have for the open web? While web pages will remain accessible, it’s conceivable that websites may become less central as AI interfaces and chatbots increasingly mediate our online experiences.

The Agentic Web: Promise and Reality

Max: The concept of the “agentic web” has been widely discussed, and it presents a compelling question.

Many have proposed solutions to realize this future vision. However, a critical aspect often overlooked is the actual user experience and the value proposition for consumers.

Currently, the appeal is limited. I’ve tested ChatGPT Atlas and Comet, and at best, they offer a marginal improvement in efficiency.

Frequently, the experience involves passively watching the AI agent navigate a website, performing tasks that most users wouldn’t typically undertake. The common example of automatically adding recipe ingredients to Instacart feels unrealistic for many.

There’s a significant disconnect between the tech industry’s focus on the “agentic web” and the practical needs of everyday users.

The Role of Search and Established Habits

Sean: I haven’t yet used these AI browsers, largely because my work heavily relies on document research, which necessitates direct interaction with specific sections of web pages and the use of Boolean searches on Google.

I might consider them if Google eliminates Boolean search functionality, but that hasn’t happened yet.

The interesting aspect of these AI browsers is that we’ve seen numerous companies attempt to compete in the browser market, consistently failing. It’s incredibly difficult to monetize a browser as a standalone product.

Some have tried charging a fee, achieving limited success, but it’s ultimately unsustainable when competing against Safari, Chrome, or Firefox.

OpenAI's Unique Position

What sets this situation apart is that OpenAI possesses virtually unlimited financial resources, allowing them to sustain development and deployment without immediate profitability concerns.

They can afford to let the browser evolve organically without the pressure of generating revenue in the short term.