The Browser Company Discusses the Future of Arc, Dia, and the Web

In a letter to its users, The Browser Company is providing more details about its shift from Arc to Dia and what that means for the future.

Depending on where you’re at, this is a big deal or a non-event: Arc browser is an innovative shift away from the overly familiar design of other web browsers, but it’s also different enough that many find it confusing. Not helping matters, it was designed first on the Mac, and it remains a better app on that platform despite a big emphasis on the Windows version last year. Worse, the small company came to realize over time that its AI ambitions for the future of the web would be difficult or impossible to realize with Arc, and so it hit reset and announced an AI web browser called Dia that was set to launch in early 2025 but has yet to appear.

The result was a lot of confusion. And, for those who fell in love with Arc, worries that their favorite web browser would simply disappear. (There’s a Firefox-based Arc clone called Zen browser for those that love the Arc user experience and can put up with Firefox’s memory and performance issues.) Plus, it’s been months since we’ve heard anything substantive from The Browser Company.

“We started The Browser Company with a simple belief: The browser is the most important software in your life, and it wasn’t getting the attention it deserved,” The Browser Company co-founder and CEO Josh Miller writes in the letter. “We wanted to build something that felt like ‘your home on the internet,’ for work projects, personal life, all the hours you spent in your browser every single day … Something with taste, care, feeling.”

Miller and his small company correctly felt that web browsers hadn’t evolved to match their increased importance, and they could not have been more right about that. But rather than evolve this interface in small ways, as we see with some third-tier web browsers, they decided to radically rethink this app and create what they thought of internally as “the Internet computer.” The result, as noted, was polarizing: Some loved it, but many found it to be too different, with a steep learning curve. Miller calls this “the novelty tax.”

The Browser Company grappled with this problem and came to understand that it would need to start over instead of trying to integrate the AI functionality in Dia into Arc. Dia became an “opportunity to fix what [it] got wrong with Arc.”

There were three core problems, Miller says. The first is complexity, so the plan for Dia is to hide complexity behind familiar interfaces. Arc is also bloated with too many features, so Dia was built with a new architecture that prioritized performance from the start; The Browser Company has “sunsetted” its use of “TCA and SwiftUI to make Dia lightweight, snappy, and responsive.” (TCA is The Compositable Architecture, a SwiftUI-based app framework.) And then there’s security: The Browser Company handled its first major Arc security vulnerability in a surprisingly responsible fashion, but that was apparently a wake-up call. So it invested in its security engineering team–through personnel, red teaming, bug bounties, and internal audits–and now wants to set a security standard for small startups.

Miller says that the company has never stopped maintaining Arc, but it has stopped actively building new features, and the long-term future is unclear. The Browser Company has “extensively considered” open sourcing or selling the browser, but doing either would be complicated. It can’t open source Arc without open sourcing its Arc Development Kit (ADK) internals, and it’s using ADK for Dia, so it’s still core to the company’s value. While this is still an open question, cough, there are no plans to shut down Arc.

As for the future, Miller says that Dia is not a reaction to Arc’s shortcomings. Instead, Arc didn’t fundamentally change the web browser in the ways the company intended, and so it’s evolving to try to get that right.

“Traditional browsers, as we know them, will die,” he explains. “Much in the same way that search engines and IDEs are being reimagined [by AI]. That doesn’t mean we’ll stop searching or coding. It just means the environments we do it in will look very different, in a way that makes traditional browsers, search engines, and IDEs feel like candles — however thoughtfully crafted. We’re getting out of the candle business. You should too.”

While Miller doesn’t feel that the web is going anywhere, webpages will no longer be the primary interface to the web. Natural language interfaces to AI chatbots are “here to stay,” and webpages will “become tool calls” in which AI accesses the apps, articles, and files on the web. An AI web browser is like peanut butter and jelly, he says, or like the iPhone combined old categories together into something new.

After trying to radically change the web browser user experience with Arc and failing to some degree, Miller says that products like Cursor, an AI code editor, and AI chatbots, which threaten Google Search, showed him that evolving current interfaces was perhaps the better approach. AI is moving fast, but the user base needs familiarity.

“This is why we’re building Dia,” he writes. “It is the opportunity to chase the product of our original ambition: A true successor to the browser—maybe even the ‘Internet Computer’ we’ve been building toward all along—only in ways we couldn’t have predicted … the next Chrome is being built right now. Whether it’s Dia or not.”

Dia is currently available in alpha to students, he says, but it will soon be made available in open access to Arc members. I’m curious to experience it.