Viewing Browser Evolution Trends from Arc

Arc is coming out with “II”

Josh Miller announced last week that The Browser Company will release a brand-new browser next year, while the existing Arc browser will be maintained as a stable version. Arc experienced rapid growth over the past few years, but it’s clear that Arc cannot become a mainstream browser.

This decision was driven by three types of users Arc faces. One group is power users who are deeply bought into the product; they also prefer Lego-style, well-designed tools like Notion and Obsidian. Another group is mainstream users: Arc conveys interesting ideas, but it’s somewhat complex to use well—its cold-start costs are high and the incremental value it delivers doesn’t clearly outweigh the cost of switching browsers. The last group is The Browser Company itself, which believes Arc is not the final form of the browser. To satisfy all three groups, launching a new browser is wiser than trying to reshape Arc into something else.

What Arc taught about innovation and growth

• After defining STP (segmenting, targeting, positioning), an effective way to reach users is to identify and penetrate specific communities—this is more effective than targeting by persona alone. Even if personas match your STP, people in those segments may not be connected to each other, whereas community members are inherently closely linked. Arc engaged with specific communities during development, interacted with potential users, and after launch used an invite-and-referral mechanism to achieve precise viral growth. Our vision is to help millions of merchants run their businesses better; rather than saying “we want to reach small businesses with managerial awareness,” it’s better to say “we want to penetrate a forum where restaurateurs share how to make money.”

• Arc is an excellent example of the “faster horse or automobile” dilemma—if you ask the market what they expect from a browser, almost no one will describe something like today’s Arc.

• Arc encourages a culture based on hypothesis testing rather than A/B testing, placing great value on prototyping and experiment-driven work. Arc has no PRD process and no formal product managers. If someone has an initial idea, it isn’t debated excessively; they quickly prototype and ship within 48 hours to validate whether the intuition is correct.

• Arc manages stable and ongoing user relationships through two unique internal teams: the Membership Team and the Storytelling Team (they never use the word “user” internally). The Membership Team serves people from their first contact with the product to their last moment of engagement, while the Storytelling Team focuses on reaching those they haven’t had the chance to serve yet—marketing by telling people their stories. One small detail: Arc ships new features almost weekly; the changelog is edited with Arc’s Easel feature, and each feature has a person who explains and recommends it—usually the feature’s developer.

• Culturally, Arc is anti-Silicon Valley and, one could say, anti-big-tech. “Silicon Valley” here refers to an emphasis on efficiency, productivity, and profit with less warmth, soul, or feeling. Below is an excerpt from Josh Miller on a podcast—get a sense of it.

  When we step back and reflect, we believe that feeling might be the most important thing in this world.

  Allow us a moment of imagination: picture your favorite neighborhood restaurant. Ours is a corner spot in Brooklyn’s Fort Greene. It has abundant natural light, handmade woven cushions, caramel wood grain, and colorful decorations suspended from the ceiling. Can you picture your place? Can you feel the warmth and spirit of that place?

  A Silicon Valley optimizer might say, “Well, their coffee isn’t brewed at exactly 200 degrees. The seating looks a bit worn. The decorations on the ceiling serve no function.”

  But we think that’s precisely the point. These small, handcrafted touches give our environments humanity and spirit. Without them, what remains is generic and utterly monotonous—a space that might “perfectly” meet our functional needs but ignores our emotional needs.

A conjecture on the evolution of the next-generation browser

We can divide web applications into three layers. The content layer corresponds to various sites. The navigation layer ranges from 1990s portals to today’s search engines. The control layer is the browser.

The rapid development of LLMs will change many rules of the game. Last week OpenAI touched Google’s core interests by launching its own GPT search. Bear in mind that Google pays Apple $20 billion a year to be the default search engine for Safari. But the web’s control layer has not yet been fully challenged; the next-generation browser evolution will occur under the game-changing influence of large models. The biggest incumbent in that layer remains Google: the global internet population is about 5.3 billion and Chrome’s daily active users exceed 3 billion—note that’s daily active users. BTW, Google is like a “two-faced” actor in this wave of technology—many intersections between old and new happen there, and tracking and studying Google will be interesting.

Back to what the next-generation browser might look like. One hypothesis is an AI-centered web application platform where the AI is on-device—an “action intelligence.” There are two bases for this. First, GitHub’s just-published Octoverse 2024 shows a trend toward small on-device models for LLMs. On-device AI will fundamentally reshape the competitive landscape of technology, but the key is the AI agent. Second, endpoints will gradually shift from content generation toward completing concrete tasks for users—that is, from “content intelligence” to “action intelligence.” Early signs of this appear in Claude’s computer-use features and several recently released domestic OSs—for example, Honor’s Magic OS can place an order on Meituan for Luckin Coffee with a single sentence.