The Little Arc of Browser History

The first mainstream browser, Mosaic, was released 33 years ago. A little earlier, in 1990, Tim Berners-Lee created the very first browser, WorldWideWeb. And if you look at it, there’s surprisingly little separating those early interfaces from the way modern browsers still look today.

Indeed, the feature set of a traditional browser hasn’t changed all that much since its inception in 1990. You can argue otherwise, but most additions over the years have been incremental rather than transformative. In my view, there have only been two genuinely big ideas: tabs and extensions. Everything else, bookmarks, history, toolbar, and so on, is essentially boilerplate, with the same interface patterns carried from one browser to another without a second thought.

I don’t ask much from my browser, and I definitely don’t need any baked-in AI features. I’m pretty content with the current feature set. What’s surprising, though, is how little fresh, thought-provoking thinking there’s been around the browser interface, and how we interact with it.

I’m not talking about AI assistants in so-called “agentic” browsers like Comet, Atlas, or Dia, or the half-baked AI sidebars showing up in the latest versions of Chrome, Firefox, Edge, or Brave. I don’t want to chat with my tabs, and I definitely don’t need yet another chat bar in my browser. As Julian Lehr puts it in his post The Case Against Conversational Interfaces:

A natural language prompt like “Hey Google, what’s the weather in San Francisco today?” just takes 10x longer than simply tapping the weather app on your homescreen.

What I mean are bold moves that rethink how we actually work in a browser, and how the browser works for us.

Enter Arc.

A product of its time (2023), when TBCNY hadn’t yet gone all-in on AI and wasn’t shy about taking ambitious swings at Chrome. It was full of genuinely innovative ideas. Not all of them stuck, of course, and some were eventually discontinued, like Arc Notes. Still, most were compelling, and I miss them every day.

I’m sure I’m not the only one who fell in love with vertical tabs because of Arc. Now every browser seems to be racing to replicate the experience. Chrome finally got them this month, in April 2026, but not a single implementation feels or works quite as well.

Swiping between separate workspaces with different profiles (!) was seamless, and I’m still surprised that no other browser has managed to reproduce this behavior.

Peek was a great way to skim Twitter and open linked articles without losing context. Easels, even though I didn’t use them much, offered an interesting take on whiteboarding that lived right in the browser. And Arc Max wasn’t just another chat bar with an AI assistant, it was a genuinely thoughtful, complementary use of AI within a modern interface.

With all that in mind, Little Arc is the feature I loved, and miss, the most. Just like Peek, it was a brilliant way to preview and triage links without opening the main browser window. Orion has a similar feature called Link Previews, but it doesn’t quite hit the same mark for me.

I believe this is why Arc became so popular and earned its cult-like status among users. It packed so many gems into one product. A glass of ice water in hell. But in his Letter to Arc members 2025, Josh Miller, the CEO of TBCNY, calls it the "novelty tax" problem and frames it as Arc’s main downside:

After a couple of years of building and shipping Arc, we started running into something we called the “novelty tax” problem. A lot of people loved Arc — if you’re here you might just be one of them — and we’d benefitted from consistent, organic growth since basically Day One. But for most people, Arc was simply too different, with too many new things to learn, for too little reward.

And I get that. When you’re set in your ways, it’s hard to adapt to something new. But I still don’t think Arc’s downfall came from its novelty or the sheer number of features. It came from something else entirely: VC money.

Once you take VC money, you’re expected to demonstrate constant growth, or else. Or you pivot to AI and suddenly it becomes much easier to raise more.

The state of web browsers in 2026 looks different: while the big players and AI-companies are busy adding AI everywhere, a new crop of smaller browsers - Orion, Helium, Horse, Zen, Pola, and others - trying to emerge and, at times, mimic a fraction of Arc’s power.

I understand that browsers like Google Chrome, Mozilla Firefox, and Safari serve millions of users every day, and that at this scale it’s hard to experiment. I understand the power of defaults and all that. But still, it’s hard not to feel disappointed by how little real interface innovation there is.

Arc showed us that something fresh is still possible. Unfortunately, it also showed that we’re all chasing the wrong things.