Dia: missing essential features and agentic power

When The Browser Company announced Dia, their new AI-powered browser, I was curious. As someone who switched from Chrome to Arc and fell in love with its innovative approach to browsing, I wanted to see what they'd come up with next.

But after trying Dia yesterday, I'm left with the question: who is this for?

The Arc experience

Arc really speed up my way of working and how I interact with the web. The sidebar navigation for both spaces and tabs and the ability to treat tabs like mini-apps made browsing feel more organized and intentional. Especially the way any browser window of Arc is just a view on the actual state of your browser is an amazing feature.

For me, Arc is always open with three spaces and a lot of pinned tabs, but my daily tabs are gone the next morning to give my day a fresh start. Anything that can get me distracted is gone. The shortcuts are well thought out and super useful. Even tiny features like 'Air Traffic Control' are a game changer.

What Dia promises vs. what it delivers

Dia markets itself as an AI-first browser that can "talk to your tabs." The core feature is essentially having a conversation with an AI about whatever webpage you're looking at. You can ask it to summarize articles, explain concepts, or help with research.

That's... it.

Sure, it works. The AI can read your tabs and answer questions about them. But this feels more like a browser extension than a completely new browser. It's the kind of feature that ChatGPT's browser extension or Perplexity could add tomorrow.

I am missing some obvious features. So the browser can read the webpages for me that are open, but it cannot actually 'browse'; open new pages on the website, dig a little deeper or do an action with some form of tool call. Like a chrome extension could press a button or put some text, you still have to manually do that in Dia.

The experience that drove me back to Arc

My journey with Dia was predictably short:

Install and go through the polished onboarding
Try the AI chat features - they didn't really work as advertised
Look for easy space and tab navigation - it's not there
Wonder where all nice Arc settings and shortcuts went - they're gone
Close Dia and open Arc again

The problem isn't that Dia is bad. It's that it doesn't do anything to replace what made Arc special. If you're switching browsers, you need something that's either clearly better or solves problems your current browser can't handle.

A browser caught between two worlds

The Browser Company seems to have concluded that Arc was "too different" and "too complex" for mass adoption. They wanted something more like Chrome that regular users could easily switch to.

But here's the thing: if you want Chrome, you'll just use Chrome. Google has infinite resources, the best web compatibility, and years of refinement. Making a Chrome clone with AI features isn't exactly a compelling value proposition.

Meanwhile, if you're someone who appreciates innovation in browsers - the kind of person who was drawn to Arc in the first place - Dia feels like a step backward. It's trading everything unique and powerful about Arc for AI features you can get elsewhere.

The inevitable pivot

This move feels predictable for a tech company in 2025. AI is the buzzword, AI gets funding, AI gets attention. But building a browser around AI chat feels like putting the cart before the horse.

The browsing paradigm is already changing, but not necessarily in the way The Browser Company thinks. Tools like Perplexity and Deep Research become the go-to research aids and this feels just like something similar, but not powerful and agentic enough.

They make it seem like browsing the web is the only thing that people use a browser for, but that's not the case. I think the majority of people use browsers to do work in online apps, environments, spaces, etc. Browsing is not just reading, it's mainly doing work. I don't need AI to help me read my tabs, that feels like 2024.

Looking ahead

I hope The Browser Company proves me wrong. Maybe Dia will evolve into something genuinely compelling that justifies the pivot away from Arc's innovations.

But for now, Dia feels like a browser built for no one in particular. It's not different enough to win over Chrome users, and it's not powerful enough to satisfy Arc enthusiasts.

Until then, I'll be sticking with Arc and hoping it doesn't get abandoned entirely in favor of the AI trend of the day.