Arc, which has been my web browser of choice for the last (almost) 4 years, is well on its way to the Great Pile of Dumped Software In The Sky. Its creators are “sunsetting it” in a couple of months to focus on Dia, their AI-first browser.
It will continue to work, of course, but the way that online threats are advancing I would not advise anyone to use an unmaintained browser.
The farewell letter is signed The Browser Company of New York, but this is obviously not their call — it’s Atlassian’s decision; they bought the company a year ago.
And you know what? In Atlassian’s position, I’d probably have made the same call.
The question is… would I have bought it?
To be clear…
Before we get too far: I have no insight on how this went down, I don’t think I know anyone at either Atlassian or TBC, and I don’t have money riding on either company. I’m just an online rando walking himself through the facts.
With that out of the way…
Got half a billion to spare?
What got me thinking about this is that Laurens Hof described The Browser Company’s focus on Dia as a prime example of AI psychosis.
Arc is on the shortlist for 'company with the dumbest version of ai psychosis' award imo, killing off their own beloved product to make a much worse version of the same thing but with ai
arc.net
— Laurens (@laurenshof.online) October 23, 2026 at 11:31 AM
[image or embed]
My retort was that Dia clearly was a calculated move: they announced an AI-focused project, back in mid-2025, when consumer AI froth was at its peak, which only a few months later led to an acquisition by Atlassian.
When I first heard the news, I was sure Atlassian had acquihired them after picking up the company for chump change. Turns out I was wrong — Atlassian paid $610 million in cash.
That’s a 5x exit, for a 5-year-old company with no revenue to speak of, whose main product for all intents and purposes competes with Google’s Chrome, Mozilla’s Firefox, Microsoft’s Edge, and Apple’s Safari. Not only do those companies not need to charge for their product, but in three out of four cases, the companies own an operating system that uses it as the default browser.
With distribution like that, who in Cthulhu’s darkest of the hillside thickets would pay that much money for a browser?
IBM, of course.
Enter The Dino
Wait. Geezer slip. I meant Atlassian.
Atlassian is a huge name in the developer tools space. They’re also a public company, valued at about $56Bn. Someone inside with enough clout had to go out of their way to convince their board that it was valuable spending a full 1% of their market valuation, in cash, to buy a loss-making company with no income prospects targeting a saturated market at 5x the amount that investors had put in.
Now, why would you wanna do a thing like that?
This is not like Microsoft acquiring GitHub. For starters, Microsoft acquired GitHub for stock (even if at a similar relative valuation, ~$7Bn for GitHub vs $720Bn for Microsoft). And beyond that, not only did GitHub have actual paying customers, but Microsoft already had an entire enterprise sales machinery they could put to work by adding GitHub services to their packages.
Atlassian is focused on dev tools — they probably couldn’t tell what the consumer market wanted if it wrote them a check and stapled it to their forehead with a purchase order.
So if we discount stupidity, what makes that make sense?
(Because you don’t survive for 20+ years, going toe-to-toe with both established players and open source software, while being stupid.)
My money is on fear.
Be Afraid. Be very afraid.
OK, so it wasn’t a slip.
Jira used to be The Go-To Issue Tracker, way back in the day. Two of my startups even paid for a license, and one even used Confluence for a while (before we decided it was overkill). It had features no other issue tracker had at the time, but more importantly, it was smooth and easy to use. Their self-service licensing strategy was no-nonsense.
Then it got bloated, slow, cumbersome, and expensive. And then some. And then they started acquiring companies to grow their user base and jury-rigging integrations between the products, which didn’t make things any better.
The way I hear it, a lot of devs are stuck in Jira and Confluence just because “it is what we use”: if your company has enough custom workflows built on top, the migration cost is non-trivial, and it’s easier to just get the licenses renewed than to fight for the budget for an internal project (not to mention the reputation risk if the migration goes to hell).
In other words, the main thing Atlassian seems to have going for them right now is inertia.
Sounds familiar?
In Atlassian’s position, I’d be looking around the space very nervously, wondering if the AI-led unbundling that came for IDEs is coming for you too.
Maybe you believe that you have the best issue tracking, code hosting, and enterprise wiki design in the business, full stop.
Does that matter in a world where agents are increasingly choosing the software they use, and where developers more and more don’t go to the issue tracker to click on a thing, but tell the agent prompt what to file for them and how to move the issues around?
Nobody ever got fired for buying IBM… for a few decades there. But how did that work out for IBM when they waited too long to enter the cloud business, and then tried to do it with their long-honed AS/400 sensibilities?
Lifeboats
In comes someone at TBC or one of their investors.
They tell you that the work is not only increasingly on the browser (which you already knew, because your bread and butter are web platforms) but on a chat interface.
They tell you that your user interfaces are increasingly irrelevant, because devs are mostly using agents (which you probably knew, if you’ve been listening to your own developers).
They tell you that agentic work is fluid, not rigidly structured. They tell you that in the future, everyone’s a developer, because the tools will progress fast enough to make custom code the future (which makes sense to you, since Jira’s customizable workflows were a key selling point). They tell you that not a single one of those newly minted developers has heard of you nor is likely to be willing to put up with your tools.
They tell you that the new market needs consumer-focused design expertise. And, wouldn’t you know it, they have this company right here that is making sure AI works for users who can’t tell a Jira from a Codex.
And you can have it for a tiny fraction of your valuation.
The message is that you are not a lifeboat for them — they are a lifeboat for you (even if a seller wouldn’t insult Atlassian by phrasing it that way).
So yeah, I’d probably have bought them too, and once I had given Arc — a gorgeous but unproductive horse — enough time to further convince me it’s not going to carry its weight, I’d have killed it too.
And who knows? The AI play may end up paying off for Atlassian. It’s too early to tell.
But there are scenarios where buying TBC makes a lot of sense and, sadly, there are even more where killing Arc does too.