The Odd Technical Battle Within the AI Browser Explosion
The Browser Company made some interesting claims about not being able to get the performance needed out of SwiftUI for Dia, but Atlas went all in on it
One of the points Josh Miller made with The Browser Company moving on from Arc was that they were limited by their choice of UI framework and that Dia needed to be a total rewrite built on AppKit instead of SwiftUI. Originally, Arc had gone all in on SwiftUI—so much so that The Browser Company spent a considerable amount of time porting SwiftUI to Windows so they could keep more of the codebase unified.
Ultimately, when it came time to start building Dia, they went back on all of this SwiftUI work and landed on Apple’s much older, but also lower-level framework for building apps on the Mac. I personally thought this was a pretty big blow to SwiftUI as BCNY was a major proponent of pushing the framework forward.
My apps are all written in SwiftUI and I don’t see myself switching away from that, but I’m also not building a browser. I like SwiftUI, but I will admit that, if you aren’t doing everything right all the time (which anyone who has ever written code before…you’re probably not doing everything perfectly all the time). When stuff like this happens, apps can load slowly, you can see a lot of stutter when scrolling, it’s a whole thing.
On the other hand, OpenAI is building ChatGPT Atlas on SwiftUI. The two products are incredibly similar, so it’s interesting to see OpenAI and BCNY disagreeing on tactics here. I’ll be interesting to see how this affects the progression of the two browsers over time but, as a user of SwiftUI, I really hope that OpenAI’s partnership with Apple combined with their choice to use the framework for something like Atlas will help bring the framework forward.
