They have enough employees to build native apps that run super quick but are still seduced by the web portability argument which, as we all know, is mostly untrue even now and which introduces all kinds of non-deterministic latencies/errors, which cannot all be handled neatly.
To be honest, this is the same in almost all apps that have any more than 10 developers working on them (my estimate!). Death by dependencies and a lack of coherent design.
As someone else said, though, some things like fastmail work OK in the browser so it is possible.
Native software is incredibly difficult to build well.
There are at least 4 platforms they would need to support: Win, Mac, iPhone, and Android.
That's 4 different software engineers at least, just for the frontend.
Then, there's various backend engineers, who could be shared, yes, but not always. Android's weird runtime requirements are bespoke enough that just because the database is written in C++, doesn't mean it's the same C++ database as what the Windows backend would use.
Finally, there's the designers, who end up consolidating all the unique things about each native platform into a common design language so they can have a shared vision on all of the platforms. So engineers end up building UI that works identically on all 4 platforms, and you're basically building a bespoke "browser" at that point.
> They have enough employees to build native apps
They'd screw those up as well.
The platform is not the issue. Competent engineering teams could blow this out of the water with a single threaded jquery web app.