In my experience it's simply developer resources. In my entire career, I've never seen a complex web app "just work" across multiple browsers without some, if not a lot of, effort.
Testing on all browsers all the time is a significant cost to velocity. It almost always makes more sense to build for a single browser first, then figure out where the comparability issues are once you have an actual product.
> It almost always makes more sense to build for a single browser first, then figure out where the comparability issues are once you have an actual product.
I agree, and most places I've worked with has been the same. But I think only once, after following that process, have we ended up in "Damn, this works 100% in Chrome but is completely broken in Firefox". Sometimes something works faster/slower in Chrome/Firefox, but without requiring something like the Web MIDI API, most things just works the same across browsers, minus some minor things regarding CSS that has to be taken into consideration.
I guess what I'm looking for with my previous comment, is not so much of a general/philosophical answer of "Why you block Firefox but not Chrome?" but rather a specific answer to "What ended up being broken in Firefox but working in Chrome?"
Mostly helpful for my own curiosity, but in the strange case of finding things that works in Chrome but not Firefox, leaving a report at https://webcompat.com/ can be very helpful for the general health of the web as well.
In my experience it's subtle bugs that don't appear on one platform but do in the other. webcompat is great for true deviations from the spec but there are plenty of cases where things work for no good reason and it's a chore to figure out why.
I'm also a bit curious why blocking Firefox is preferred versus stating which browsers are supported.
Testing on all browsers all the time is a significant cost to velocity. It almost always makes more sense to build for a single browser first, then figure out where the comparability issues are once you have an actual product.