To the OP, I suggest you launch with 3.x and upgrade immediately after to 4 and start taking note of deprecations.
Yeah, but that pain will come regardless. Better I think to wait until the dust settles on turbolinks et al and plugin authors update. You could always update first to 4.0.x (for mild to large values of x) or what have you before jumping over to 4.1.
Point is, if you've got a large app it's not worth pushing back any deadlines. Someone's gotta do it! and much to love to the rails-core team, but being 3 months behind the release schedule still counts as "bleeding edge" in most of the rest of the software world ;).
I think this is the most likely outcome. Thanks again to everyone for the discussion too.
If anyone has gems that don't work with 4 yet, please let me know. Most of them at least have a branch going, but I'm always curious, and am willing to help out to upgrade things.
If you're on 2.3.x upgrade, otherwise wait a bit.