VST3... tastes just like chicken...
Let me state the following for the record, since if you're not a developer it may not be obvious: You, The User, gain NOTHING from VST3. This should be clear from the outset. The marketing points in the Steinberg press release are referencing shit that is included, almost entirely, in VST2.4. There is nothing in the VST3 kit that makes anything better, faster, stronger, lower CPU, lower latency, more likely to get you laid, anything. Don't believe the hype.
Take some time to get comfortable with that concept. Go ahead. I'll wait.
It is clear, to me at least, that this entire Switch rests firmly on what Ableton does. If they decide to support VST3 next week, and VST3 only (and there's no reason, historically, to suppose that they won't), then we'll figure this shit out in short order. If they don't, then Steinberg can go fuck themselves sideways, basically. Move along. Nothing to see here.
Of course, it isn't as cut-n-dried as that, but that's the basic feeling in the industry, at least among the people I communicate with. That said, and taking in to account my initial issues with the kit, it doesn't look to be that complicated of a process to port our shit to VST3. Not any more complicated than porting it to AU, anyways. So if we must do it, then it can be done. But as far as I can tell, there's no pressing need. None of our products will see any gain from moving to VST3, as we tend to design with a lowest-common-denominator feature set, as far as communicating with the host goes, so that we can build both AUs and VSTs from the same code base.
For what it's worth, through all the changes on the OS X platform in the last 3 years, we've had to do nothing on Windows. I suppose you don't get a free ride anywhere these days. I truly feel sorry for someone just getting in to VST development this week, though. They're in for a bumpy learning experience.