Regarding the future of Harman AIR, though I would absolutely love to stick with it, I have a few reasons to be a little skeptical.
First, I'll assume that the Harman team currently has the best intentions.
Part of the reason Harman got/purchased AIR from Adobe is that they themselves need and use it for their embedded enterprise solutions.
Now that they got it for themselves, they can also try to make a commercial product, offering AIR SDK for payment. The revenue should at least try to pay for the Harman AIR dev team.
How many AS3/Flash/AIR devs are there still? Veterans would most likely be willing to pay, and continue running on AIR. But what about new devs? How can Harman attract new devs? There's no official IDE and support for Animate is dubious (not to mention having to pay for Animate separately). The strength of AIR (coupled with Animate) is that it is probably one of the easiest and fastest dev platforms in the world, for non-hardcore coders.
Again, assuming the best intentions of Andrew and the Harman AIR team, if the AIR subscriptions revenue does not match Harman expectations, in 6 - 12 - 24 months, higher management might just pull the plug for the public side, and just keep it for themselves, for their embedded requirements, where probably nobody is forcing them to upgrade to 64 bit.
Just some thoughts.
Again, I love AIR and would love to continue using it, but Harman needs to step up and regain the confidence that Adobe completely crushed via their miss-management of AIR/Flash.