so, since you guys are into it, I remember my RF prof mentioning some crazy designs where using bond wire inductance for VCO, any thoughts?

I think bondwire inductors can have somewhat (but not vastly) higher Q than those done in the metal layers (and perhaps not better than the redistribution layer). Bondwire inductors will be much less well controlled, as the shape of the wires will vary with technique rather than being fixed in a mask layout. Anyway we won't get any bondwires unless someone with a wire bonding machine (me?) decides to wirebond to pads on their chip, which I would prefer to avoid, as it seems like a lot of extra work per chip. Using bondwire inductors would make a design much harder to reuse, so I would see it as a last resort.