I think that's the disconnect. Software isn't understood as patterns in hardware. It would be futile to try to explain software in these terms. And just as there are different modes of programming (e.g. imperative, declarative, inductive), this entails not only explanations that are far removed from the hardware in which these programs are instantiated, but also in the invention of different explanatory modes.

To your point, it's trivial to say that the software must interact with the hardware. Not only does "patterns in hardware" have zero explanatory power, it wouldn't be a useful explanation of software even if it were possible to discern.

We can certainly talk about software explanations without reference to hardware explanations. And as with other emergent phenomena, we'll find ways to do the same with the mind-body problem.