Mathematical systems are outside the purview of physics. I can make true statements about infinite series and n-dimensional spaces all day long and never have to actually go point at any in the real world.

Except... the decidedly non-trivial exception to this rule is that whatever is actually implementing the mathematical reasoning is constrained by physics, whether it's my brain or an i5 quad core.

Which is how I interpret the quote: the internal rules of abstract systems are not subject to physics, but the systems themselves sooner-or-later are. This is particularly relevant to programming because you can use non-physical abstractions if and only if you (or the people who wrote your tools) can explicitly represent them with physical processes.