Abstract
Unconventional computing rejects the assumption that computation is substrate-neutral, yet it has pursued that insight outward, into memristive, photonic, chemical, and biological media, while conventional silicon has largely remained its controller and benchmark. We argue that this control substrate is not physically neutral: ordinary CMOS is noisy, dissipative, hierarchical, and addressable. We propose that it be treated as a candidate unconventional-computing substrate whose participation is experimentally testable.


