Evolvable hardware circuits and algorithms rarely take inspiration from biology beyond the idea of the genetic algorithm. I have applied key concepts of self-organization and complexity theory (primarily Stuart Kauffman?s "auto-catalytic sets" [1, 2]) to the problem of circuit ontogeny, in which a genotype becomes a phenotype through a process of growth rather than a direct mapping. I present two novel hardware architectures, one analog and one digital, as well as a biologically-inspired method of encoding the genome of a circuit, and processes for circuit growth. Among the findings are the remarkable result that "adult-hood" is simply a stable basin of attraction reached by iterating the (nonlinear) growth process, and that the same process continued past circuit maturity implements a form of fault-tolerance.