Nobody claims that Turing machines in particular are well-suited to the task. What we claim is something more open-ended: Turing-complete languages can simulate

Nobody claims that Turing machines in particular are well-suited to the task. What we claim is something more open-ended: Turing-complete languages can simulate each other given a universal encoding (a program we typically call an "emulator") and the simulation is only polynomially slower and larger. Well, that's the Church-Turing Thesis, which most people accept but isn't being discussed by me or the OP - the topic being "theories of the mind", I think.