>But I'm unconvinced that this is a good measure of software complexity; for one, there are only three nontrivial complexity levels (2, 3, 4) But can we even r

>But I'm unconvinced that this is a good measure of software complexity; for one, there are only three nontrivial complexity levels (2, 3, 4) But can we even reason beyond 4 levels? This point made a lot of sense to me. My intuition is that this comes down to our fundamental inability to generally solve quintics. There are also topological proofs to this which fit the author's point perfectly [0]. 4th degree is the limit of formal proof. [0] https://en.wikipedia.org/wiki/Topological_Galois_theory https://en.wikipedia.org/wiki/Topological_Galois_theory