I was not thinking primarily about code, but design documents, which are peer-reviewed for qualities beyond "correctness" - for example clarity and completeness

I was not thinking primarily about code, but design documents, which are peer-reviewed for qualities beyond "correctness" - for example clarity and completeness, which are important for communicating your ideas to many people. And even code has a subjective quality beyond "correctness", so your first claim is false and has no bearing on real-world software projects. OP was not talking about correctness and I am not talking about correctness, but the economics of collaboration and publication in academia.