> But this seems to me like a contrived example: how often in real
> code do people pass around these builtins, rather than calling
> them directly?

From experience developing PyPy, every argument that goes "this theoretically breaks obscure code, but who writes it in that way?" is inherently broken: there *is* code out there that uses any and all Python strangenesses.  The only trade-offs you can make is in how much existing code you are going to break -- or make absolutely sure that you don't change semantics in any case.