I think he is generally consistent here: he wants papers with flawed proofs to be retracted, and he does not want papers to be rejected for not being sufficient

I think he is generally consistent here: he wants papers with flawed proofs to be retracted, and he does not want papers to be rejected for not being sufficiently significant (if they are otherwise correct).