The point is that you can prove that a thing is self-consistent without proving it is true. And I think what are calling an "infinite regress" is, in this case, self-consistency.

Proving self-consistency is a decent achievement, and is certainly error. But it is only weak evidence in favour of a position.