How can it ever play against itself on novel software tasks? First it has to come up with the task. Then it can write tests but then it needs to verify that the

How can it ever play against itself on novel software tasks? First it has to come up with the task. Then it can write tests but then it needs to verify that the tests are correct, mixture of experts can come to wrong conclusions, etc...