I’ve heard AI folks talk about verifiable domains (like coding) as being more suitable for applying AI. Do you address that in your textbook? They seem to mean that a verifiable domain is one where the computer can check validity automatically and with high confidence.
