One of the largest hurdles in my mind to increasing the number of full nodes operating is the startup cost related to downloading the blockchain. A similar but less important (in my mind) issue is the amount of disk space required by the full blockchain.
In Would moving a transaction from an old block to a new one allow more pruning? Murch suggested that some assumptions be checked, and I think one of them was my assumption that it would be safe to rely on a UTXO Set provided by other users. Given the answers to another assumption Murch suggested I check, it seems that the answer is "Not usually," but if there were a standard proof, such as a signed UTXO set or a UTXO Committment in the header, then the answer seems to change to "Yes!"
In any case, Murch suggested that the assumptions be checked, so I've created this question to check that one.