I think it's a near perfect solution to the problem of recording and transferring ownership of deeds and titles.

In what way do you think it is superior to the current system (say, an Oracle RDBMS running in a county land office)?

Do you think the current security risks are an impediment to it being a perfect solution? I don't want my house deed transferred to someone else just because the provider that managed my wallet that held the deed to my house got compromised?

Wallets are not an essential component of a blockchain. Blockchains are simply a linked-list of data, with each new block of data ensuring the integrity of its previous block. That's it.

If you want to verify that a particular operation on the data was done by the person who is allowed to do it, you need some sort of public key crypto, right? If someone steals your private key, they can pretend to be you on the blockchain. Whether that private key is called a wallet or something else.

There was some buzz recently about Wal-Mary using blockchain to track pork through the supply chain. I thought it would be cool if they e.g. created tokens by hashing a pig's DNA.

That's not really true. We had hash chained data structures for ages before BTC came out. Nobody retroactively called these "blockchains".

Not with the current energy consumption.

Not all blockchains need to use energy-intensive proof-of-work mining to reach consensus.