Smart contracts are amazing, as long as its a contract between two crypto people doing crypto stuff. But (and please, someone correct me if I'm wrong) the oracle problem still stops them from using anything in the real world as an input; I think this is what GP meant by naval gazing.

The example contract in the link I posted doesn't rely on third-party data sources (with the exception, perhaps, of time).

Does the oracle problem still apply?

It would seem that by definition, if you want smart contracts or anything "on the blockchain" to interact with the real world, you will need to find intermediaries you trust. Meaning you're back to square one if you want to cross the crypto-real world membrane.