That is how you break into vaults, fyi.
You can encrypt something that will need hypothetical quantum computers / processing until the heat death of the universe to decrypt without the key.
http://www.telegraph.co.uk/news/uknews/defence/8017041/MI6-Q...
> When forty spectators assembled for an outdoor trial, they reported that the safe seemed to be “on the point of explosion” and the gas issuing out of holes in the bottom of the safe meant it was “lifted some inches off the ground” forcing observes to retire to a “place of safety behind the building.”
Nothing really prevents you from having a safe that after N failed attempts from destroying the contents with explosives.
I would hope that a number of laws prohibit my carrying on my person a small safe filled with explosives and a known-effective trigger. Hypothetically, we could create such a safe.
Practically, however, we could not create one that would be as easily and broadly used as an iPhone. Therefore, the nature of the social question presented by a perfectly secure (for all intents and purposes, or at least a future iteration that is) mass market device is fairly novel.
This is where we disagree. What you've enumerated are attacks of convenience against a cryptographic implementation, not cryptography itself. And this is exactly what ease of access dramatically shifts.
The real twist isn't that Apple is suddenly providing quality encryption. We've had unbreakable encryption since the first one-time pad. IMHO, the FBI et al. didn't care because a statistically relevant number of people didn't use it. The twist is that Apple suddenly packaged that up into a consumer device with a quality implementation and all the hard details handled. And the FBI et al. do care because a very statistically relevant number of people use iPhones.
So yes, ease of access is a balance altering change. Because really, I don't think the government cares if hard encryption exists: it cares if lots of people use it.