I was imagining mobile operators that cooperated to some extent with the changes I was proposing, or at least didn't obstruct them. If it's using existing GSM protocols, the IMEI would have to be rotated frequently (and it's not that obvious how to do that without making the connection between the old IMEI and the new IMEI apparent), and the SIM technology would have to change. (What it's trying to prove in a privacy-friendly communications system is more like subscriber entitlement, not subscriber identity!)

There's also the "netheads and Bellheads" theory from the 1990s which can be taken to say that phone companies would never make technical changes to make themselves collect less data, or to be less helpful to government surveillance. Sometimes I think this is right. I still remember how I took part in a meeting with a mobile phone industry association or industry consortium of some sort about a year before the Snowden stuff. Someone on my side said "so, let's talk a bit about surveillance issues", and someone on the other side replied "sorry, that's something we don't talk about". Imagine an industry meeting with privacy advocates where the industry people are completely precommitted to not talking about surveillance!

> There's also the "netheads and Bellheads" theory from the 1990s which can be taken to say that phone companies would never make technical changes to make themselves collect less data, or to be less helpful to government surveillance.

You've got to sell them on something that's useful for them. Present the case that eliminating data collection simplified their network, saves money, reduces staffing, and reduces interaction with government.