As I've been down the customer discovery rabbit hole for @StoffelMPC, one thing that comes up time and time again that I don't see being addressed is differential privacy. I know many of you, especially pure cryptography folks, eschew differential privacy. But in web2, it's singlehandedly the most widely deployed privacy tool to protect individual-level privacy at a wide scale. Cryptography alone doesn't solve this at all. A simple example to demonstrate is the multiplication of a set of numbers held by 2 parties. 2 parties want to multiple their individual numbers without sharing their number with each other. They can do an MPC protocol in order to achieve this. At the end of the execution of the MPC protocol, each party gets the final result of the multiplication. However, since each party knows what their own number was, they can easily find the number of the other party! The above example is contrived but illustrates that even with the use of advanced cryptography, one is able to deduce an individual level input. This is where differential privacy comes in! Naively, if differential privacy were applied in addition to the MPC protocol, then each party would add some noise to their individual inputs and then undergo the MPC protocol in order to find the multiplication of their numbers. Now, with the result, they wouldn't be able to easily deduce the exact number the other party had because of the additional noise that was added! Obviously, there's an issue here in that the noise can introduce inaccuracies in the final result but there are practical ways around this. Hopefully, you can get a sense of why Big Tech loves differential privacy and why it's as widely deployed as it is. As orgs get more serious about other forms of privacy, they are going to be looking to combine advanced cryptography in addition to differential privacy as opposed to swap one for the other. TLDR: Even advanced cryptography needs differential privacy