[AlexandrB]

> A second strategy out of bounds for privacy preserving products: Interoperability

This part had me scratching my head. Most established non-privacy-preserving products have been slowly killing interoperability because data lock-in provides a moat against users leaving and against potential competitors accessing valuable user data. There's no economic reason why privacy-preserving products should have worse interoperability than privacy-violating ones. Especially in product categories where interoperability does not imply sending PII to third parties.

Edit: Arguably, providing interoperability is easier for products that don't gather a lot of user data because there's less risk of an embarrassing leak of PII if the API is not properly secured.

[jessems]

> This part had me scratching my head. Most established non-privacy-preserving products have been slowly killing interoperability because data lock-in provides a moat against users leaving and against potential competitors accessing valuable user data. There's no economic reason why privacy-preserving products should have worse interoperability than privacy-violating ones. Especially in product categories where interoperability does not imply sending PII to third parties.

I agree. I don't see a strong economic reason this would be the case. But there's a strong practical reason (which perhaps has economic consequences): loss of control. If you make it easy for your users to interface with other services which don't have the same privacy guarantees, you're increasing the risk of their privacy being violated. If you implement an interface that's so secure, that no leakage or abuse is possible, then you win. But if that's not possible, and you end up restricting things that would otherwise lead to cool, productivity features, then you've hit the trade-off I touch on in the essay.