Here at HN, I think most people see it differently (me included): having clear math proof of "confidentiality" is usually seen as both cheaper and more trustworthy.
Yes, there might be a breakthrough or a bug in encryption, and jnless you've been targetted, you can respond. But we've seen and experienced breakdowns in human character (employees spying on customers, stealing data...), government policies and company behaviour to trust the complexity and cost (lawyers) of enforcing accountability through policy.
In general, you do need both, but if you've got one, to engineers, technical solution is usually more appealing.
Yes, there might be a breakthrough or a bug in encryption, and jnless you've been targetted, you can respond. But we've seen and experienced breakdowns in human character (employees spying on customers, stealing data...), government policies and company behaviour to trust the complexity and cost (lawyers) of enforcing accountability through policy.
In general, you do need both, but if you've got one, to engineers, technical solution is usually more appealing.