Jason Kuznicki at Cato Unbound:

To my mind there are two ways to do libertarian activism.

One approach is easy, deeply satisfying, and - at least on our current margin - it’s basically ineffective. The other approach is difficult, usually thankless, and - I dare say it - revolutionary when it works.

Let’s call the first way libertarian moralizing. We know it by what it aims to produce: The intended product is more libertarians. Eventually we’ll persuade everyone, or at least enough of everyone, and then we’ll change the world.

[…]

To pick a completely incendiary name, I will call this second type of activism libertarian social engineering. By this I mean the deliberate attempt to create, on an incremental, case-by-case basis, the new, voluntary institutions and practices that a society would need if it were to become significantly more private, more decentralized, and more free. I mean here institutions like cryptocurrency, which is already well known; private institutions of assurance and trust in consumer satisfaction and safety; and Alexander Tabarrok’s idea of the Dominant Assurance Contract, which is exceptionally obscure, but which stands to my mind a fair chance of making almost all state action obsolete.