3- to see long-term potential in bitcoin is to come to the conclusion, already expressed in the original paper, that it is irrational for an actor or cartel to perform a 51% attack on the network, as it costs more to successfully carry out the attack than what can be expected to be gained by it.
Well, I don't agree with that.
The "proofs" I have seen usually assume that the cartel has some specific goal (say, pulling a double spend), and a specific simple-minded strategy, and argue that
that attack would not pay off, of
that strategy would fail.
However, there are many other goals that a cartel may have, and better strategies to carry out the "attack". I can see some "attacks" that would probably succeed, would have to be accepted by the users, and would results in substantial payoffs for the cartel. But I have worn my fingers arguing this issue in other threads, so please forgive me for not doing it again here.