1. To deanonymise a hidden service connection you need to observe the traffic of all of the nodes in the circuit.
2. OK, let's say your adversary controls all of the nodes in the circuit and deanonymises the endpoints. Now what? You're no worse off than you would be if you weren't using Tor in the first place, so it's not an argument against Tor at all. All it's saying is "the absolute worst case of using Tor is no worse than the best case of not using Tor".
> the absolute worst case of using Tor is no worse than the best case of not using Tor
While this is true I just wanted to point out that one does in fact not need *all* the nodes. It is possible to perform traffic analysis and infer which nodes are used by a certain user even if the attacker only controls a part of the nodes. [1]
While this of course doesn't change the fact that using tor is a good idea, one should not let themselves be lured into a wrong feeling of security when using tor.
