There is such a thing as personal variance, you know. The fundamental attribution error kicks in when you're trying to judge by single events. If you observe someone for a year, it's a lot more reasonable to expect them to be much the same next year. I would agree with the statement "people are oddly consistent", it's just that it's not so easy to judge from single events, and when we do judge we're likely to miss consistent background causes or consistent reactions, and think in terms of consistent personal attributes.
I'm always pleasantly surprised when I meet someone who actually improves over time in any visible way. I don't go around expecting it.
My understanding is that the per-minute risk of car travel and plane travel are roughly equivalent. Per mile, of course, planes are a heck of a lot safer. But if you wear your seatbelt in your car, you should be willing to sit at the back of a plane, right?
The Simulation Argument is not that we are living in a computer simulation. It is that you must reject at least one of the 3 propositions:
(1) we are not living in a computer simulation
(2) the human species will give rise to descendants that have the power to run lots of simulations
(3) at least a few of them will run ancestor simulations.
Which of these three do you think is untrue? Alternatively you could reject the Simulation Argument that says the three propositions are mutually incompatible - but if so, how?