Great post. :) I think the simulation argument significantly reduces the expected computational power in our future (more discussion here), but the far future is probably still extremely important in expectation.