For example, it's not impossible for the past to always be more different from the present than the future is; they could be exponentially approaching a stable self.

Yes, they report a decline in change with age [1], but even allowing for that, they find that people underpredict how different they will be. Age X predicts less difference in the next 10 years than age X+10 recollects in the last 10 years.

[1] It should be noted that these studies were not longitudinal, which introduces another class of possible error. They never compared any individual's actual traits at different ages, only different individuals of different ages at the time of the tests.