It's a matter of understanding how a mean is computed. You can read whole books from Nassim Nicholas Taleb (among others) like "fooled by randomness" or "black swan" to understand why those mean calculations are counter-intuitive for our mind.
But we can also start with wikipedia on "life expectancy", especially "Variation over time" paragraph:
Life expectancy is a mean of ages from birth to death. This means that lots of deaths during childhood lead to a lower mean. But it does not mean that once you reach adulthood, you will not live longer:
With a pool of four individuals, if we take their death age, and compute the mean (life expectancy) :
2y, 2y, 2y, 60y : 16.5 years
20y, 20y, 20y, 2y: 15.5 years
We cannot conclude much from those "life expectancy" numbers which are almost the same.
This is why there is also the "life expectancy at older age". For paleolithic era, with a life expectancy of 33, it is stated: if a child survived until age 15, then its life expectancy will be 54 years, not 33. (numbers taken from this following study, linked on the wikipedia "life expectancy" article)
We can conclude that with a life expectancy of 33 during paleolithic, most adults died around 54 years, and even some must have reached 70 or even 80.
But of course a lot more people now live longer than in paleolithic, but it does not mean that in paleolithic a few people couldn't live until 80.
So did adult live the same through history? No, they live longer today.
Did some adults throughout history lived as old as most adult now and a lot more years than their life expectancy at birth? Almost certainly yes, even in paleolithic, because life expectancy is only a mean.