QUOTE(turetzsr @ Aug 21 2009, 03:19 AM)

...Umm, wouldn't it be the reverse -- we have even
less time, 'cause we've used up 2.5 million of our (maximum) 7.8 million?

<g>
...Actually, I'm willing to include in my ancestors the
Australopithecines, which reduces our maximum time left to about 3.9 million years - half!
Boy, are we ever O/T! But no, looking at the 'balance of probabilities' consequent on the 'we're in no special space and time' assumption (the Copernican principle) means we rule out the probability of being in the lower 2.5% of our lifetime - 1/40th - (lifetime of the species, in this case) because it is 'unlikely' and we similarly rule out the upper 2.5% so to leave the (most likely) inner 95% which accordingly ranges from 1/39th of our known existence to 39 times our known existence. So, if you increase the period of known existence you proportionally increase both ends of the predicted future range. Thus (the lower ends) if 200,000 -> 5,128 then 2,500,000 -> 64,103, etc.
"95% confidence" is the usual (minimum) acceptance level for an extrapolation/prediction, it is the 'practically safe bet', as used by pollsters - the honest ones
1 - and other statisticians when they need to draw defensible conclusions out of uncertain conditions. Bayesian probability estimates are said to produce similar results to these simple probability ones but I wouldn't know about that offhand.
In any case, the more you include non-human ancestry, the more you include non-human progeny/descendants (pre human->human->post human, Vs human ->human).
1Not to point fingers but from what I have seen pollsters in the US are the most honest around, in terms of statistical methodology and transparency. Probably because the US public is comparatively well-educated in such matters and would routinely bring them to task should they publish grossly misleading results. There may be bad pollsters in the US but I am definitely happy with the work of the ones I have seen (in terms of the 'basics', not considering questionnaire design which is actually critical). Some other places, maybe not so much.