[rossdavidh]

So, if we looked at all of the "oldest yet found" discoveries (that held up to later scrutiny), and plotted their date vs. the date they were found, would we see it asymptotically approaching some (presumably actual true) value? A job for an enterprising data scientist/archaeologist out there.

[garmaine]

What would that show?

[rossdavidh]

Perhaps, what the actual value could be expected to be. So, for example, you don't do a lot of work on top of the current estimates for oldest tool-using humans outside of Africa, that is likely to be overturned in ten or twenty years.

[garmaine]

It wouldn't be convergent.

[rossdavidh]

Interesting if true. I could easily believe what you say, but do we know that? I don't know the field well enough to say.

[garmaine]

Yes it is rather trivially true. The age is getting exponentially further while discoveries are happening faster.

But that reflects improvements in technology, not anything about what’s being measured.