In Guns, Germs, and Steel, Diamond says that agriculture was invented at least 4 separate times and possibly (IIRC) up to 7. All of these happened within about a 5000 year period at the beginning of the Holocene.
The populations have been separated as long as 40K years, so it’s unlikely to be a genetic change. What’s the likelihood that a similar genetic change happened in three or four completely isolated populations? (Australia, New Guinea, the Americans, and the Old World)
Any necessary common genetic change would have to be 40Ky old, since that’s when the Australian and New Guinea populations split off. So, I don’t buy that.
I’m sticking with climate as being most significant, but another factor (mentioned above as “standing on the shoulders of giants”) could be that a certain level of cultural development (e.g., sufficiently sophistcated language) was necessary, and which could have developed naturally and independently in the various populations.
But still, 30Ky and no obvious ag, and then in a 5 or 6ky span, at least 4 separate cases of the same novelty, leads me to suspect a synchronizing factor, which would be climate.
Above, there are posts about the rapid rise of ag in the Americas. As stated in my post here, I agree with what I think is the point there. However, the spread of ag in the Americas was very slow (assuming Mexico-area and Peru-area development was independent; which is arguable but not conclusive). Diamond points out that ag spreads quite quickly east/west, but much more slowly north/south, due to climate. He shows that ag did seem to spread quickly in the Old World, and much more slowly in the Americas, and posits that the north-south topography of the Americas is the reason.
(BTW, Azimov would have loved Diamond’s approach to archeology, which is the closest thing I know of to psychohistory. And probably the closest that it’s possible to get, outside of fiction. )