I wonder how advances in CPU design compare to slight variations in arrow design. In 10,000 years they might consider computer evolution to basically be mechanical, vacuum tubes, transistors, optical, then quantum. Even if we see huge advantage to smaller and faster transistors it's still mostly learning how to make a better version of the same thing.

I don't doubt that you can say "in the last 40 years" and you will still be right...

There's a most excellent vintage 1980s lecture by Robert Anton Wilson titled "The Jumping Jesus Phenomenon". The gist of it can be found at the rawilson.com website as well as elsewhere.

Wilson defines the "jesus" (small j) as the number of scientific facts known at the time of Jesus. Then he goes on to count how many years it takes to double that number. By his count, humanity accumulated 2j by about the year 1500. Then 4j by the year 1750, and so on. I forget the exact numbers but the point of the lecture was that our j-factor was increasing at an ever-increasing rate. By Wilson's estimation the curve would go vertical sometime around ... wait for it ... the year 2012.

This reads a lot like Kurzweil's singularity hypothesis (among others) but I'm not sure who developed the concept first.

It's Terence Mckenna's Timewave theory (from the 1970s).

Kurzweil was following along after a lot of other people.

Vernor Vinge, for instance, came up with the term Technological Singularity in 1993, Teilhard de Chardin had the Omega Point in 1950, etc.