that is a pretty important part of the equation. what if the universe is the minimum viable machine for creating intelligence? if you think of the universe as computer and evolution as a machine learning algorithm then we already have an example of what size of a computer and how long it takes for ML to create AGI. it seems presumptuous to believe that humans will suddenly figure out a way to do the same thing a trillion times more efficiently.
>it seems presumptuous to believe that humans will suddenly figure out a way to do the same thing a trillion times more efficiently.
Nature isn't efficient. Humans create things many orders of magnitude more efficient than nature as a matter of course. The fact that it didn't take millions of years to develop even the primitive AI we have today is evidence enough, or to go from the Wright Brothers' flight to space travel. Or any number of examples from medicine, genetic engineering, material synthesis, etc.
You could say that any human example also has to account for the entirety of human evolution, but that would be a bit of a red herring since even in that case the examples of humans being able to improve upon nature within relatively less than geological spans of time are valid, and that case would apply to the development of AI as well.
> it seems presumptuous to believe that humans will suddenly figure out a way to do the same thing a trillion times more efficiently.
Why?
I think it might be confusion on your part on how incredibly inefficient evolution is. Many times you're performing random walks, or waiting for some random particle to break DNA just right, and then for that mutation to be in just the right place to survive. Evolution has no means of "oh shit, that would be an amazing speed up, I'll just copy that over" until you get into intelligence.
that is a pretty important part of the equation. what if the universe is the minimum viable machine for creating intelligence? if you think of the universe as computer and evolution as a machine learning algorithm then we already have an example of what size of a computer and how long it takes for ML to create AGI. it seems presumptuous to believe that humans will suddenly figure out a way to do the same thing a trillion times more efficiently.