Our current approach to AI is massively inefficient. Human brain needs .3 kWh, even 1000x worse should be able to do with 300kw. Shouldn't we spend a trillion on figuring out how to do it better before burning the earth?

Good point. Let me know if you figure out how to measure

Wind back 10000 years I don't think you could show humans reason. Tool use fills the gap which is quickly developing in ai