How serious of a problem is this for Kurzweil’s predictions? It seems like it would be fairly important to account for this (and of course add in a couple more decades for planning fallacy). But how much faster do computers need to get to upload a human brain or run an AI architecture?
It is not clear that anyone outside of the computer world is aware of this. The chip makers PR departments are trying their best to hide this fact. (“It’s a Core Two Duo Pro X2400! Don’t ask how fast it is!”)
The writer seems unaware of the right-hand turn.
Took me awhile to find anything on this. Here is an article from 2005.
http://www.gotw.ca/publications/concurrency-ddj.htm
How serious of a problem is this for Kurzweil’s predictions? It seems like it would be fairly important to account for this (and of course add in a couple more decades for planning fallacy). But how much faster do computers need to get to upload a human brain or run an AI architecture?
Moore’s law still holds. But you need to be able to write highly parallelizable code.
It is not clear that anyone outside of the computer world is aware of this. The chip makers PR departments are trying their best to hide this fact. (“It’s a Core Two Duo Pro X2400! Don’t ask how fast it is!”)