The main impact of whole brain emulation will probably come long after we have superintelligence, when origins emulations (to help predict the actions of unknown aliens) become important.
Whole brain emulation impact earlier seems likely to be low—few will want to emulate a crappy human brain once we have superintelligence.
That’s Eliezer’s claim, but other people think that EM’s might be the first type of true artificial intelligence. It depends how far we are away from solving the software problems of creating a GAI from scratch, and the truth is no one really know the answer to that.
The main impact of whole brain emulation will probably come long after we have superintelligence, when origins emulations (to help predict the actions of unknown aliens) become important.
Whole brain emulation impact earlier seems likely to be low—few will want to emulate a crappy human brain once we have superintelligence.
That’s Eliezer’s claim, but other people think that EM’s might be the first type of true artificial intelligence. It depends how far we are away from solving the software problems of creating a GAI from scratch, and the truth is no one really know the answer to that.
It depends on an ability to compare the difficulty of the two paths.
To me this seems to be a much more practical task—even if it is relatively challenging to put an absolute scale on either of them.