What Would Superintelligence Look Like?

I have been known to mock those* who think that selective breeding plus maybe a bit of CRISPR could produce humans with IQs of 1000. Whatever that means.

Whatever does that mean?

My only, or at least main, guide, is what I would wish to be able to do better, intellectually speaking. I'd like to have a better memory, an eidetic memory, that could recall everything I had learned. I'd like to be able to understand complicated arguments and equations at sight, and I'd like to see remote analogies and connections better and faster. Most of these seem to reduce to better memory and faster processing.

Does that sound like anything that we've ever heard of?

If you said computers, you might be right, but there are still some things people do that computers can't do (yet!). The most important is constructing sophisticated generalizations from what they know, and storing them as meta knowledge. I think that we humans usually do that by creating narratives - from Genesis to Darwinian natural selection to geometry. I have very little idea how we do that, but a supermind ought to be good at it.

My latest excuse for musing on this sort of thing is rereading a bit of A. E. van Vogt. His superminds seem to have super speed and super precise senses.

*Like Steve Hsu and Lumo.

Comments

Popular posts from this blog

Inequality and Technological Progress

Technological Advance and Capitalism