Tuesday, October 14, 2014

Superintelligence, simulations, etc.

There's an interesting review of Nick Bostrom's Superintelligence: Paths, Dangers, Strategies at the website of the Institute for Ethics and Emerging Technologies, by piero scaruffi (whose name appears in lower case). I haven't read Bostrom's book, though I have been keeping tabs on the subject (see here and here) and found scaruffi's skeptical take on the dangers of AI intriguing. I then went to scaruffi's own site, and from now I will no longer worry, as I sometimes have, that my own blog incorporates an unduly broad range of interests; it's nothing compared to scaruffi, and I say that respectfully.

On another Bostrom-related note, I was a bit disappointed by Seth Shostak's piece "Is Life an Illusion?" which is an uncritical take on Bostrom's ideas about living in a simulation. I would've expected Shostak, who's spent decades contemplating probabilities involving alien life, to have something more sophisticated to say about efforts to calculate whether we are characters in a computer game. Years ago, I expressed some reasons for skepticism about that subject, which strike me as still making sense. (I don't, however, claim to be up to date on the technical debate over the simulation argument, and I see that Bostrom in 2011 co-wrote a "patch" for his original case.)

UPDATE: "Elon Musk: Robots Could Delete Humans Like Spam." Strikes me as not a compelling analogy, given how difficult it is to get rid of spam.

No comments: