|
|
|
|
|
|
by sebastialonso
1944 days ago
|
|
|
But points against what? The very idea that a hyperintelligent being is _possible_ very much lingers in the land of science fiction. Yes, even today. Anyone with a little education in artificial intelligence knows that it's either impossible or somewhat possible but not likely. We truly are no closer to artificial consciousness that we were 500 years ago, it's all still smoke and mirrors. The thing that's actually a dangerous scenario is the dumb artificial intelligence we can build to start misbehaving, due to human programming errors or statistical bias: self-driving cars running over people; military systems confusing friends from foe; this stupid predictor keyboard not predicting correctly what I'm trying to type. This is the stuff of real, possible and demonstrable fears. Anything grandeur has no basis in reality. I haven't read Bostrom's book yet, but from what I've read from others about the book, it's all about this nonsense that we can create a new kind of intelligence who will instantly destroys us.
This is cool for people ignorant of the basic tenets of what A.I. is (so far) but for people who actually know or understand the current state, they should know better. Worrying about an hyperintelligent A.I is like worrying about ghosts while living in a neighborhood filled with serial killers. edit: grammar. |
|
|
Here is a survey (first one I googled) showing most believe it is relatively soon: https://aiimpacts.org/agi-11-survey/
I think if you want to critique Bostrom that is fine, but better to do so on the merits. His work is straightforward to read.