Trusting without understanding
The problem with a superintelligence is not that it would be smarter than us. It is that its reasons might be true, and still out of our reach.
The problem with a superintelligence is not that it would be smarter than us. It is that its reasons might be true, and still out of our reach.
Aristotle already told apart the ability to reach a goal from the ability to know which goal is worth it. Our machines excel at the first. The second is still our job.
We ask machines to prove that they are conscious. Let us turn the question around: could we prove that we are?
If machines outdo us at almost everything that produces a result, what remains is whatever is valuable in the doing itself. Perhaps that is where meaning always was.