imagine a hypothetical in which computer vision is perfected and the main challenge of self-driving cars is decisionmaking
-
-
Slightly more worrying: how many would answer in affirmative even if tradeoff clearly bad, if it reduces the monetary cost of the system?
-
how bout: tradeoff bad *and* more expensive but funnels pork to contractors in district and creates appearance of decisive action
-
The interesting point here is control. PHB control over complex software is very limited. Lots of oppos for sabotage and/or subtle influence
End of conversation
New conversation -
-
-
wouldn't this amplify system of negative selection, where at one point system would feel consequence of ousting outliers, and then will
-
design some kind of solution of averaging outlier exclusion to a minimum?
End of conversation
New conversation -
-
-
Great thread ty
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The role humans still need to play in things is to decide whether the technically correct outputs of these systems are moral to act upon.
-
We come up with immoral but technically correct solutions to things and reject them all the time. This tech SHOULD just make step 1 faster.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.