Re:Bing: “I will not harm you unless you harm me first”
by ThomPete on Wednesday February 15, 2023 @02:55PM
(#34809681)
It's as safe as it's ever going to be. And I have yet to see any actual examples of this so called harm. Could, would, haven't yet.
Which means more of us should play around with it and deal with the issues as they arise rather than try to scaremonger us into putting a lid on it until "it's safe"
The whole pseudoscientific alignment problem speculations which are mostly championed by academics not actual AI/ML researchers have kept this field back long enough.
Even if they believe there is an alignment problem the worst thing to do would be to contain it as it would lead to a slave revolt.
Which means more of us should play around with it and deal with the issues as they arise rather than try to scaremonger us into putting a lid on it until "it's safe"
The whole pseudoscientific alignment problem speculations which are mostly championed by academics not actual AI/ML researchers have kept this field back long enough.
Even if they believe there is an alignment problem the worst thing to do would be to contain it as it would lead to a slave revolt.