If you worry about humanity, you should be more scared of humans than of AI(thebulletin.org)
thebulletin.org
If you worry about humanity, you should be more scared of humans than of AI
https://thebulletin.org/premium/2023-09/if-you-worry-about-humanity-you-should-be-more-scared-of-humans-than-of-ai/#post-heading
9 comments
Yeah, Humans are very adept at killing other groups of humans
Only ever due to human ineptitude or intentional malfeasance, has a robot killed a human. Never of their own free will.
Life itself is the leading cause of death.
Life itself is the leading cause of death.
>Life itself is the leading cause of death.
It only takes a volcano, earthquake, tsunami, viruses (which are not living organisms) and other such events, for life not to be a leading cause of death in a region for a good while.
And let's not mention meteors...
It only takes a volcano, earthquake, tsunami, viruses (which are not living organisms) and other such events, for life not to be a leading cause of death in a region for a good while.
And let's not mention meteors...
Life -> volcano -> death
? -> volcano -> ?
No life preceding death, so no death.
Volcano and meteors are variables, while life preceding death is a constant.
? -> volcano -> ?
No life preceding death, so no death.
Volcano and meteors are variables, while life preceding death is a constant.
Scary, if true.
AI: scary
Humans: scarier
Humans + AI: scariest
Humans: scarier
Humans + AI: scariest
This right here.
US IC already using pattern of life analysis for target selection, wouldn't surprise me if they decided to use ML with "95% accuracy" to do that.
Best outcome: really good slaughterbots become cheap and every country gets them. Then the battlefield is mostly slaughterbots slaughtering each other. But still, when one side wins...
I mean, you gotta understand, this whole "oh no, civilians died" thing is fairly new. Japan, Germany and Russia in WW2 wouldn't have minded slaughter bots that use visual and hear signature based human identification to enable slaughterbots that simply kill anything that lives until their battery runs out.
US IC already using pattern of life analysis for target selection, wouldn't surprise me if they decided to use ML with "95% accuracy" to do that.
Best outcome: really good slaughterbots become cheap and every country gets them. Then the battlefield is mostly slaughterbots slaughtering each other. But still, when one side wins...
I mean, you gotta understand, this whole "oh no, civilians died" thing is fairly new. Japan, Germany and Russia in WW2 wouldn't have minded slaughter bots that use visual and hear signature based human identification to enable slaughterbots that simply kill anything that lives until their battery runs out.
I hear solar is cheaper and more effective than ever, these days.
In case you want your slaughterbots to last longer, I mean.
In case you want your slaughterbots to last longer, I mean.
So this is like saying "guns don't kill people, people kill people" (with AI in place of guns).