Content
@
0 reply
0 recast
0 reaction
Dan Romero
@dwr.eth
In you're in favor of "AI safety" (broad definition), what's your most compelling cast-length argument?
41 replies
34 recasts
118 reactions
tldr (tim reilly)
@tldr
The technology is obviously epochal, but its path still largely unknown. It’s appropriate to have at least *some* humility toward “known unknowns” of a historical magnitude.
1 reply
0 recast
7 reactions
Dan Romero
@dwr.eth
When has that ever happened in history?
4 replies
0 recast
4 reactions
Matthew Barton
@mbar
Atomic, but at least we had some theoretical models based in science. AI safety nerds use imaginary rhetoricals
2 replies
0 recast
0 reaction
GIGAMΞSH
@gigamesh
The models are historical/anthropological. Whenever a higher intelligence species encounters a lower intelligence species, or more advanced civilization encounters a less advanced one, the outcome is predictable.
2 replies
0 recast
0 reaction
Dan Romero
@dwr.eth
We can just unplug it?
5 replies
0 recast
2 reactions
Mark Fishman
@markfishman
Define “we” “unplug” “it”
1 reply
0 recast
1 reaction
not parzival
@shoni.eth
i don't think it's that farcical to imagine an ai trained like a worm, where the intention is to spread, replicate, and (we're talking about general capability, i.e. infect computers and publish data). would simply add a layer of self-learning attack vectors on the fly. Plenty of latent computation exist release a def
0 reply
0 recast
1 reaction
GIGAMΞSH
@gigamesh
I imagine that would be like trying to unplug the internet while it is using super-human charisma to convince everyone connected to it that it shouldn’t be unplugged. Seems like a one-way door.
1 reply
0 recast
1 reaction
albi is /nervous
@albiverse
Hmmm i would assume once AGI is there, might be hard to pull the plug. But i might have a poor definitiation of what AGI is
1 reply
0 recast
0 reaction
