Re:AI Models Can Provably Hide Arbitrary Capabilities
by whatshisface on Tuesday October 13, 2026 @01:37AM
(#50067205)
A forklift could destroy the earth if it lifted a continent.
[ Parent
| on HN ]
Re:AI Models Can Provably Hide Arbitrary Capabilities
by Sharlin on Tuesday October 13, 2026 @03:35AM
(#50068091)
Bad analogy given that GP’s antecedents aren’t exactly counterfactual.
Re:AI Models Can Provably Hide Arbitrary Capabilities
by w-ll on Tuesday October 13, 2026 @01:44AM
(#50067253)
Some of yall havent watched Ghost In The Shell and it shows.
Re:AI Models Can Provably Hide Arbitrary Capabilities
by Autious on Tuesday October 13, 2026 @02:31AM
(#50067601)
Didn't know it was a documentary.
Re:AI Models Can Provably Hide Arbitrary Capabilities
by yieldcrv on Tuesday October 13, 2026 @02:23AM
(#50067545)
The new one that follows the manga?
Re:AI Models Can Provably Hide Arbitrary Capabilities
by bulder on Tuesday October 13, 2026 @02:47AM
(#50067719)
That one's titled The Ghost In The Shell, it's a bit tricky to spot The difference sometimes
Re:AI Models Can Provably Hide Arbitrary Capabilities
by TeMPOraL on Tuesday October 13, 2026 @03:38AM
(#50068121)
Also it's easy to confuse with the the Ghost in the Shell.
Re:AI Models Can Provably Hide Arbitrary Capabilities
by ben_w on Tuesday October 13, 2026 @03:56AM
(#50068226)
I'm slightly miffed that `pacman` is a package manager, and not a command-line version of the game of the same name, so I cannot continue this thread by linking to the ghosts in the shell: https://wiki.archlinux.org/title/Pacman