Re:OK, I can partly explain the LLM chess weirdness now
by dartos on Thursday November 21, 2024 @04:00PM
(#42208602)
There isn’t much utility, but tbf the outputs aren’t identical.
One danger is the human assumption that, since something appears to have that capability in some settings, it will have that capability in all settings.
Thats a recipe for exploding bias, as we’ve seen with classic statistical crime detection systems.
One danger is the human assumption that, since something appears to have that capability in some settings, it will have that capability in all settings.
Thats a recipe for exploding bias, as we’ve seen with classic statistical crime detection systems.