zboží
(prázdné)
-
Jako člověk
Russell, StuartNení skladem 299 KčDo košíku -
Human Compatible
Russell, StuartNení skladem 290 KčDo košíku -
Künstliche Intelligenz
Russell, StuartNení skladem 2 382 KčDo košíku -
Human Compatible
Russell, StuartNení skladem 531 KčDo košíku -
Adolf Hitler Frontsoldat im 1. Weltkrieg, 1 DVD-Video
Russell, StuartNení skladem 284 KčDo košíku -
Sztuczna inteligencja. Nowe spojrzenie T.1 w.4
Russell, StuartNení skladem 591 KčDo košíku -
Sztuczna inteligencja. Nowe spojrzenie T.2 w.4
Russell, StuartNení skladem 456 KčDo košíku -
Artificial Intelligence
Russell, StuartNení skladem 6 306 KčDo košíku -
More Heroes, Villains & Victims of Hull and the East Riding
Russell, StuartNení sklademDo košíku
Recenze knih autora
RNDR.Mgr.Přemysl Stejskal napsal/a o Human Compatible
This book actually covers various topics from the human psychology, social anthropology, cognitive sciences, game theory and logic. I must say that I personally would recommend reading it rather than listening to it since that way you would probably be able to get much more of the content.
In any way, I appreciated the most the author's effort to actually offer some proposal... číst celé as to how to solve the AI control problem. There is not much of a question whether AI will ever become 'smarter' than the humans, but rather the 'when that happens, how can we, the humanity, still remain in control?'. Russell proposes developing a beneficial AI built on a particular understanding of human preferences. That way, without having exogenously inserted any specific goals into AI, it would be possible for it to update its knowledge at all times and, most importantly, accept that there may be situations in which it has to seek humans' approval of selected course of action, be overriden or even shut down.
In any way, I appreciated the most the author's effort to actually offer some proposal... číst celé as to how to solve the AI control problem. There is not much of a question whether AI will ever become 'smarter' than the humans, but rather the 'when that happens, how can we, the humanity, still remain in control?'. Russell proposes developing a beneficial AI built on a particular understanding of human preferences. That way, without having exogenously inserted any specific goals into AI, it would be possible for it to update its knowledge at all times and, most importantly, accept that there may be situations in which it has to seek humans' approval of selected course of action, be overriden or even shut down.
Je tato recenze užitečná?
0
25/10/2021