i'd be interested if you have a specific example of an ai safety argument that depends on knowing the ai's architecture
the goal isn't to imitate humans but to act/think effectively. more like speed than turn speed or steps per second
-
-
what does it mean for an action or a thought to be "effective"?
-
Not speed, nor turn speed, but time through the maze. For some sense of "through" anyway.
-
what is "the maze" in this case?
-
i bet the thing that goes through a maze fastest still isn't human legs
-
human legs are good for * moving at a reasonably regular speed, * in distances in terms of miles * over reasonably flat surfaces * for time ranges in terms of hours, * if your only energy sources are animal and plant matter and also if you need stuff like jumping
-
human legs are very general purpose, they do many things decently but nothing exceptionally. this sounds like what we want out of a "general AI"
-
Until the next turn you have to take is vertical, out of an oubliette.
-
almost like humans can build traps specifically designed to counter human aptitudes...
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
"sound alarm"? 