|
I'm gonna say decades or sooner. The 'go' playing AI was a general intelligence and not really programmed with much structure. It viewed a series of go games, played against itself a bunch against itself, then smurfed a bunch of humans. Recursive learning, although narrow atm is already here. I also don't doubt that ai will be a thing. High resolution brain emulation could be a thing in a few decades, giving your ghost a path to immortality. It opens the possibility of a bunch of things, like having a philosophical discussion on combating entropy or you can have a reflection of yourself enjoying 24/7 IR sex orgy minigames. Idk go read the control problem if you want these ideas fleshed out better with more effort.
|
# ¿ Nov 28, 2016 23:14 |
|
|
# ¿ May 15, 2024 13:14 |
|
Condiv posted:the shockingly beatable intelligence of the average human is actually beyond our most advanced computers at the moment and we're not even close to being able to simulate such intelligence. also, the AIs that are "creating" art really aren't, they are simulating "art" based on their creators' notions and preconceptions about what art is and what is worthwhile art. UR RIGHT every human that picks up an instrument/brush was doing so without a prior influence or cognitive sample that they were emulating from
|
# ¿ Nov 28, 2016 23:16 |
|
Rush Limbo posted:Its also the case in, say, video games and other AI simulations that they handle even fairly basic things like stairs incredibly badly, to the point where various workarounds are created including them just plain ignoring such changes in elevation from a purely AI standpoint and just moving them while having animation take over. Video game ai is a bad metric for measuring where we are regarding ai. Development costs for stairs ai constrains stair ai. Also making that hardware run as a side process on 400$ consumer platforms constraints it.
|
# ¿ Nov 29, 2016 00:01 |