Forums before death by AOL, social media and spammers... "We can't have nice things"
|    alt.cyberpunk    |    Ohh just weirdo cyber/steampunk chat    |    2,235 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 350 of 2,235    |
|    joss wright to All    |
|    Re: AI (again)    |
|    28 Oct 03 17:37:27    |
      ekrodomos.net> 6557858e       From: joss@nospampleasewerebritish.nekrodomos.net              first off, i'm not that regular a poster (but have lurked for quite a       while), but i'd say that off-topic is quite a difficult thing to attain on       this forum. alternatively, you could argue that (given historical       evidence) anything that isn't "what is cyberpunk" is off-topic :o)              we seem to mainly agree on the AI front. but i would pick up on one point       that you made. a self-coding AI will still be limited, even though the       accidental limitations of its own code may well be reduced. the easiest       point to pick out is that any computational hardware on which we could       currently run an AI program would be a Turing machine. it is known that       there are limits to the computational powers of Turing machines, although       we do not have a more powerful model of computation. so an AI will, i       think, always be limited at some point. (arguably it could invent a more       powerful computational model and cause that to be created and upload       itself into it, but i think that is both unlikely and somewhat...       pathetically conjectural. it's like saying "it could be hit by lightening       and come alive".)              as to an AI being closer to an operating system, i think that that is a       fair and interesting point. the human brain is separated into distinct       processing areas for specific functions and each of these could arguably       be viewed as a sub-program running in a larger system. (my knowledge of       neuroscience goes no further than that, however, and so i'll stop making       wild theories).              oh, and i agree about the stating of the basis of the learning system       idea. it must receive, then process data and alter its behaviour based       upon this data or i fail to see how it can be said to learn. bear in mind       though that in a FPS, the "environmental data" is itself modelled on a       "real" universe, so in that situation the monsters are probably closer to       "seeing" and "hearing" than "perhaps" "you" "meant" "to" "put" "across".       damn... quotation marks are running away from me again. suffice it to say       that any inputted data is a form of sense. obviously       sight/taste/smell/hearing/touch are not the only possible input methods.              and as to all these games mentions: as someone who plays chess, go, and       every board game that i can get my hands on (atmosfear, mysteries of old       peking, the crystal maze board game, scrabble), i can only sigh in       satisfaction. :o)              also, allow me to apologize for spending an entire post violently agreeing       with you. i promise that next post i'll find something totally innocuous       to be offended at and start a flame war ;o)              joss              --       "A theory however elegant and economical must be rejected or revised       if it is untrue; likewise laws and institutions no matter how       efficient and well-arranged must be reformed or abolished if they       are unjust" - Rawls, "A Theory of Justice"              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca