|
| Home | The Boards | Messages | Thread |
|
|||||||
| by: ColonelZen in reply to 15 by ColonelZen | IP: 251.46 | rated: 0-0 | posted: 2007-11-30 13:35:09 | ||||
|
(Ron followed up with a $10 bet that it will be at least a century before we develop a Turing Test capable AI) Of course I'll take that bet. And I'll tell you that if I lose I'd be delighted to be around to pay off I'd be PO'd ... I suspect that in a hundred years I could get close to programming an AI .. unfortunately I likely only have 20-40 in proper person with declining faculties and for the next 15-20 will be spending too much time working for the man doing programming a shade less interesting. But there are lots of BYM out there looking. I once said rather cynically that I expect AI to come from the gamers, and probably serendipitously, rather than academia. For that matter has anyone (as per MS, IBM, Oracle, Intel ...) ever picked one of the leading lights of AI given say a 100M grant and said "pick a team and go get it done!"? Do that with a half dozen or so teams, and I'd bet that in 10 years it would be done ... and the ROI would be huge (and for IBM and Intel and maybe Oracle, the financial gain would come from hardware and service contracts on expanded facilities even if the result were completely open sourced). Marvin Minsky's "The Emotion Machine" is almost a blueprint at the current stage of our ignorance (I must reread, there's too much there that I didn't fully absorb the first time through). The NL problem probably overlaps the general AI problem almost completely. English certainly requires semantic cues and reflexive references to properly parse. I have no doubt that every other natural language is the same even without counting non-grammatical idioms. There are issues for AI not directly related - general learning and knowledge representation, but the NL processor needs thes e to some extent as well (it has to remember, i. e. learn, the reference of pronouns, and needs a knowledge store that says "Sally" is likely a she). -- TWZ |
|||||||