home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!opl.com!hri.com!spool.mu.edu!sdd.hp.com!usc!news.service.uci.edu!unogate!stgprao
- From: stgprao@st.unocal.COM (Richard Ottolini)
- Newsgroups: comp.ai.philosophy
- Subject: Re: Minds, Computers and Searle
- Message-ID: <1993Jan27.230655.10105@unocal.com>
- Date: 27 Jan 93 23:06:55 GMT
- References: <1993Jan22.181429.8759@psych.toronto.edu> <chrisk.728156188@fester> <1993Jan27.210202.22368@psych.toronto.edu>
- Sender: news@unocal.com (Unocal USENET News)
- Organization: Unocal Corporation
- Lines: 27
-
- In article <1993Jan27.210202.22368@psych.toronto.edu> christo@psych.toronto.edu (Christopher Green) writes:
- >In article <chrisk.728156188@fester> chrisk@fester.dell.com (Chris Kostanick - X6736) writes:
- >>
- >>I'm curious as to what you think is possible in AI, weak or strong.
- >>Let's assume that compute power will continue to grow for another
- >>couple/three hardware generations. This should make 100 Mip machines
- >>readily accessable. What do you think an AI program running on
- >>one (with suitable robot effectors and vision aparatus) could do?
- >>
- >I'm not sure why sheer computational power would be relevant to the question.
- >Weak AI can do for cognitive science exactly what computational models do
- >for other sciences: formalize and idealize parameters thought to be
- >crucial, and show you in pretty short order what kinds of behaviors
- >such a model implies. Compare to data, revise model, and off to the races
- >we go.
-
- Much of the seminal work on AI was done in the late 60s and 70s on machines that
- KOP speeds (thousands of operations per second). There was a hypothesis that some
- of AI's problems could be solved by building a "fast" computer. An industry was
- created and died in the 1980s to build fast LISP (later PROLOG) machines.
- (This industry died because general purpose RISC micros could run LISP faster than
- special purpose machines.) The 10,000 - 100,000 fold increase in computing
- speed in the past twenty some years did not accelerate progress of AI
- in many people's opinion. Japan's 5th generation project was the most
- ballyhoed disappointed along this approach. I personally am hopeful in
- regards to strong AI, but faster computers were not the answer. Some more
- fundamental understanding of the nature of intelligence is the likely answer.
-