Shane, I have also considered using massively distributed processing a la [EMAIL PROTECTED] for Novamente; but in a Novamente context, this issue is not closely tied to "open-ness."
This is because we could use massively distributed processing for aspects of Novamente cognition, without releasing the vast bulk of the Novamente source. Our architecture would involve a central Novamente cluster, doing types of cognitive processing that are better done centralized (mostly probabilistic inference), and then a massively distributed periphery doing things that are better done massively distributed (mostly evolutionary learning). -- Ben G > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > Behalf Of Shane > Sent: Friday, March 05, 2004 9:04 AM > To: [EMAIL PROTECTED] > Subject: RE: [agi] Open AGI? > > > Hi Ben, > > I'm not really interested in "open source" in the software sense in > particular, but rather openness in general. Of course if you open > source the code then the project is very open in general too! > > I see that you run sort of an intermediate approach here, as does Pei. > Peter takes a more closed approach with A2I2, which probably reflects > his background in business rather than academia. Others like James > Rogers take a very closed approach; in fact I don't think I have ever > seen a document describing what he is working on? If there are more > closed projects out there.... well we probably wouldn't even hear > about them in that case ;-) > > I understand your desire to limit the number of people working on your > code out of purely practical reasons also. However with the emphasis > on genetic algorithms in vetta I have another reason to open up the > project --- I might actually need the participation of many people > donating their CPU time in order to obtain sufficient computer power. > > I figure that if the Search for ExtraTerrestrial Intelligence (SETI) > has managed to get 2 million years of CPU time then perhaps a few > people out there would also be interested in donating CPU time for a > Search for Artificial General Intelligence. Unfortunately SAGI is > a really bad sounding name! So vetta it remains. With millions of > years of computer time simulating billions of generations on a very > large population of learning networks perhaps some interesting > evolution could start to take place. However by taking this road > the project would be very open as I'd be literally distributing the > design of the system as it evolved to a large number of computers > all around the world on a regular basis. I'm not sure that this > much openness would be a good thing... this is why I got thinking > about this question. > > Cheers > Shane > > > > > > ___________________________________________________________ > Yahoo! Messenger - Communicate instantly..."Ping" > your friends today! Download Messenger Now > http://uk.messenger.yahoo.com/download/index.html > > ------- > To unsubscribe, change your address, or temporarily deactivate > your subscription, > please go to http://v2.listbox.com/member/[EMAIL PROTECTED] > ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
