On Sun, Jan 21, 2007 at 11:49:00AM -0500, Benjamin Goertzel wrote: > However, I don't think these brain-theory-simplifiers are right. I
Nobody in neuroscience camp is taking them seriously. For a very good reason. > doubt the brain's general intelligence is based on a few simple > tricks/mechanisms. My guess is that it is fundamentally based on a > complex, unhol-ily messy combination of very many mechanisms, and that > we **will** need pretty advanced brain-scanning technology to puzzle It's just money, and time. It's a straightforward roadmap. > it all out. Because of this, my guess is that even with vast amounts > of $$ and scientific wizardry, the brain-emulation-based approach will > NOT get there first. We're certainly going to see who's right within the next 30 years, or so. > >There are shorter paths, but nobody knows where they are. > > but I say: Don't assume everyone is as ignorant in this regard as you are > ;-) Instead of ad hominem arguments, how's for a little demonstration. If there are shorter paths, and you folks know what they are, and in fact people have been saying this since 1950, it can't be all that difficult to build a working prototype. So where is it? Where's at least a success track, which shows you're barking up the right tree? > > He also wrote: > > >That's the key point of it: the world is complicated. Dealing with the > >world takes lost of machinery. > > Yes, but not necessarily the same kind of machinery as the brain has. No disagreement, but nobody knows the precise shape of that machinery. The parameter space of candidates alone is vast you can get lost in it forever. > I don't think the bias toward oversimplification is any more present > in AI than in other scientific disciplines. And the brain-based No disagreement, but such disciplines have ever dealt with a complex system successfully. > approach to cognition has certainly suffered from at least as much > oversimplification as the computer-science-based approach, cf Hawkins > and Hecht-Nielsen. Strange, I would have put both of them into the usual strong AI camp just as everybody in here. > >No validation or further evidence required; it's all obvious. > > I don't know whom you're attacking here, but I have certainly never > claimed that achieving AGI via computer science rather than brain > emulation is **obvious**. If it weren't very hard someone would have > done it long ago. But "very hard" doesn't mean implausible if you are > very smart and work very hard... I agree that somebody is going to succeed, at some point. > Historically, all approaches to AGI have failed so far. The wet neuroscience people have hardly failed, in fact they're clocking a great rate of progress. Their simulations agree with what the tissue is doing -- in fact, it's the only way to know you're making progress. > I really don't think ego is the issue here. Maybe not, but watch how Peter snapped at me, quite uncharacteristically. There seem to be some pretty powerful emotions at play. > >but it's the only game in town, as far as I can see. > > It is not the only game in town, there are plenty of us taking the > computer-science approach to AGI as well. Perhaps I should have put it as "the only game in town with any winning chances". But, the future will show. > Your point of view seems even more extreme than Kurzweil's, as he e.g. > mentioned Novamente favorably in "The Singularity is Near." > > I find it frustrating that your attitude is at least approximately > shared by so many others, including Kurzweil and some of the high Hey, don't blame me for the AI winter. The folks brought it on themselves. > muck-a-mucks at IBM. However, I also find it gratifying that there Why do you think Markram & Co are doomed? > are many other smart folks such as Pei and Peter Voss and others on > this list who have a more open-minded attitude regarding various AGI > approaches. I've first started reading about AI as a 14 year old, in 1980 or so. It's the abysmal lack of progress that made me give up on them in beginning on the 1990s. -- Eugen* Leitl <a href="http://leitl.org";>leitl</a> http://leitl.org ______________________________________________________________ ICBM: 48.07100, 11.36820 http://www.ativel.com 8B29F6BE: 099D 78BA 2FD3 B014 B08A 7779 75B0 2443 8B29 F6BE ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
signature.asc
Description: Digital signature
