True AI When?

kojikun · Post by **kojikun** » 2003-05-20 11:49pm

the guys who published the study are about as intelligent as the chimps they want to make into H. Trgologytus or whatever the fuck they want to call them. Humans and chimps share a common ancestor before the Homo genus came into existance, and while that line between Homo and Australopithacus is blurry, its not blurry to the extent that the edges include 7-8mya!

What constitutes the genus Homo is the descendants of the common ancestors of H. habilis, H. erectus, H. [sapiens]neanderthalensus, and H. Sapiens Sapiens; atleast until the differences become to great that species and sub species no longer work as good taxonomic categorizations. With the Pan genus, the differences are great enough to warrant a completely different genus.

I think a good way of determining whether they should be in the same species/genus/etc is if the nearest species-common-ancestor (meaning the common ancestor of H. sapiens sapiens and of H. [sapiens] neanderthalensus) can interbreed with the respective common ancestors of chimps and their nearest sister species.

Humans and Neanderthals might have been able to interbreed, with luck, but most likely couldnt, so we're seperate species. If our common ancestor and the latest respective common ancestor of the Pan species' in question could have interbred, then theyre the same species/genus. This is not the case, as the latest common ancestor of any two given chimpanzee species, and the latest common ancestor of any two modern human species could not have interbred. Thus, the species' in question are of a seperate genus.

NapoleonGH · Post by **NapoleonGH** » 2003-05-21 12:53am

well we are closer to chimps than Donkeys are to Horses, and donkeys and horses can interbreed and produce ofspring that while not viable (fertile) are still able to live a decently long time.

now as far as i know we cant interbreed with chimps and bonobos, but we share greater genetic similarity to them than two species who could interbreed.

Im pretty sure the point of what they were saying is that the Pan and Homo genuses should be merged into a larger genus considering the smiliarities.

Also perhaps, we should base taxonomy on genetic similarity rather than on the order of evolution/closest common ancestor

Admiral Valdemar · Post by **Admiral Valdemar** » 2003-05-21 08:36am

Just a pointer, new evidence suggests that we never were one with neanderthals and never bred with them, that or bred so little that none of their genome came into ours.

kojikun · Post by **kojikun** » 2003-05-21 05:46pm

we and neanderthals do have a common ancestor tho, valdy. and can someone PLEASE split this thread?

kojikun · Post by **kojikun** » 2003-05-21 05:49pm

ON TOPIC

What do you think the best route to AI/AS is; pre-designed neural networks/simulations, naturally organising neural nets/sims, or more computerised stuff like actual programs that are sentient not that produce virtual-brains that are sentient?

Cthulhu-chan · Post by **Cthulhu-chan** » 2003-05-21 07:16pm

I'd say self-organizing is the way to go. Keep all the basic functions simple and flexible so that they can be readilly adapted. The top down approach is a dead end, for now. We just don't have enough knowledge of how complex neural networks opperate to try emulating them pre-programmed.

kojikun · Post by **kojikun** » 2003-05-21 07:36pm

Cthulhu-chan, you know anything about programming, or might want to learn? We might be able to start up out own little SDN AI study group :p