Ben: In my work in bioinformatics and finance, for instance, I have no trouble assessing empirical evidence for hypotheses. AGI is a different sort of endeavor, in my opinion, because the steps required to find evidence validating an hypothesis like "OpenCog can lead to human-level general intelligence" are very complex and lengthy to undertake
It’s strange that you can so split your mentality and think in terms of evidence in only one sphere. The key to AGI is to think in terms of generality/”take off” – a machine that if it can obey one command, can obviously ipso facto obey an infinite series of other diverse commands. Aim at that – the problem of AGI - and you’ll be able to test and produce evidence whether you’re succeeding or not. (But you haven’t aimed at the problem – you’ve aimed at producing an “AGI architecture” without worrying what the take off problem is, or whether your architecture can solve it., and provide evidence for its utility) I can express that problem for you and the kind of solution entailed, quite simply., (I won’t try to make a serious, extended argument/case here). If you have a robot that can obey the command : “move your body over here” then you have an AGI, that will immediately be able to take off. For it will have the capacity to move its body over here/there in endless diverse ways - to “go [move its body over] to the kitchen” pace Woz, or to pick up endless diverse objects, or to tread on endless diverse terrains. An infant works on the principle of “Moving its body over there”. It may not immediately be able to “go to the kitchen”, but it self evidently has the capacity to develop that ability – to progress IOW from simpler to more complex journeys (from simpler to more complex “forms of moving its body over there”). Even thinking can be understood – and I’m confident, *is* understood – by the human mind in terms of “moving its body to a goal” , e.g. “use your head”, “ have a look at..[this problem]” To solve the problem of how a robot can understand and execute “move your/my body over there”, you have to solve the problem of how the conceptual system works, (and it’s got bugger all to do with the narrow AI database approaches that AGI-ers are currently using) - of how “move” and “body” like all concepts are open to endless diverse interpretations – endless diverse configurations and “movements” of the body and its limbs and faculties. ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
