On 12/18/05, Eugen Leitl <[EMAIL PROTECTED]> wrote: > On Sun, Dec 18, 2005 at 03:36:59PM -0500, Pei Wang wrote: > > > I'm afraid the issue is not as simple as you belief. Your argument is > > based on the theory that to get what we call "intelligence", a > > necessary condition is to get a computer with the computational power > > of the human brain. > > There are two issues. Bootstrap, and operation complexity. > With current lack of knowledge on constraints on operation > of biological infoprocessing system the onus is on bootstrap > (plowing through infertile parameter space). Another issue > is how much your substrate is helping you. Some operations are > intrinsically cheaper in some specific substrates. So "computational power" > is not well-defined. This especially applies to the structured > substrate between our ears. The computational power metric > on that substrate is not well-defined, in absence of a very > specific metric. > > Another issue entirely is how much crunch (however you might define > crunch) is required for intelligent action, and how much crunch you > need to obtain the specific state required to perform said intelligent > action. > > I would, personally, consider throwing hardware on bootstrap > and obtaining constraints from live critters > easier than just magically extracting required knowledge from > much contemplation of mathematical navel fluff, and introspection.
This is a much better argument, though to me, your understanding of "intelligence" is still too close to "human intelligence as produced by the human brain". It is a valid research goal, though not the one I'm interested in. > It is a self-evident fact that we're intelligent, and that a certain > process brought us here. Lack of progress in respect to AGI on part > of the mainstream is similiarly evident. In absence of practical progress > the onus of proof is on the side of those with extraordinary claims. Agree with all the above statements. However, it still doesn't mean that the approach you suggested is the only possibility. > I think people with a bad track record shouldn't risk a lip, > unless they have a killer demo. Especially, since everybody's > funding is at stake. > > Do you have a killer demo? I don't know what is your standard for "bad track record", and I don't have a killer demo yet --- I don't think anyone has. But if you think only the people who have built an AGI can talk about it, why should we have this discussion? Pei ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
