Ben, On 12/12/08, Ben Goertzel <b...@goertzel.org> wrote: > > >> > There isn't much that an MIMD machine can do better than a > similar-sized > >> > SIMD machine. > >> > >> Hey, that's just not true. > >> > >> There are loads of math theorems disproving this assertion... > > > > > > Oops, I left out the presumed adjective "real-world". Of course there are > > countless diophantine equations and other math trivia that aren't > > vectorizable. > > > > However, anything resembling a brain in that the process can be done by > > billions of slow components must by its very nature vectorizable. Hence, > in > > the domain of our discussions, I think my statement still holds > > I'm not so sure, but for me to explore this area would require a lot > of time and I don't > feel like allocating it right now...
No need, so long as 1. You see some possible future path to vectorizability, and 2. My or similar vector processor chips aren't a reality yet. I'm also not so sure our current models of brain mechanisms or > dynamics are anywhere near > accurate, but that's another issue... I finally cracked the"theory of everything in cognition puzzle" discussed here ~4 months ago, which comes with an understanding of the super-fast learning observed in biological systems, e.g. visual systems the tune themselves up in the first few seconds after an animal's eyes open for the first time. I am now trying to translate it from "Steveze" to readable English which hopefully should be done in a week or so. Also, insofar as possible, I am translating all formulas into grammatically correct English statements, for the mathematically challenged readers. Unless I missed something really BIG, it will change everything from AGI to NN to ???. Most especially, AGI is largely predicated on the INability to perform such fast learning, which is where experts enter the picture. With this theory, modifying present AGI approaches to learn fast shouldn't be all that difficult. After any off-line volunteers have first had their crack, I'll post it here for everyone to beat it up. Do I hear any volunteers out there in Cyberspace who want to help "hold my feet to the fire" off-line regarding those pesky little details that so often derail grand theories? >> Indeed, AGI and physics simulation may be two of the app areas that have > >> the easiest times making use of these 80-core chips... > > > > > > I don't think Intel is even looking at these. They are targeting embedded > > applications. > > Well, my bet is that a main app of multicore chips is ultimately gonna > be gaming ... > and gaming will certainly make use of fancy physics simulation ... Present gaming video chips have special processors that are designed to perform the 3D to 2D transformations needed for gaming, and for maintaining 3D models. It is hard (though not impossible) compete with custom hardware that has been refined for a particular application. Also, it would seem to be a terrible waste of tens of terraflops just to operate a video game. and > I'm betting it will > also make use of early-stage AGI... There is already some of that creeping into some games, including actors who perform complex jobs in changing virtual envrionments. Steve Richfield ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=123753653-47f84b Powered by Listbox: http://www.listbox.com
