Maybe someone has some information regarding the time series evolution from "Seed AI" to "Trans-Human AI"
I was trying to get some idea optimistic versus pessimistic time series per some MIPS rate to get some idea of the learning curve versus time. Dan Goe ---------------------------------------------------- >From : Charles D Hixson <[EMAIL PROTECTED]> To : agi@v2.listbox.com Subject : Re: [agi] Timing of Human-Level AGI [was: Joint Stewardship of Earth] Date : Sat, 06 May 2006 12:21:55 -0700 > Ben Goertzel wrote: > > Hmmm.... > > > > The inimitable Mentifex wrote: > >> http://www.blogcharm.com/Singularity/25603/Timetable.html > >> 2006 -- True AI > >> 2007 -- AI Landrush > >> 2009 -- Human-Level AI > >> 2011 -- Cybernetic Economy > >> 2012 -- Superintelligent AI > >> 2012 -- Joint Stewardship of Earth > >> 2012 -- Technological Singularity > > > > > > Regarding the 2009 schedule for human-level AGI, based on what I see > > in the field and in my own project, it seems somewhat wishful > > thinking, though certainly not impossible. > > > > I believe the Novamente design is basically a solid design for AGI, > > but there are still plenty of details to be worked out --- and I > > believe that even if we got a lot of funding right now, achieving > > human-adult-level AGI by 2009 would be about the most optimistic > > scenario that is at all plausible. > > > > Unfortunately though, I have to say that we'd have to get lucky to get > > there by 2009, even with a lot of funding. > > > > Working out the details of a complex AI design just takes time -- and > > I suspect this will be true for any AI design, be it computer science > > based like Novamente, or neuroscience based like the designs that > > Hawkins, Guillen and others are dreaming of. > > > > Heck, look how long it's taken Microsoft to work out the details of > > their new operating system release ;-p ... a far simpler thing, though > > admittedly occupying far more lines of code... > > > > **Maybe** of course, there is someone out there with a much simpler > > AGI design that will take less long to get fully implemented and > > refined -- or maybe there is someone out there with an AGI project > > that's much further along... but I have seen no evidence of this... > > > > Next , regarding the 2012 thing, I have a strong feeling that 2012-ism > > is going to start seeming very silly sometime around, er, 2013. > > > > And this prediction of mine is intended as valid whether or not we > > have a Singularity before 2012 (which I very much doubt). > > > > Please let's restrict discussion on this list to topics more directly > > related to Artificial General Intelligence: no 2012-ism!! There are > > other lists for that, I'm sure.... > > > > I realize list traffic has been light lately -- and I've been too busy > > with various sorts of practical work (including AGI work, thankfully) > > to post a bunch of messages ... but I'd rather have light list traffic > > than off-topic list traffic of this nature!! > > > > thanks > > Ben > > > > ------- > > To unsubscribe, change your address, or temporarily deactivate your > > subscription, please go to > > http://v2.listbox.com/member/[EMAIL PROTECTED] > > > Not that I think 2009 is probable...but..... > To me it seems plausible that we are still missing fundamental > techniques in constructing an AI. It is plausible that once those > techniques are developed, the AI will self-assemble. This doesn't mean > that other approaches won't work, chemistry doesn't stop working because > viruses self-assemble. > > Of course, since I also don't really believe in a "general > intelligence", but rather in a toolkit of techniques to me it appears > that this will lead to lots of "small intelligences of specialized > nature", and it is these that would self-assemble into a larger "pseudo > agi". It wouldn't really be a general intelligence, but it might well > come closer to the mark than people do. Timeline? To me it appears > that ANYONE who talks about a timeline here, with an intent to be > specific, is misleading his audience. Truthfully, I wouldn't say it > couldn't happen tomorrow, but if it did the thought processes would be > very slow, since the only plausible scenario that I can see is a bot-net > that wakes up. Blue-folding (or whatever IBM's latest offering is > called) isn't going to be dedicated to "blue-sky" research for any > sensible fraction of it's time, and the same is true of the other > intentional supercomputers. (I may think that the estimates of the > computational needs for human scale intelligence are vastly overstated, > but they're still pretty large. [I think much to most of the > computation the brain does is related to handling the body. And I also > feel that a good "agi" design will modularize this code out of the agi, > and have it reside in body specific modules that can be swapped when > bodies are swapped. And that ideally the body specific code with run on > dedicated cpus...located in the body. This would facilitate > consciousness switching from body to body as needed.]) > > ------- > To unsubscribe, change your address, or temporarily deactivate your subscription, > please go to http://v2.listbox.com/member/[EMAIL PROTECTED] ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
