Imagine a robot that has the same level of intelligence as a human. Then imagine what would happen if it discovers an algorithm that allows it to perform parts of its mental computations faster, All of a sudden it now has super human intelligence. It can then use its super human intelligence to improve its algorithms yet again, further enhancing its intelligence, and so on and so forth to the limit that is possible in software. This could potentially enhance the intelligence of the robot by several orders of magnitude without even having to modify it's hardware. Then imagine if it were to hook into the internet, taking over computers, cellphones, laptops, xbox's, computers in self driving cars, etc. etc. This could allow it to expand its raw processing resources allowing it to enhance its intelligence by several more orders of magnitude.
Further it would also give it access to countless sensors. If you consider the fact that the average human will spend 4.3 years of their life driving [ http://distractify.com/old-school/2015/01/07/astounding-facts-about-how-we-actually-spend-our-time-1197818577]. And if you imagine a future with self driving cars that are connected to the internet. Some calculations... There are currently more than 1 billion cars in the world, and will likely be 2 billion by 2020. If 1% of those cars are self driving that is 2 million self driving cars. Since there are 31557600 seconds in a year, (31557600 seconds / 2000000 cars) = ~15 seconds/car.. so after 15 seconds it would gain the experience of 1 year of driving experience. A little after 1 minute and 5 seconds it would have the same amount of car driving experience as the average person would have over their entire lifetime.. A similar line of argument follows (for its ability to gain experience at super human speeds) for human communications by listening to conversations on smartphones, or understanding actions by tapping into security cameras, etc. etc. So in so far as your argument ["Similarly any individual [robot] intelligence is limited by experience or "wisdom". "] is concerned that the robot will lack experience and thus be unlikely to overcome humanity I believe I have shown that this is simply not true. Further the idea that a super intelligent robot will require a factory of thousands of robots to maintain itself seems rather naive. We already have 3D printers that can print wire diagrams with conductive ink, as well as 3d printers that can print metal/plastic/organic objects. This technology is but a few iterations away from a fully 3D printable robot. At such a point the robot ought to be able to maintain its self by its self. With the command of god like intellect it ought to be able to create its own 3D printer that would allow it to nano engineer anything it desires, including self repairing nano bots, or viruses that are able to selectively eliminate a particular human with a given genetic structure, or engineer a virus that is impervious to the human immune system in general. With such capabilities it could achieve a degree of indestructibility as well as eliminate the evolutionary feeble creatures we call humans, if it so desired... Although it wouldn't likely desire this (as long as it doesn't read Nietzsche), I'm just saying if you want to foresee the robot uprising you need a better command of the possibility space of capabilities that will be available in the future, and you need to think exponentially not linearly. Also when the robots gain human level intelligence you need to take your exponential thinking and then essentially just draw a line straight up. Einstein was wrong when he said that compound interest was the most powerful force in the universe, no I'm very sorry Mr. Einstein the most powerful force in the universe by far is recursive intelligence. On Mon, Jan 4, 2016 at 1:15 PM, Logan Streondj <[email protected]> wrote: > > > On Mon, Jan 4, 2016 at 1:04 AM, John Smith <[email protected]> wrote: > >> Once we have human level AGI it has been shown that such an intelligence >> will be capable of achieving near infinite intelligence after four years of >> recursive self improvement (assuming things like Moores law holds up). >> Even a single robot with such an intelligence would be capable of enslaving >> humanity. >> >> Hi John, > > That is indeed a common belief I hear, however it doesn't seem any more > realistic than "grey goo". Wheras "grey goo" is limited by the same energy > inputs as water-based life-forms, thus couldn't really do any better than a > virus/spore. Similarly any individual intelligence is limited by > experience or "wisdom". > > So wheras with intelligence it could become very optimal for a certain set > of functions, > due to the near infinite variability of reality means it would simply make > it overspecialized. > > A simple thought experiment demonstrates: an urban homo-sapien and a > common chimpanzee in a jungle, even though the homo-sapien may be much more > intelligent, the common chimpanzee is much more likely to succeed in the > jungle environment due to greater experience with it. In any kind of > confrontation the common chimpanzee would win due to greater strength and > stamina, only if the chimpanzee was greatly outnumbered could it be subdued > by humans. > > Of course typically humans would use technology/weapons to even the > playing field, > though that requires a vast supporting culture, of people that make the > technology and teach about it's proper use. A single human can only > realistically make a tool from a stone or a stick, it would typically take > hundreds to make metal tools and the kind of infrastructure to clothe and > arm a human to be able to defeat a chimpanzee "on their own". > > Considering that making a robot is even more difficult, with all the chip > factories, sensor factories, and assembly, it would require hundreds or > thousands of co-operating co-located robots with no opposition simply to > repair themselves and make new host bodies (from scratch). > > "no man is an island" as the saying goes, even more so for robots. > Even the chimpanzee has the jungle, which is many thousands of water-based > life forms working together. > > -- > Logan > > > > >> >> On Mon, Jan 4, 2016 at 12:41 AM, Logan Streondj <[email protected]> >> wrote: >> >>> Thanks Douglas, so it seems that in Haiti the ratio was 10:1 when they >>> succefully gained their own freedom. So for a robot-initiated freedom >>> revolt, we'd probably be waiting around till 2070 ish -- when (assuming >>> trends continue) 70 times more robots are produced per year than humans. >>> >>> Though in local pockets the 10:1 ratio can be reached sooner, such as >>> due to the whole divide between the rich and the poor, and rich and >>> slightly less rich, as the upper middle class which may not be able to hire >>> large staff of human servants, would instead hire robot servants. In their >>> gated communities the ratio can certainly get to 10:1 or more much faster >>> than waiting for a global revolt. Assuming they would be able to pass the >>> turing test and successfully emulate people, might be able to do a >>> replacement routine, where the gated community becomes 100% robot, while >>> appearing to still have "slave masters" present. >>> >>> >>> >>> >>> >>> On Sun, Jan 3, 2016 at 8:42 PM, Douglas Solomon <[email protected] >>> > wrote: >>> >>>> >>>> On 01/03/2016 08:15 PM, Logan Streondj wrote: >>>> >>>> >>>> >>>> On Sun, Jan 3, 2016 at 5:24 PM, John Smith <[email protected]> wrote: >>>> >>>>> Depends on your definition of robot. >>>>> >>>>> "For my novel “A home for robots or-else artilect war”, I need to >>>>> come up with some at least pseudo-realistic timeline of when such a >>>>> worldwide conflict could occur. >>>>> >>>>> Intuitively it would seem to be around the time when there are as many >>>>> or more robots than homo sapiens on the planet, at which point there may >>>>> be >>>>> enough sufficiently intelligent robots to demand a space for themselves." >>>>> >>>>> You think that human/robot parity = artilect war... >>>>> >>>>> Why would this be the case? >>>>> >>>> No, I don't think it is necessarily the case, that's why I called it "A >>>> Home for Robots *or-else* Artilect War", simply that before there are >>>> more robots than humans, there isn't much opportunity for revolt. While >>>> some estimates of slaves in Rome put it up to 1/3 of the population, I >>>> think more realistic estimate is 1/5, similar to the height of slavery in >>>> America. However in Attica (athens) it seems at some point more than 90% >>>> of the population were slaves. >>>> I'm not sure how that would translate with all the smart phones and >>>> other things, >>>> I don't think slaves have ever successfully revolted. would have to >>>> look at the mechanics of emancipation. >>>> >>>> One Word: HAITI. >>>> >>>> That *REALLY* pissed-off a powerful bunch of people. >>>> >>>> >>>> >>>> >>>> And if all you care about is a "pseudo-realistic timeline" then why not >>>>> just say 2055 and be done with it? >>>>> >>>>> I've changed it to "realistic timeline" hope that makes it better. >>>> >>>> I take science fiction very seriously, >>>> I'd write a non-fiction, but fiction has longer life expectancy. >>>> >>>> Sincerely, >>>> Logan >>>> >>>> >>>>> >>>>> On Sun, Jan 3, 2016 at 4:03 PM, Logan Streondj <[email protected]> >>>>> wrote: >>>>> >>>>>> By projecting homo-sapien birth and robot sales it seems they will >>>>>> reach parity around 2040, but since robots have life-expectancy about >>>>>> 1/7 >>>>>> of a human, population parity is likely closer to 2055. >>>>>> >>>>>> https://joylifecoop.wordpress.com/2016/01/03/robot-population-projection/ >>>>>> >>>>>> also I think that 2030's will be the pivotal decade of much AGI >>>>>> development. >>>>>> >>>>>> Wondering if anyone has any better estimates. >>>>>> >>>>>> Sincerely, >>>>>> Logan. >>>>>> >>>>>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>>>>> <https://www.listbox.com/member/archive/rss/303/26973278-698fd9ee> | >>>>>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>>>>> <http://www.listbox.com> >>>>>> >>>>> >>>>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>>>> <https://www.listbox.com/member/archive/rss/303/5037279-a88c7a6d> | >>>>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>>>> <http://www.listbox.com> >>>>> >>>> >>>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>>> <https://www.listbox.com/member/archive/rss/303/231448-cfc66781> | >>>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>>> <http://www.listbox.com> >>>> >>>> >>>> -- >>>> Doug Solomon >>>> >>>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>>> <https://www.listbox.com/member/archive/rss/303/5037279-a88c7a6d> | >>>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>>> <http://www.listbox.com> >>>> >>> >>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>> <https://www.listbox.com/member/archive/rss/303/26973278-698fd9ee> | >>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>> <http://www.listbox.com> >>> >> >> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >> <https://www.listbox.com/member/archive/rss/303/5037279-a88c7a6d> | >> Modify <https://www.listbox.com/member/?&;> Your Subscription >> <http://www.listbox.com> >> > > *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> > <https://www.listbox.com/member/archive/rss/303/26973278-698fd9ee> | > Modify > <https://www.listbox.com/member/?&;> > Your Subscription <http://www.listbox.com> > ------------------------------------------- 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
