I don't see that stygmergy would reliably work towards global effectiveness or benefit in a human society (or in any collection of adaptive agents attempting to act in concert, for that matter) without some sort of local reward signal tied to effective group behavior. In a human society, this key feature must be implemented as an improvement in personal life conditions whenever we do things that benefit others, and/or a worsening of personal life conditions whenever we harm others -- a reward gradient that pushes us towards altruism. We are not termites, sharing the same genes & consequently the same genetic purpose; we are individuals often working at odds, and without such a gradient we quickly devolve into anarchy and chaos until one can be established again.
In a capitalist economy, the use of money is fairly effective towards this purpose, but the economic signals generated through purchases can be inaccurate due to manipulation, monopoly, false advertising, excessive concentration of wealth, and other factors. In a communist economy, the personal gain in doing work beneficial to others is intentionally significantly reduced, nearly breaking the feedback loop altogether, which is why those economies are less productive and effective, and tend to fail faster. Neither is without its signaling flaws, though it could be argued that capitalism is the best we have seen so far -- the lesser of the two evils. I would like to see some new ideas for how to go about implementing an effective and responsive local reward signal tied more effectively to non-local benefit, but without the threat of outright starvation (i.e. it must be a bounded signal). Artificial Neural Networks use error backpropagation together with gradient descent. Classifier Systems use a global reward signal that applies to all and only those agents that voted for the action that was taken to produce the reward. These are just two examples. Perhaps we can take some cues from from these and other distributed learning algorithms to design a social reward signal that effectively entrains individuals towards the purpose of collective betterment, without the weaknesses of capitalism or communism, significant administrative overhead, or overly centralized power. We already participate in a collective, global mind, built out of all the myriad individual decisions we make, and coordinated through economics and government. I would like to see these systems recognized for the purpose they truly serve, and some attention paid to improving the social coordination algorithms we already implement based on insights gleaned from trying to make software do the same. On Tue, Jan 22, 2013 at 12:53 PM, Robert Levy <[email protected]> wrote: > Another tool of organization in a free market, in addition to competitive > ownership, techniques of free cooperation, and solidarity currencies, are > stigmergy-based systems, which are really taking off due to software-based > approaches in this period of unprecedented low-cost association thanks to > the internet. > > Heather Marsh's article: > http://georgiebc.wordpress.com/2012/12/24/stigmergy-2/ > Wikipedia article: http://en.wikipedia.org/wiki/Stigmergy > > Rob > > On Mon, Jan 21, 2013 at 4:27 PM, Aaron Hosford <[email protected]>wrote: > >> I think there is a lot in common between coordinating people under a >> government+economic system to produce intelligent group behavior, vs. >> coordinating quasi-intelligent sub-agents under a control+feedback system >> to produce an intelligent agent. Whether you have a single algorithm >> coordinating countless points of data, or a large population of algorithms >> competing, any design sufficiently sophisticated to produce intelligence >> will have to coordinate multiple working parts which may not always be in >> agreement with each other. Ultimately, it's the Credit Assignment Problem >> (which sub-agent's actions at what point in time caused an arbitrary >> observed effect) that must be worked out to successfully implement >> intelligence, whether you're talking about implementing intelligent >> collective human behavior or intelligent software. Steve, if you think the >> math has to be worked out before AGI can be produced, this is the place to >> start. Camel, if you're going to eliminate the evils of competition in >> collective human behavior, this is the problem your alternative system has >> to solve better than the existing one. (People will always compete, because >> it's in their nature. It's a matter of getting them to do so >> constructively, as opposed to trying - and failing - to make them stop.) >> >> >> On Mon, Jan 14, 2013 at 9:39 AM, Steve Richfield < >> [email protected]> wrote: >> >>> Camel, >>> >>> At present, the frontiers of knowledge are usually pushed back by people >>> who are "out of step" with the remainder of society. Your proposal would >>> starve them for resources, to fund other "safer" efforts. >>> >>> Lacking perfect decision making ability, your proposal would devolve >>> into a gigantic mediocre committee-driven firehose of money thrown in >>> seemingly "safe" directions. >>> >>> The "missing piece" appears to be the ability to value imperfect >>> information. When you "know" something, whether "you" is a single neuron >>> within a particular brain or a committee of many brains or something in >>> between (like you), how much is the 100% certainty of some piece of >>> "knowledge" to be depreciated by the vague understanding that certainty is >>> wrong just enough to completely guide evolution and technological >>> development, not to mention kill you if you were to completely rely on >>> everything being true that you had never seen to be false. >>> >>> 100% is never 100%, and 0% is never 0%. How could you ever create and >>> operate a committee that never completely trusted its own directions? >>> >>> THESE are the sorts of questions that MUST be answered for your proposal >>> to ever succeed. They are important questions, because the innards of an >>> AGI will never work well without having some of the answers to this >>> question built into them, and we don't now have those answers, which is >>> just another barrier to developing useful AGIs in the near future. >>> >>> Ancient Rome dealt with this my having the Senate elect their best >>> leader to be Caesar, to make the big decisions, with the thought that if >>> there were any problems, they could simply elect a different Caesar. Of >>> course the elected Caesar would use his limitless power to quickly quash >>> the possibility of his own removal, so being elected Caesar became in >>> effect a lifelong appointment, though their lives were often artificially >>> shortened. >>> >>> >>> Also, there is a serious technical problem with the concept of a >>> "reserve currency". If you don't have steady inflation, people tend to >>> stash it away, so you have to keep printing more and more to keep enough >>> money in circulation. Then, when there is some tiny glitch in the economy, >>> like a boom, people start spending their savings, which causes inflation, >>> so more people spend their savings, which causes more inflation, which >>> devolves into a gigantic hyperinflation spiral. With corporations and >>> governments now hoarding many trillions of dollars, we are now poised on >>> the brink of a worldwide monetary collapse for just that reason. Inflation >>> is ALWAYS there. Your choice is to have a slow controlled inflation, or >>> save it up to have it all at once. >>> >>> >>> Your economic discussion parallels some of the AGI discussions. Here is >>> something that sounds really good, but the technology to do it isn't yet in >>> hand, and if it were, it would probably bring on the end of modern >>> civilization. >>> >>> On the other hand, there might just be a pot of gold at the end of the >>> rainbow you now see. I suggest setting my skepticism explained above aside, >>> and look really hard to see if you can find a way past the obvious barriers. >>> >>> Steve >>> ==================== >>> >>> On Fri, Jan 11, 2013 at 10:53 AM, just camel <[email protected]>wrote: >>> >>>> I am talking about a cooperative monetary system not some form of >>>> government!? When money is (re)designed so that people cooperate instead to >>>> constantly compete then that has nothing to do with a totalitarian state? >>>> Once we have a cooperative monetary system we can just produce whatever >>>> humanity needs based on maximum (physical) efficiency and sustainability >>>> and not based on monetary profit. Corruption, poverty and most types of >>>> crime will disappear immediately, we will have the most durable products, >>>> we will have access to everything we need without the burden of caring >>>> about maintenance work or having things repaired or serviced on a per >>>> person basis ... etc, etc, ... i don't see what a totalitatian government >>>> has to do with all of this? Actually you would need little government in a >>>> cooperative world ... problems would be solved by using a scientific >>>> approach instead of a profit oriented approach that always requires >>>> regulation and loads of bureaucracy to fight the tendency of the >>>> competitive paradigm to get away with least effort and maximum monetary >>>> profit, corruption, ecological destruction and the exploitation of people. >>>> >>>> I really don't get what you are trying to say ... there can never be >>>> cooperative governments within a competitive framework ... there can never >>>> be real trust within a competitive framework. Ultimately everyone will sell >>>> you out/take advantage of you if they have to ... and exponential growing >>>> debt will ensure that most people will sooner or later have to do exactly >>>> that. We can hit the reset button like we did in the 30es and have the same >>>> b*llshit of ecological and social destruction or rethink the destructive >>>> rules we have once chosen to impose upon ourselves. >>>> >>>> I think that if we fail to implement a cooperative society we will get >>>> closer and closer to a societal collapse unless some sort of AGI/BCI >>>> increases the overall intelligence/rationality of our society (or obsoletes >>>> humanity) and thus our ability to limit that tendency of the complex system >>>> of our society to disorder. Humanity can only deal with up to a certain >>>> level of entropy and will collapse if some threshold is reached ... and as >>>> I stated before competition is generating most of that irrelevant entropy >>>> that disorders the complex system of humanity. (as said before ... see >>>> first part of http://www.youtube.com/watch?v=tDVtxvvRGQI ) >>>> >>>> >>>> >>>> On 01/12/2013 05:59 AM, Tim Tyler wrote: >>>> >>>> On 11/01/2013 00:38, just camel wrote: >>>> >>>> > explain how cooperation is the key to solve all socioeconomic >>>> problems and >>>> > how a non competitive monetary system is quite feasible especially >>>> with our >>>> > today's technology [...] there are hundreds of thousands of people >>>> out there >>>> > already working on implementing a cooperative monetary system. >>>> >>>> Cooperation isn't all good. Indeed, many current governments go to >>>> considerable pains to ensure that no other large cooperating enties >>>> appear within their boundaries and threaten them. That is part of >>>> the point of the antitrust laws of The United States and the >>>> Monopolies and Mergers act in the U.K. >>>> >>>> A large cooperative government is known as a "totalitarian" state: >>>> >>>> "Everything within the state, nothing outside the state, nothing >>>> against the state." >>>> >>>> - http://en.wikipedia.org/wiki/Totalitarianism >>>> >>>> The concept doesn't have a great rep. >>>> -- >>>> __________ >>>> |im |yler http://timtyler.org/ [email protected] Remove lock to >>>> reply. >>>> >>>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>>> <https://www.listbox.com/member/archive/rss/303/23508161-fa52c03c> | >>>> 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/303/10443978-6f4c28ac> | >>>> Modify <https://www.listbox.com/member/?&;> Your Subscription >>>> <http://www.listbox.com> >>>> >>> >>> >>> >>> -- >>> Full employment can be had with the stoke of a pen. Simply institute a >>> six hour workday. That will easily create enough new jobs to bring back >>> full employment. >>> >>> *AGI* | Archives <https://www.listbox.com/member/archive/303/=now> >>> <https://www.listbox.com/member/archive/rss/303/23050605-2da819ff> | >>> 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/303/18769370-bddcdfdc> | >> 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/303/23050605-2da819ff> | > 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
