food for thought

22 responses to “food for thought”

1. 

   boring fact

   Oct 4, 2005 at 1:48 am

   someone is very artistic … opens all the right brain channels to receive some hefty lefty stuff :)Thanks for sharing with us

   Reply
2. 

   Todd Huffman

   Oct 4, 2005 at 3:40 am

   Two really essential components for effective problem solving with diverse crowds:

   1-A sufficiently diverse set of individuals to explore the solution space. In the cartoon above, this is covered, but they add the caveat the solvers must be smart. Smart solvers aren’t a definite need; intelligence can be compensated for though inexpensive agents. Take ants, not intelligent, just a basic rule set and some ‘random’ activity to add diversity. The less intelligent the agents, the more you need a…

   2-Good fitness function. The group of agents needs some method for coming to a consensus. For ants and other simple agents, the fitness function can be crude and still effective. For more amorphous activites, such high level planning and business strategy in a multinational corporation, the fitness function is a bit more difficult. Obvious functions, such as market worth or other financial indicators, are not necessarily useful, as the time lag is too long and the numbers too easy to manipulate. Many subtle intermediate fitness functions are needed. Often, the best fitness function is the good judgement of the agents, as they collectively compete and give quarter to the better ideas, long before any implementation. Which brings us back to the necessity for smart solvers…

   Reply
3. 

   LouisL

   Oct 4, 2005 at 3:44 am

   Reminds me of two more Gigerenzer books in the queue!

   Reply
4. 

   angry shock

   Oct 4, 2005 at 5:08 am

   ahhh… mind challenges… they are so darn HOT! Gimme, gimmeeeee!

   I go for my warm-up, be back soon.

   =)

   Btw, jurvey, are you the cartoonist?

   😛

   Reply
5. 

   Todd Huffman

   Oct 4, 2005 at 5:25 am

   Todd make coherent comment after sleep. Eight hours, more with the wakefulness. 🙂

   Reply
6. 

   Victor1

   Oct 4, 2005 at 5:47 am

   Perspectives… each point of view is a type of ‘measurement’, and a quality measurement is one made enough times to form a statistical ‘consensus’. So a variety of perspectives is the only way to improve the ‘measurement’ of something, hence, any one point of view is has a lower probability of being ‘correct’ in that sense. The ‘fitness’ of a perspective then, has to do with its diversity and the amount of evolutionary activity the various perspectives engaged in before the outcome viewpoint was arrived at. Even then, it can be completely ‘wrong’, as the majority does not determine what is most correct, but what works best for the majority. Evolution favors the fitness of the middle ground.

   So the solution of a problem is deemed ‘correct’ if it advances the middle ground of the perspectives, leaving the least and most intelligent viewpoints out of the picture…. till that perspective/solution fails, and the ‘curve’ re-evolves and is advanced in the direction of what the smarter perspectives knew in the first place.

   Example: 1/2 the unites states backed bombing Iraq and snubbing the Germans and French for not joining in on the stupidity. Now, with 2,000 americans dead (10k disabled) and 30k Iraqi’s dead (100k disabled), that ‘middle ground’ perspective has shifted toward what the smarter people knew to start with – it wasn’t a good idea to push democracy on a people who never asked for it (even as the ‘cover story’ for what other’s see as creeping imperialism sold as democracy). At this point, Saddam kept better peace than America does.

   Example: Clinton’s ‘stain’ on his own reputation (and Monica’s dress) was excused ‘oh well, everyone does that’, but a year later a poll found that the middle attitude changed dramatically toward being less tolerant of that whole episode – you know – lying to the supreme court and getting on TV and lying to the whole world. But being followed by Bush, makes him look good, and by extension, Carter is looking better all the time (winning a nobel prize isn’t in Bush’s future).

   The result? In any case, its called ‘reality of the moment’ – the middle muddle which is the evolutionary adaptation fitting (or enabling) the middle of the curve, not the edges.

   Smart scientist undergo the same problem in bringing radical improvements forward against the established thought, whether it takes years or centuries. Who is it that said something like ‘real advancement only happens when those holding the ‘established point of view’ finally die off and make room for the younger more modern thinkers who knew the next version of reality sooner’ ?

   Reply
7. 

   benjiman

   Oct 4, 2005 at 6:11 am

   I swear I saw Schrödinger’s cat in here somewhere. Does every persepective change everything? More perspectives, more diversity, more realities, better problem solving.

   This thread feels like its tip-toeing toward Goswami’s ideas about conciousness creating physics and everything else.

   Reply
8. 

   sweet thing

   Oct 4, 2005 at 6:18 am

   My favorite course in school (artificial programming in LISP) involved a competition to build an agent to solve a difficult game.

   Most of the great performers played tens of strategies in paralell and submitted it’s actual moves based on the strategy that was fairing the best at that time. The best agents used the most diverse and numerous strategies.

   Reply
9. 

   Victor1

   Oct 4, 2005 at 6:51 am

   Do the best agents also stay up too late writing on Flickr when they should be getting enough sleep?? !!

   Good thoughts folks, and good night.

   Reply
10. 

    jurvetson

    Oct 4, 2005 at 4:40 pm

    oddwick: very interesting. Is the essential component a fitness function alone, or do you also need a process of voting that preserves independence of opinion? I’d posit that diverse group performance comes not from convergence to the mean on a single parameter scale, but the factoring of diverse and orthogonal perspectives. Diversity brings more variables into the multivariate regression of teams.

    According to Scott Page, “People in diverse groups are less happy. Their views are challenged, and they feel like the outcomes were manipulated. Based on their experiences, they will self-report that it was not better than when they were on a homogenous team.”

    Regarding smart agents, “Imagine how physics would be if electrons could think” – Murray Gell-Mann

    Victor: As you increase diversity, complexity goes up, but then it drops and you get the central limit theorem. There is a sweet spot with just the right interplay between agents. Also, there is not one dimension that perspectives lie along. Diversity captures orthogonal perspectives and more adjacencies. The better the perspective, the less rugged the landscape (in terms of finding the global optimum and not getting trapped in local optima). Consultants can hop across local peaks without being any smarter or more experienced in their client’s business. The goal is not regression to the mean.

    Page’s conclusion that diversity is as important as ability seems pretty profound.

    His argument for diversity in complex adaptive systems seems to be to be the underpinning of that popular book by Surowiecki, The Wisdom of Crowds.

    Thinking about the wisdom of crowds as an emergence, this is the question I have been wrestling with:

    Does the minimal threshold complexity for interesting emergent phenomena necessitate inscrutability of results by members of the system?

    For example, if a group of diverse people routinely beats the experts, where does the learning occur? It seems to be at the system level, and not the individual level. The decision may make no sense to the individual members, but the decision making process does. The “wisdom” of the process could be taught to others, but not the outcomes.

    This generalization about emergence seems to hold for evolution, brains & neural networks, hives, and cultural memetic drift (more on this). In interesting systems, the emergent phenomena are at a different layer of abstraction, and may only be recognized by “in-process” or nodal members by pattern or proxy.

    Reply
11. 

    Todd Huffman

    Oct 4, 2005 at 6:40 pm

    "Does the minimal threshold complexity for interesting emergent phenomena necessitate inscrutability of results by members of the system?"

    Maybe. I have wrestled with this question before in the past, and have yet to come up with a satisfactory answer.

    As a human, I am perfectly capable of thinking on a ‘meta’ level. I can think in abstract terms about social dynamics and emergent phenomenon, et cetera. Contrast this to other systems which utilize collective decision making (ant colonies, fish schools, bacteria, neurons, avalanches, robots), they have limited self-awareness and little evidence they are aware of the ‘whole’.

    Plus! I can observe emergent phenomenon! Sit me down with a few beers and I will give examples of emergent phenomenon until I pass out, including ones involving humans. I can even give examples of systems composed of member systems composed of humans which exhibit emergent phenomenon, throw in whiskey and I can add at least four or five meta layers on top. I am a member in some of these systems, and in some cases on multiple levels of abstraction. If that doesn’t demonstrate scrutability, I dont know what does!

    But selection bias is a motherfucker… Perhaps I am only aware of a sub-class of the emergent phenomenon taking place. In fact, I am willing to bet this is the case. How complete of a definition can a member make of the sub-class they occupy without a definitive knowlege of the total? Is there some trick that could be used to bootstrap extra knowledge, even if admittedly incomplete?

    For example, autistic children generally lack the ability to view the world from another persons perspective, they lack a general theory of the mind (see en.wikipedia.org/wiki/Theory_of_mind ). Is something like this analagous to our inability to percieve emergent phenomenon? Could we examine this disorder and the conceptual framework it imposes upon the patients mind to build a theory delineating our own ignorance of emergent phenomenon in large complex systems? Autism comes in degrees, as does awareness of emergence. Is there a way to assess the level of awareness an individual has of emergent phenomenon, or is this like asking an autistic child to create thier own diagnostic criteria?

    Assuming transhuman intelligences arise, will they view us the way we view autistics? Beings mentally limited by the physical organization of thier brain? Will I be put into a preschool full of other humans, tended by the more caring superbeings?

    Reply
12. 

    Victor1

    Oct 4, 2005 at 11:31 pm

    Steve: to Murry Gell mann – never mind electrons…imagine what a society we’d have if enough PEOPLE could think.

    My thoughts on perspectives assumed an interaction of multiple linear or non-linear approaches, which when ‘multiplexed’, create additional perspectives.

    Diversity IS a form of ability, which is why the tendency toward elimination of diversity, puts the system at risk (Microsoft and Monsanto’s ‘terminator’ technology are examples that come to mind… the latter puts the human species at risk should a germ ruin the plants they’re forcing on us which produce no seeds of their own, the former puts software intelligence growth at risk through a similar hegemonical centralized control). All forms of consolidation (corporate, government, religious, thought-system, are the ebbing part of the cycle of diversity’s ebb and flow).

    Oddwick: the autism example is perfect, and to me explains why extra-terrestrials don’t interact with us. We’re their ant farms and such lucid discussion like ours here are an anomaly which won’t prevent the middle muddle from repeating the same lemming-like self destruction cycles we pretend are ‘one time’ events. WW1 was called ‘The Great War’ and Nobel’s invention of dynamite was supposed to end war becuase hoorors like WW1 and dynamite would be indigestably horrible to society and we’d never repeat that mistake.

    Where’s Benjimin? We need a good Tao perspective here. My stab at it: from the metacognitive perspective that Taoist (~right hemisphere) consciousness provides, the left hemisphere wranglings and endless re-definitions, are like watching a video of an endless fractal, swirling about in ‘diversely similar’ ways.

    I think we all concur that the central theme here, Diversity >> better outcomes, is sound. Its even more sound when the metaconsciousness ’emerging’ through meditative ‘thought’ validates it. However those who only know this consciousness via description won’t concur, since their knowledge set doesn’t include this experience, which isn’t discrete quantifiable knowledge but experiential, through practice – its a skill you develop, not a ‘product’ you can consume.

    8*}

    Reply
13. 

    angry shock

    Oct 5, 2005 at 12:49 am

    Ok, I admit that I haven´t read but a couple of words from the cartoon and the thread -stricking enough -… I really don´t know if this has something to do with the whole topic, but I was thinking about this, it appeared in my mind as a possible answer, an hypothesis to one current question: "Why I can so clearly see what to do and how to do it to solve a crucial problem in my life, but I am still here unable to do something about it? I can think about it easily, but can´t act it out!" I thought this… What if:

    …We are caught in the crossfire between the neocortex and the rest of our brain? As a result of still undeveloped aspects of this new feature -neocortex- we have been given thousands years ago. We can use it clearly, but didn´t master it yet, basically because it isn´t completely ‘finished’ to work perfectly.

    What if… We are given the faculties to think -in the most general meaning of thinking, a conciousness, Reason-, but we are still governed by our primitve brain -uncoiunciousness, perception, feelings? And this lack of or failure in an efficient communication of these two places/aspects of our brain is what immerses us in the alienation we all know about. We are aware of something we want to change, we ‘know’ how to resolve it, we know where to move, where to begin from, but we can´t operate after it.

    What if… This is probably the reason for many of our current and not so current problems as individuals and as a whole community? Our brain is still under construction, and due to this we are in a certain point in which we have a conciousness but we still aren´t able to fuse it, to interconnect it with the rest of ourselves with ‘happy’ outcomes. That´s why we can be great thinkers, brilliant and clear minds and yet be so difficult for us to act likewise. [Among many other side effects we ‘suffer’ and you may easily think of, too]

    What the metaphor about men being moved only by their hearts and not their heads tells us about.

    What if… We are structurally torn inside? How could we do otherwise but show it in everything we do?

    ————

    Apologises in advance for mixing this in the middle of the conversation, feel free to continue with the thread. Just wanted to share the thought before it blew away. =)

    Reply
14. 

    Victor1

    Oct 5, 2005 at 4:41 am

    Gi: Since this does change the topic a bit, and I don’t want to drag it further away, see my response on the latest photo on my page – I’m hijacking the thread! It answers this quite completely. 🙂

    Reply
15. 

    angry shock

    Oct 5, 2005 at 5:08 am

    great, V! thank you very much! Gonna read now. =)

    (sorry again for my offtopical behavour)

    EDIT: Shortcut to Vic, photo-thread

    Reply
16. 

    jurvetson

    Oct 5, 2005 at 7:11 pm

    Alieness & Dirty V: you know that my brain is still immature and child-like… =)

    The cortex is relatively new development by evolutionary time scales. After a long period of simple reflexes and reptilian instincts, only mammals evolved a neocortex, and in humans it usurped some functionality (e.g., motor control) from older regions of the brain. The reptilian brain is like a “logic”-centric era in our development that then migrated to a memory-centric model. (from Hawkins blog)

    Vic: “SimEarth” has God mode…. Someone is playing with the weather dial.

    Oddwick: I should have clarified “inscrutable” and “interesting.” We certainly can observe and reverse engineer (with time) emergent phenomena, and we can recognize patterns. But, unlike a designed system, the role of subsystems and comprehensibility of the system hierarchy are not immediately available. When we evolve an algorithm, like a sort algorithm, we immediately know how the evolutionary process works, but not how the resulting sort algorithm works. Danny Hillis did this, and the evolved sort algorithm was inscrutable. Sure, he could invest serious time and energy into reverse engineering the evolved system behavior. But will this help him understand the next complex system that he evolves? He would probably learn more about the process, but not necessarily about outcomes (in a meaningful engineering sense that provides a generalized insight).

    Hope that clarifies my question a bit. It relates to the dichotomy of evolution and design. Perhaps the “Wisdom of Crowds” is complex enough to be interesting within social systems. How should the individual member behave to affect outcomes? Is it obvious and apparent or deduced from trial and error and generalization?

    I think cultural norms and suprahuman emergence may be misleading examples to focus on because of their glacial time scales, giving us nodal members plenty of time to analyze and recognize patterns. Neural networks, evolved software, and maybe the wisdom of crowds are better examples.

    Another “perspective” on this… from a computational equivalence point of view, computational shortcuts for emergence and evolution do not come for free. In some cases, might they never be discovered?

    Reply
17. 

    angry shock

    Oct 5, 2005 at 9:57 pm

    I like you, for you never quit a question once you have formulated it…

    Again, u r like my cherished Prince. The child-like mind thingy, y´know.

    "…"The thorns–what use are they?"

    The little prince never let go of a question, once he had asked it. As for me, I was upset over that bolt. And I answered with the first thing that came into my head:

    "The thorns are of no use at all. Flowers have thorns just for spite!"

    "Oh!"

    There was a moment of complete silence. Then the little prince flashed back at me, with a kind of resentfulness:

    "I don’t believe you! Flowers are weak creatures. They are naïve. They reassure themselves as best they can. They believe that their thorns are terrible weapons . . ."

    I did not answer. At that instant I was saying to myself: "If this bolt still won’t turn, I am going to knock it out with the hammer." Again the little prince disturbed my thoughts:

    "And you actually believe that the flowers–"

    "Oh, no!" I cried. "No, no, no! I don’t believe anything. I answered you with the first thing that came into my head. Don’t you see–I am very busy with matters of consequence!"

    He stared at me, thunderstruck.

    "Matters of consequence!"

    He looked at me there, with my hammer in my hand, my fingers black with engine-grease, bending down over an object which seemed to him extremely ugly . . .

    "You talk just like the grown-ups!"

    That made me a little ashamed. But he went on, relentlessly:

    "You mix everything up together . . . You confuse everything . . ."

    He was really very angry. He tossed his golden curls in the breeze.

    "I know a planet where there is a certain red-faced gentleman. He has never smelled a flower. He has never looked at a star. He has never loved any one. He has never done anything in his life but add up figures. And all day he says over and over, just like you: ‘I am busy with matters of consequence!’ And that makes him swell up with pride. But he is not a man–he is a mushroom!"- [full]"

    Another offtopical entrance… this is becoming a pathology. I must visit a threadapist. =D

    Reply
18. 

    Todd Huffman

    Oct 5, 2005 at 11:08 pm

    Steve, so what you are saying is that evolution doesnt comment code very well? Maybe this is a suprise to computer scientists, but as a biologist, I have been trying to understand poorly commented evolved solutions for years. 😉

    Re: the Hillis work. He could have reverse engineered it, so it wasn’t inscrutable, just inconvenient. The biologist in me empathizes, but I believe there is merit the reverse engineering.

    Perhaps Hillis and co just arent looking at the problem through the right filters. In biology, we are constantly discovering facts which alter our perceptions, and make the whole system much clearer. A recent example, it was discovered the human genome has far fewer genes than anyone predicted. This forced biology to re-examine the process of protien expression and we are currently realizing how important post translational modifications are, and discovering a whole slew of new gene expression regulatory mechanisms nobody really looked for until now.

    Learning how evolved systems work is difficult, they are black boxes filled with intertwined and recursed black boxes. If you pick a subsystem and open it, its function will not be readily apparent. The next box you open will require going back to all previous boxes and reformulating your knowledge. The process is long and painful.

    Once one of Hillis’ systems gets (cyber)cancer, I bet the tune about inscrutable changes.

    Reply
19. 

    benjiman

    Oct 7, 2005 at 5:35 pm

    Holy crap. I leave for a day and we’ve all evolved. Great thread. I’m going to have to catch up on this one…

    Reply
20. 

    pentimento

    Oct 20, 2005 at 2:20 pm

    well done!

    Reply
21. 

    periodic wealth

    Nov 2, 2005 at 5:36 am

    My bumper sticker says, "I never want to live in a world so small my mind can comprehend it."

    Infinity is really big. Any idea you can encompass has already been thought. All new ideas are already older than time. And this just encourages me to inquire ever deeper. Ever further. Into reality and meta-reality. But it is the fact that there ARE mysteries — deep inscrutables, so to speak — that cause me to thank God and delve deeply, knowing I will never span the breadth of reality regardless of how evolved I get, or not.

    Reply
22. 

    jurvetson

    Dec 5, 2015 at 8:03 am

    and 10 years later…

    

    Reply