P=NP Proof/Conjecture
April 22, 2010 11:34 PM   Subscribe

P=NP Proof/Conjecture
P=NP is one of the greatest open math problems, and has vast consequences in a wide variety of fields. While certainly posed in the terms of precise mathematics, could the solution actually be posed in a very different format? While my formulation and proof may not look like anything out of a math book, the terms used are from the frameworks of economics and information theory which have deep mathematical connections. I challenge anyone to attack the substance of the theory, feel free to email me at lind@yahoo.com. - Jason Lind
posted by jlind0 (101 comments total) 9 users marked this as a favorite

I know you seem to think this theory is a big deal, but as someone who has done a decent amount of computational complexity proofs this seems very Time Cube-y. Aside from the fact that a P=NP proof is very unlikely, I would at least expect a proof to at least mention polynomial time or something else directly related to the problem.

Also, I hope that you didn't join MetaFilter just to post this.
posted by burnmp3s at 8:00 AM on April 23, 2010 [2 favorites]


Thanks for the reply, I actually state that P = NP if Artificial Intelligence is NP-Complete. Honestly I don't know if that's true or not, I'm clearly not an academic; this is a thought experiment that came out of some other work. However that is a very good point, and I will address that.

Also while I did join mefi specifically to post this I've been a reader for a long time and do have more to contribute.
posted by jlind0 at 9:29 AM on April 23, 2010


I added a couple paragraphs regarding the connection to P versus NP as requested by burnmp3's, same URL. Thanks!
posted by jlind0 at 9:46 AM on April 23, 2010


Pretty timecube-y. I'd suggest you remove your real name and university from the post so this doesn't come back to haunt you in a few years when you're looking for a job.
posted by miyabo at 12:51 PM on April 23, 2010


I realize it's a bit out there, but I am pretty confident it's not crack pot. Whether it is completely correct or not I don't know, but my willingness and effort to inspire proper peer review should be a sign that I have not lost perspective.

I actually run a successful software consulting practice and haven’t been in school for some-time, I have no problem with anything that would come up in a search engine when you search my name including this.
.
posted by jlind0 at 1:26 PM on April 23, 2010


I found the layout of this a little hard to follow (to the point where I couldn't really read through it) - the column is narrow, the font's vary in size quite a bit. Also, minor quibble, but I don't find the reference system robust (no references, hyperlinks etc).

I don't mean to be overly negative - just providing feedback about how I find this challenging to understand and process.
posted by a womble is an active kind of sloth at 7:32 PM on April 23, 2010


That's not a bad point, I'll think of a better way to present it.
posted by jlind0 at 8:25 PM on April 23, 2010


As requested: http://idealorg.wikia.com/wiki/Ideal_Organization_Theory_Wiki
posted by jlind0 at 9:11 PM on April 23, 2010


Maybe you should read a little bit about complexity theory.
posted by chunking express at 9:41 PM on April 23, 2010 [4 favorites]


This is not explained clearly enough to evaluate it. I mean, obviously if you had completely formalized this you would have already won a Fields Medal, so I can't ask you to go that far. But you need to do better at explaining how your argument works, so a skeptical outsider can know what you're trying to talk about.
posted by grobstein at 9:55 PM on April 23, 2010


I added an arguments section on the front page, does that help at all?
posted by jlind0 at 12:04 PM on April 24, 2010


Ideal Organization Theory is a P algorithm for forming organizations whereas artificial intelligence is probably NP-Complete.

jlind,

I don't agree with your proof at all, for two reasons.

Your notion of Ideal Organization Theory (IOT) is not a P algorithm. Because your Actors in IOT already have bona fide Natural Intelligence. In other words, your IOT system is cheating.

Similarly, the problem of creating an AI is not NP-complete. The only known way of verifying an AI is with another AI, or with an NI (namely us). This distinction exceeds the definition of NP-completeness.

It is kind of neat to wonder about the connection between computation and man-made economic systems. Thanks for bringing it up on a boring Saturday afternoon. :) Also, there is probably some academic research out there already in this area--you should try to check that stuff out.
posted by polymodus at 1:59 PM on April 24, 2010 [1 favorite]


polymodus: Actually the actors in IOT only require self-determination, which I would argue randomness is a form of, insomuch that the actors and not the system is making the determination. Even rudimentary AI, which has been achieved, would be considered self-deterministic.

The fact the verification of the answer is not in P is an interesting point, however the solution seems to be. Once verified however the intelligence would take on a life of its own and further verification would not be necessary, which means the underlying collections creating the Artificial Intelligence would still be running in P.

I'm not completely sure what the ramifications of what I have posed are, P = NP was not the goal of this work when I started but the connection seemed obvious.

If there is any specific work you recommend I would be very interested in checking it out, keep in mind I'm not an academic and this is more or less a hobby of mine.

Thank you for the great feedback!
posted by jlind0 at 2:26 PM on April 24, 2010


Clean up the possessives and plurals (those apostrophes will short circuit most people's
judgment before they even finish the sentences). And if you haven't already, take a look
at Douglas Hofstadter's Godel, Escher, Bach: An Eternal Golden Braid. He got a
Pulitzer for that thing, and it looks more like where you are headed than a P = NP proof.
posted by the Real Dan at 9:03 PM on April 24, 2010


I also noticed such things as "fly's" when you meant "flies" in 16.0.

I found it too confusing to follow otherwise. For example, when you say "self-determination" above, is this meant to mean something akin to "deterministic?" When you define A.I. as "passing the Turing test," the rigorousness of the definition is lacking (which evoking Alan Turing doesn't bring in to existence.) Ultimately, mathematics isn't about the world but is about mathematics so if you are going to reference the world, you need to do so in a mathematically rigorous way. E.g. a point or a line in geometry is only named the same as a point or a line in the real world to make it more interesting to mathematicians who (presumably) live in that world. They could be named "cat" and "dog" with no loss of generality (or even switched "line" for "point" and "point" for "line" as in projective geometry.)

For your entertainment, I include a link to a (flawed) proof that P != NP I made once in Metachat.
posted by Obscure Reference at 7:56 AM on April 25, 2010


It's cool to really think about new ideas in the P=NP problem, but this is a little bit oddball. It would be best to remove the part about Wikipedia because it seems to completely misunderstand (or just not understand) why this kind of article would not be allowed on Wikipedia - that Wikipedia is not the place for 'original content,' but rather as an encyclopedia about things that exist. And using it as a place for 'peer review' is misguided - articles are not reviewed like peer review, and for that, you would better run this by a few CS professors. Also I would really recommend not defining QED and referring to such characters as Galileo Galilei just by themselves - giving parenthetical citations at least, and not identifying them as "founder, Scientific Method." And the quotes-method of proof is just quite weak. One must imagine that Shakespeare didn't defend his "world's a stage" quote and give a reasonable proof from which your proof can draw.
posted by tmcw at 8:20 AM on April 25, 2010 [1 favorite]


tmcw: I actually had a great email chain with Jimmy Wales about this, as I did try to post this on Wikipedia and both of us concluded while Wikipedia is not setup for this today it’s something that could happen in the future. He ended up putting me in contact with someone over at the Wikimedia Foundation to possibly add a process of peer review of original work, such as this, accessible to the Wikipedia community.

I kept the reference in there as if this actually did inspire true peer review that took this from the loose and abstract nature it is in today and made it into a full blown proof, which I contend would happen, then that is proof of this in and of itself.

The thesis is above all an equation, and while I know exactly what it says I realize other people do not, so my formalizations attempt to shine light on the underlying concepts while inspiring debate. Identifying them by the major work also is part of the formalization.

Shakespeare by the way was a genius and I consider him the original information theorist.

The model suggests the fastest way to achieve consensus is through open development, and I'd argue through the comments on this chain that is already occurring. I welcome you or anyone else to add to the theory on the Wikia page where you see fit.
posted by jlind0 at 9:48 AM on April 25, 2010


Obscure Reference: I'd argue the real world is mathematics, and therefore a proof in the real world is just as mathematical as a proof in the mathematical world. Invoking a quote from Alan Turning, assuming it’s the right one, is perfectably acceptable to define Artificial Intelligence, I do not see any flaw in his definition at the level of abstraction I'm working at.

Obviously if someone would like to formalize these definitions further the wikia site is open for editing.
posted by jlind0 at 9:54 AM on April 25, 2010


The model suggests the fastest way to achieve consensus is through open development, and I'd argue through the comments on this chain that is already occurring.

This is not the venue to get your theory reviewed. Perhaps you could sit down with a computer scientist or mathematician in a nearby university to discuss it.
posted by odinsdream at 12:34 PM on April 25, 2010


odinsdream: That's a heck of a statement to make with no research or theory to back it up. I'm talking with professional researchers as well, but according to the model a forum such as this is much more efficient as the number of actors is way greater.
posted by jlind0 at 1:07 PM on April 25, 2010


That's a heck of a statement to make with no research or theory to back it up.

Comical, coming from someone claiming to prove P = NP without a wisp of math.

What "model" are you referring to? The reason I suggest this isn't the place is that, while your project's existence on the web may be interesting to the Metafilter community, that interest is not going to translate into scholarly utility because Metafilter, though frequented by a bunch of smart people, is not a scholarly community.

If you're actually looking to learn something about your theory (I use this term very, very loosely in reference to your website), you should talk to someone who has spent years of their life studying it. That person is likely going to be able to point you in useful directions because they understand the math behind it completely. Scholarly research is not what Metafilter is designed for.
posted by odinsdream at 1:40 PM on April 25, 2010 [1 favorite]


P vs NP is a mathematical problem. A proof either way would have to be extremely rigorous. You're bringing up very ill-defined concepts, like artificial intelligence which is mostly a mission statement. Intuition is useful only when the underlying concepts are sound enough that you are able to get rid of it.

If you want to study P = NP, work with the limited framework of computational complexity. If you want to study IA, find a small problem that can be solved by the application of a practical technique. If you want to study philosophy, don't hope to pass it off as mathematics.
posted by Tobu at 2:08 PM on April 25, 2010 [1 favorite]


odinsdream: Just because something was not designed for such use does not mean it cannot be used in such a way. The feedback received thus far has been largely helpful, and indicates that MetaFilter is a place for scholarly research.

I continue to attack on all fronts; MetaFilter is only one of those. Several experts are reviewing this work as we speak and consider it credible enough to not outright dismiss.

Tobu: While P vs NP is formulated as a mathematical problem, its implications permute virtually everything not the least of which philosophy. I've formulated the philosophy into an equation that can be evaluated by various branches of science. I contend if this equation can be demonstrated as true, which I feel is clearly self-evident, than P is in fact NP.

Even if Artificial Intelligence is not NP complete, my definition of artificial intelligence has profound consequences and I feel that rapid peer review is needed to determine if what I posed indeed has merit.

I am not an academic, this is more or less a hobby, and I do not have the training nor inclination to turn this into accepted scholarly format myself. I leave that to those who do have that training, I am merely attempting to promote what I feel is a valid and interesting approach with vast applications.
posted by jlind0 at 2:22 PM on April 25, 2010


Also Ideal Organization Theory (the right side of the equation) is a model for how the best organizations are run. It indicates that market-oriented organizations (like MetaFilter) outperform hierarchal organizations (like a university), however to achieve the best results a combination of the two is required.

MetaFilter is probably too far to the market side of the wall, as is Wikipedia. Wikipedia is a biit more structured and should be far more efficient and effective at evaluating something like this. However MetaFilter is still in a better position than most organizations.
posted by jlind0 at 2:34 PM on April 25, 2010


jlind0; since this is your first foray into this community, please observe that this area of the site is designed as a kind of staging area for the main site, the idea being that someone reading this may find it interesting enough to post it there.

While I grant that the main area of Metafilter may be a nice place to debate the merits of your hobby, this is absolutely not the place. At the most, this is the place to debate why your hobby should be posted on the main page of metafilter, and that's arguable. It's absolutely not the place to get scholarly feedback on what you seem to think is the answer to a quite famous mathematical problem. An answer that includes, literally, no math.

Even if Artificial Intelligence is not NP complete, my definition of artificial intelligence has profound consequences and I feel that rapid peer review is needed to determine if what I posed indeed has merit.

Rapid? I'll save you some time. It's not even internally coherent. Sometimes your sentences barely make sense individually. It's not spelled correctly. You have no understanding of what an equation is. If you have the time, take a community college algebra course.
posted by odinsdream at 8:28 PM on April 25, 2010 [3 favorites]


odinsdream: I'd like to start by thanking you for your aggressive approach to what I have posed. I'm not trying to be overly aggressive here but it may come off as that, and I would like to think that your comments so far have likewise been constructive and not hostile.

As I've stated before this is a meta-equation, but an equation none the less. Artificial Intelligence is on one side and Ideal Organization Theory the other, expressing equivalence. Just because you've never thought of this kind of equation doesn't make it any less of an equation than x^2 + y^2 = 10.

I actually completed Differential Equations by the time I was 15 and was a full time student at Marquette University in Electrical and Computer Engineering by 16 so I'm sure a course in algebra would serve no purpose for myself. I would suggest you take a course in Boolean Logic if you don't understand why what I have posed is an equation.

If there are areas that confuse you maybe you should ask before just calling it incoherent, as there are many people much deeper in their fields than you or me, who while calling this badly organized don't call it incoherent.
posted by jlind0 at 10:31 AM on April 26, 2010


The P vs. NP problem is a computational complexity problem. Your essay should probably discuss algorithms and running times at some point, since this is how the problem is defined, in terms of complexity classes. Your essay comes off more than a little half-assed and rambling. At times, it seems muddled. Your opening statement about artificial intelligence is incoherent. It might also be badly organized.

"Ideal Organization Theory is a P algorithm for forming organizations ... "

You say so without demonstrating this is the case. For something to be in the class of problems P you need a polynomial time algorithm that solves "Ideal Organization Theory". And based on your description of the problem it's really not clear whether this is solvable in polynomial time or not.

"... whereas artificial intelligence is probably NP-Complete."

Saying something is 'probably NP-Complete' is pretty meaningless.

For something to be NP-Complete, there is a reduction from the problem to any other NP-Complete problem. That's what it means for something to be NP-Complete. Can you think of a way to solve the traveling salesmen problem with an oracle that can pass the turing test? That's something to think about I guess.

"If this equation is demonstrated to be true, and artificial intelligence is in NP-Complete, then P = NP."

I guess so.

Here is my proof for P=NP: If [some random problem in P] can be reduced to [sum random NP-Complete Problem] than P=NP. Although I am using words and skipping over details, the mathematics underlying my proof are sound.

I actually think my proof makes more sense than yours.
posted by chunking express at 1:49 PM on April 26, 2010


Except that I say this problem in P (Ideal Organizational Theory) reduces to this problem in NP (Artificial Intelligence). I am not an academic, clearly both sides of the equation need to be shown what complexity group they belong in and I'm not capable of doing that.

My understanding of complexity theory, coming from my experience as a software engineer, gives me an abstract high level understanding of P and NP, certainly enough to normally make the correct classification without the need to formally prove it. I am very confident IOT is in P and AI is not, if someone wants to formally disprove those statements then this work obviously has no bearing on the P=NP question. Even if that is the case defining AI such as this should have applications in multiple disciplines.

However what should be clear is that both sides are indeed saying the same thing and are equivalent, I could never do as much justice to the concepts behind those words as the people I quoted.

I have taken another stab at explaining myself, please see the wiki entry (click on the words in the equation for an explanation of them).
posted by jlind0 at 2:24 PM on April 26, 2010


My understanding of complexity theory, coming from my experience as a software engineer, gives me an abstract high level understanding of P and NP...

Clearly it doesn't.

... certainly enough to normally make the correct classification without the need to formally prove it.

That's kind of the whole point of a proof.

I am very confident IOT is in P and AI is not, if someone wants to formally disprove those statements ...

Why would anyone do that, when you are too lazy to formally prove those statements? I'm not trying to be a jerk, but your approach to this 'proof' seems ass-backwards. It's no one else's job to disprove your hand-waving.
posted by chunking express at 3:23 PM on April 26, 2010


I think that I've posed something interesting and useful, I don't have the skills to do formal mathematical proofs of this nature, my hope is people who do will find it interesting enough to attack with the full rigor of their training.

My approach was to prove with highly abstract concepts that Artificial Intelligence is what I call Ideal Organizational Theory, and if someone truly understands all the concepts involved I contend my thesis is proof enough.

This is deep stuff, and there are not many people with the backgrounds required to immediately grasp the substance of my attack, however if you sit back and attempt to understand what that equation really says I think that you will agree it is obviously true.

If one sits at the micro level of their discipline and only attempts to understand this concept in the terms of their discipline they will never grasp my attack. Read my guidance, and the selected quotes, surrounding the terminology. Really think about them and how they relate.

If this equation is correct it has profound consequences that affect all of us, and requires all disciplines understand it which is far more than any one person can expect to accomplish. I have made the initial formulation and now I am championing it for people to attack with their knowledge, it may not be your job but if you see something here I'd argue it is your responsibility to help.
posted by jlind0 at 4:22 PM on April 26, 2010


Your web page keeps promising a proof, but never seems to actually give one. You should probably avoid using the word "proof" without giving a mathematical proof. Maybe say "argument", instead?

When you say things like this:
10.0 Groups of actors can be defined by their formal and informal associations. “…a group is an algebraic structure consisting of a set together with an operation that combines any two of its elements to form a third element.” – Wikipedia
you give the distinct impression that you don't know what you're talking about. Why are you quoting Wikipedia about the abstract algebraic structure called a "group"? It doesn't have any relevance to the way you're using the word "group". Many of your other quotes have this problem.

When you say things like
If one sits at the micro level of their discipline and only attempts to understand this concept in the terms of their discipline they will never grasp my attack. Read my guidance, and the selected quotes, surrounding the terminology. Really think about them and how they relate. If this equation is correct it has profound consequences that affect all of us, and requires all disciplines understand it which is far more than any one person can expect to accomplish.
you come across as a crackpot. (Sorry, there's no nice way to say that.) If you want people to take you seriously, you have to stop dismissing their objections by saying they don't know enough. This is doubly true when you say, right off the bat, that you don't know enough to express things formally.
posted by samw at 4:49 PM on April 26, 2010


samw: I think I've formalized things a bit better on the wiki, you may want to check that out. I can see why people do not see the proof in it, but I maintain that's a lack of understanding of the format and the concepts. If there a specific questions I can answer them.

I use the group theory definition of group since organizational groups are actually an abstraction of that concept. Groups are formed when two or more sets come together (the operation being collaboration) which then results in a new element.

I tried to use quotes from people as best as possible, unfortunately when it comes to group theory I was unable to find any satisfactory quotes driving at what the term "group" means in the context of the equation, so the horribly imperfect citation of Wikipedia was all that was available.

Thank you for your specific question I will update the wiki to reflect the very valid point you brought up and I hope I better explained. As far as being a crackpot, maybe I am I don't know, only time and proper review will tell. But I'm really interested to find out.
posted by jlind0 at 5:11 PM on April 26, 2010


This is deep stuff, and there are not many people with the backgrounds required to immediately grasp the substance of my attack, however if you sit back and attempt to understand what that equation really says I think that you will agree it is obviously true.

Your tack here is unbelievably arrogant. Perhaps it's equally possible that you're deluded by the levels of obfuscation you've created to "explore" this problem and have reached a point where you are unable to recognize your hubris.

If this equation is correct it has profound consequences that affect all of us, and requires all disciplines understand it which is far more than any one person can expect to accomplish. I have made the initial formulation and now I am championing it for people to attack with their knowledge, it may not be your job but if you see something here I'd argue it is your responsibility to help.

I'm going to be polite and not call your work crackpottery, but know this: It's indistinguishable from crackpottery. It has all the hallmark signs: You believe you alone have come up with a brilliant solution to a massive problem without any formal training. You believe that your solution is so brilliantly complex that to criticize it would require an equally brilliant mind. In fact, to even glimpse the brilliance of your solution you'd have to be a genius. You believe your solution has somehow been ignored all this time precisely because people trained in examining this problem are too muddled in the details. All you're lacking is a section on some ivory-tower organization's efforts to quash your original research. Oh wait! Curse you Wikipedia elitists! If only you accepted my ramblings as the obvious fact that they are, you could include it on your site!

Here's the more likely explanation: Your webpage is incoherent. It doesn't consist of anything resembling logic. It contains literally no math at all. You cannot even begin to discuss this problem without math. Do not pass go, do not collect $1 million dollars.
posted by odinsdream at 7:54 PM on April 26, 2010 [5 favorites]


BTW, welcome to Metafilter. We have a lot of interesting things here. Please check out the rest of the site. Ask Metafilter may be something you would enjoy.
posted by odinsdream at 7:59 PM on April 26, 2010


odinsdream: I hope you are enjoying this as much as I am, I really do :)

I'm not saying you need to be a genius to understand this, maybe understand it at first glance with no additional research, but I think that if anyone takes the time to really understand the terms I used they will understand what I'm talking about. I've done my best to guide people to the understanding I have of the terms, and continue to refine it as people bring up reasonable and specific arguments.

As I said Jimmy Wales and I had a great email chain, his objection had nothing to do with the accuracy of my work it had to do with the ability of his organization to verify it. He also recognized that this is something that could change with modifications to the site, and eventually agreed with me that peer-review of original material was actually a natural evolution of his organizations mission.

I freely admit I could be wrong here, when dealing with subject matter this complex it is often impossible for even the author to determine whether it is brilliance or madness. Personally I have been unable to find a flaw in the logic of the equation.

Whether it truly solves P=NP even I admit that's determined by the complexity theory classification of AI and IOT which I am unsure of. However while I posed it as such to get people’s attention, the main point of interest is if I've defined AI in a new way, which I believe I have.

So is this crackpot, I don't know. That's for the public domain to determine not me, I believe that time will quickly prove that there is something here, and I can't wait to find out.
posted by jlind0 at 8:26 PM on April 26, 2010


You mix precise, technical, mathematical terms (P, NP, complexity theory, etc) with unprecise, nonmathematical terms/ideas (AI, whatever it is you mean by "group"), and then you try to draw precise, technical conclusions (P = NP). It doesn't work that way.

You are so confident about your result, but you simultaneously admit you have no grasp of the details. It doesn't work that way.

You claim this is all deep and complex, so where's the detailed explanation of the argument? The explanation of the background to the problem? The acknowledgement of holes left to be filled, problems left to be solved?

Here's what it looks like from my end: you've tossed out some idle speculation, and not done any work to make it make sense to anyone but yourself. I'm not saying that's the case, but that's what it looks like on my end. Given how it looks to me, why would I put much thought into any of it? You don't appear to have!
posted by samw at 8:55 PM on April 26, 2010 [1 favorite]


I'm not writer, and this is a hobby, so I really don't have the time nor inclination to deal with this at more than abstract level right now. This is a thought experiment I've been grappling with for basically my entire life, I have a deep practical background in information theory, a deep informal background in economics and as such a background in set theory.

Given the training and experiences I have had I have been able to express this theory very succinctly in my thesis and been able to provide a cursory background on the terminology. As I'm sure you can see further formalizing this abstraction gets technical and complex very quickly, and I am not trained in most of the accepted methods of doing that.

While my form is radically different from accepted practice I contend it is sound and valid. As far as the equation is concerned there should be no holes or problems left to be solved, the holes and problems left to be solved are only relating to the P=NP portion of the problem.

I probably shouldn't have even brought up the reference to it, but the connection is so apparent to me, as I'm sure it would be to anyone if they understood what that equation says, I felt it should be included. It is pretty clear I jumped the gun on that one, I would ask only that you analyze the definition of Artificial Intelligence as it relates to IOT.
posted by jlind0 at 9:11 PM on April 26, 2010


I would ask only that you analyze the definition of Artificial Intelligence as it relates to IOT.

But you haven't done so. Why should I? This is your hobby, right? Not mine. I suspect you will claim you have, in which case all I can say is you need to work on your writing. If you want to get anyone interested in what you have to say, you need to communicate it clearly. What you've written comes off as nebulous and arrogant. That's no way to get someone interested.
posted by samw at 9:49 PM on April 26, 2010


Odin,
This is bad
I've seen worse...from people with PhDs in CS
It doesn't matter, it's only the Internet
posted by miyabo at 6:33 AM on April 27, 2010


While my form is radically different from accepted practice I contend it is sound and valid. As far as the equation is concerned there should be no holes or problems left to be solved, the holes and problems left to be solved are only relating to the P=NP portion of the problem.

Your contention that it's sound and valid is based on nothing concrete. You have not created an equation. Stop saying that.

Word problems are not an alternate way to represent equations. They are equations in their own right. The reason word problems work is that they are still based on mathematical properties, like associative, commutative, distributive, etc.

For example, the word problem:

Joseph has two apples. He gives one apple to Susan. How many apples does Joseph have?

Would represent the equation:

2 - 1 = x

Nothing about the words or the equation violates any basic mathematical properties.

Here's another word problem:

Joseph has two apples. Susan gives him five blue circles. How many buckets of water does Joseph have?

This cannot be represented in an equation because the associative property cannot be applied in the case of apples, circles and buckets of water. This is what you're doing.
posted by odinsdream at 6:35 AM on April 27, 2010 [5 favorites]


Link (act three).

I don't mean to be disrespectful. I just want you to see how you seem to be coming off to some of the laypeople here, myself included. If you're truly committed to this, you may need to change your approach to avoid being identified as this kind of thinker.
posted by Valet at 8:31 AM on April 27, 2010


samw: I've attempted to communicate it as clear as I possibly could through that equation. In electrical engineering they teach you to work with "black boxes" when designing analog circuits, and how to test what those black boxes do. These words can be treated as black boxes but only you can determine what they mean in the context of the way that you think.

I've added some guidance on how I intend the terms to be used, however what I state is unwavering and literal, meaning that the equation captures everything that those can mean when placed together. I realize how abstract that is.

odinsdream: I am saying Artificial Intelligence is equivalent to IOT, I'm not sure how much more of an equation I can make it.

Valet: I do not mean to come off as disrespectful, it is difficult to interact with someone if they don't have a specific question and are just attacking the obviously abstract nature instead of the substance.

miyabo: What exactly are your problems with it so I can make the improvement’s?
posted by jlind0 at 9:31 AM on April 27, 2010


I'm not sure what 'attacking the obviously abstract nature' means. I have no ability to judge your ideas on even an abstract level. I'm talking strictly about your social approach -- the way you're sharing your ideas with others. I'm actually trying to help, but I understand that you may have started to feel defensive, so I'll bow out. Good luck.
posted by Valet at 9:38 AM on April 27, 2010


Valet: I did not mean to imply you were attacking the abstract nature of what I have posed, there are others on this discussion who have been. I was referring to how as far as the social dynamics go I don't see any better way to engage those people than to challenge them to leave the microcosm of their discipline.

When there's a specific question I can answer that easily. When some is attacking the validity of abstract thought, that's more difficult. I am obviously being defensive, but that's out of design, people are attacking my ideas, which is what I want, and therefore I have to defend my position.

If you have any concrete suggestions on how I could handle this discussion better I'd very much like to hear them.

Thanks for the effort!
posted by jlind0 at 11:38 AM on April 27, 2010


I've attempted to communicate it as clear as I possibly could through that equation.

And I'm telling you you've failed. It's incumbent on you to fill in the black boxes enough to make it begin to make sense to others. You have not done so, so don't complain when people can't make sense of it. If you leave it to others to fill in the black boxes, they'll fill them in with "this guy doesn't know what he's talking about".
posted by samw at 12:39 PM on April 27, 2010


It would be helpful if there was a specific box you didn't understand and I could attempt to clarify it through your misunderstanding.
posted by jlind0 at 12:58 PM on April 27, 2010


Here's the problem you need to address with your site: nothing you've said makes me think there even is anything in the boxes. Your comments in this thread only reinforce that impression. If that's the impression you give, why would anyone spend time trying to get you to explain what you mean?
posted by samw at 2:57 PM on April 27, 2010


Ohh, I get it now.

Here are a few things from time spent in the humanities that might be useful:

1) Try to carefully elaborate self-defined terms. Even though it can be more work, use terminology from conversations that have gone on before you -- the ideal is to use existing terminology first, and then your own in parallel, so that people who care (and are familiar with the state of the conversation about a topic) can use the old conversation as a sort of bridge. Sometimes it seems like 99% of the disagreements between authors and readers come down to "that's not what I mean by X".

2) In the same way, build on previous results or previous conversations -- even if to say, "I find previous, specific result X (with citation) to be completely wrong". This doesn't have to do with the truth of a thing, but its expression -- people will naturally be trying to narrativize an academic discussion, and you should be providing that narrative.

3) Do your reading and footwork. People always talk about great advances coming out of nowhere, but in many fields, that's a relative observation -- Grigori Perelman has a Ph. D., worked at the USSR Academy of Sciences, and travelled around the US for years teaching classes and giving lectures. Even though he has a research position at a relatively small organization in St. Petersburg, he is still a part of a (far-flung) intellectual community, and doubtlessly has people he trusts and respects to read his work, to talk to, and to exchange ideas with. Put another way, he keeps up with what's going on *outside* his mind. It's intellectually important (that's just my opinion) -- but it's also professionally important (that seems clear to me from experience).
posted by Valet at 6:52 PM on April 27, 2010


Valet: Great!

(1): Trying to, it's a process and so far it has not been going as well as I'd hoped. However there have been some great comments on this thread that have enabled to clarify a bit, although it seems still not enough.

(2): Yeah good idea.

(3): Oh for sure, I read and discuss with a wide variety of people regularly, I'm always down for an intellectually stimulating conversation.
posted by jlind0 at 9:03 PM on April 27, 2010


burnmp3s's first comment has it, I think.

You're talking about computational complexity without mentioning algorithms.

Some specific points:

In the unwavering and literal terms of set theory, economic theory and information theory; for all sets, under all conditions where the actors have self-determination the author poses:

What do you mean by "unwavering and literal"? This sounds like a fortune teller warming up his gullible audience.

What do you mean by "for all sets"? You don't mention the chosen set anywhere else in the article. What happens if I choose the set {3}?

Those are the first two mathematical issues I have with the article.

This could also be called “Ideal Organizational Theory” as it is the contention that the classification of sets, involving self-determining actors, cannot be optimized further than the thesis.

What do you mean by "the classification of sets"? How are "self-determining actors" involved? What do you mean by "optimised"? You're not talking about algorithms until you've defined the problem being solved and the steps taken to solve it.


There might be some interesting ideas in here, but I can't understand it. As far as I can see, it's another Time Cube.
posted by richb at 9:09 AM on April 28, 2010


richb: By unwavering and literal I mean I'm using the words in all that they mean, regardless of level of abstraction. These terms can be reduced to a core concept, and through the quotes and my simplistic descriptions I'm attempted to convey what I see as the best reduction.

When I talk about sets, I'm referring to sets of actors which is implied by the use of finite/non-finite. Also I make the distinction the sets require self determination, so {3} wouldn't be valid. {Math.Random()...Math.Random()} would be.

So when I say classification of sets I refer to the definition of the structure of those sets, and optimizing that definition so that it is the structure.

Does that make more sense?
posted by jlind0 at 10:34 AM on April 28, 2010


If you're using the words in all that they mean, then "for all sets" should mean just that.

If I see a mathematical proof which starts "for all sets", then usually they continue something along the lines of "for all sets X where Y, X has the property P(X)". The first thing I often do is choose a set with Y, say "{3}" and check if I can evaluate P({3}) to confirm that a) I understand P and b) that the proof holds for a trivial case.

I know you're not claiming that this is a mathematical proof, but the same process ought to apply for word proofs.

You've said "for all sets, under all conditions where the actors have self-determination".

Does that mean "for all sets of actors with self-determination"? Or is the word "conditions" important here? If so, could you explain how, ideally with an example where given the wrong conditions, the theory doesn't hold & an example of conditions where it does?

Assuming that you do mean "for all sets of actors with self-determination", for me to "plug in the numbers" and see if the theory holds for an example, I would need to choose a set of actors with self-determination. You haven't explained what you mean by "self-determination", so I can't do that.

I don't think that there is a generally accepted understanding of what "self-determination" means, so any author which uses that term must explain what they mean by it to be understood. See Valet's comment above.

I'm intrigued that you think that Math.Random() has "self-determination". This makes me think that you mean "self-determination" in a distinctly non-standard sense. You might like to use that example in your explanation, as well as some counter examples.
posted by richb at 2:10 AM on April 29, 2010


If the set is random, I would say by definition it is determining itself. Even under limited conditions where an actor has the choice of Yes/No, in order for it to be self-determining the answer is random as far as the classification scheme is concerned, ergo the classification scheme doesn't really care if the self-determination is by a primitive algorithm or by intelligent choice.

So I should further clarify that while the actors are sets, so are the organization structures, in fact the organization structures would look like something out of group theory (as I've mentioned above). So the structure/classification would be set(s) that are built from sets of self-determining actors. That is what I meant by "conditions..." in the original statement.

Honestly the key components of this theory was written when I was in what’s considered a manic state, which I have been out of for about a month. So even I am trying to figure out the details of what I meant, I don't know if you've ever had a maddening moment of clarity but that's where this came out of.
posted by jlind0 at 1:30 PM on April 29, 2010


Honestly the key components of this theory was written when I was in what’s considered a manic state, which I have been out of for about a month. So even I am trying to figure out the details of what I meant, I don't know if you've ever had a maddening moment of clarity but that's where this came out of.

This is not a good sign, and this is readily apparent as I've read the "theory" as well as your replies in this thread.

I've had "maddening moments of clarity" and they rarely amount to anything unless I'm making art or music or something similarly nonscientific, and even then it's iffy. Genius is easy. But that's maybe 10% of the game - the other 90% is a lot of hard, disciplined and careful work and even then it can go pear-shaped in a hurry.

Mania doesn't give you super-powers of cognitive ability or clarity. If you were able to re-read your theory from an outsider's perspective as well as your own dismissive, defensive responses that are more or less paraphrased as "You simply cannot get it because you're not brilliant enough to understand what and/or how I'm saying it" you'd be very frustrated with yourself as I am when I read this thread. It's all very condescending and dismissive. You're convinced you're right but you either refuse or unable to provide proof.

Formal peer review will not help you prove this "theory". That's not what peer review is for. Peer review is there to try to prove you wrong and viciously attack the theory. Peer review is not "lazyweb". You need to provide the proof and then defend it from attack with those proofs, and they need to be concrete and objectively verifiable and reproducible. There are no proofs here that can be tested, verified or reproduced.

I'm not heavily into math outside of armchair abstraction but I'm fascinated with theoretical number theory, complexity theory and P/NP, and your post and this thread make about as much sense as if someone had put Hofstadter's Godel, Escher and Bach into a blender and then made homemade paper out of it.

You may have a solid, provable idea but you're failing to either prove it or communicate it clearly, and this is exactly the sort of thing that people spend their entire lifetime trying to prove or communicate clearly, but as it is now there's no there, there.

Also - Jimmy Wales isn't known to be an expert in anything other than being in the right place at the right time to start Wikipedia by stealing/borrowing the idea from wikiwikiweb. He certainly is not a mathematician nor a scientist. I would also ask you to consider that you likely exchanged emails not with Mr. Wales himself but with an assistant working in his name.
posted by loquacious at 4:57 AM on April 30, 2010 [7 favorites]


Well hence the informal peer review, I see no better way for a non-academic such as myself to refine my ideas other than engaging those in various communities. I realize that there is a communication barrier and that is what I'm trying to resolve.

I'm not saying people don't understand it because they aren't brilliant enough to, I'm saying they don't understand it because they don't have the background to do so. I'm trying to provide that background the best as possible.

Actually I don't think it's difficult to grasp at all once one really understands the terminology. Now to really understand probably takes a bit more research, and I don't even pretend to understand the full consequences of what I have posed. However I am confident that I can defend the work.

Any suggestions on how I could communicate better please let me know and if there's a specific area you don't understand ask and I'll do my best to explain my reasoning behind it.

As far as Jimmy Wales goes, I've found most people actually respond to their emails themselves unless its in a very formal format.

And as far as mania goes, yes it often leads to really bad things but there's a moment of extreme clarity where everything you've been working on comes into extreme focus. All theory is as much art as it is science, and while art and music may be your medium applied theory is mine.
posted by jlind0 at 9:33 AM on April 30, 2010


However I am confident that I can defend the work.

Your interaction here proves exactly the opposite. You've answered questions with circular reasoning or obfuscation by creating your own definitions for terms that have otherwise clear meanings.

All theory is as much art as it is science, and while art and music may be your medium applied theory is mine.

This is a lie you're telling yourself. Research requires work to produce results. Proofs are rigorous exercises, not poetry. You have not done any work. You have listed some terms (most of which you've either made up, or used definitions completely divorced from reality) on a website, some of which are loosely collected in what could charitably be called sentences.

You're discussing a math problem without any math at all. You have not explained how your terms are related to algorithms or their runtimes. For example, you've said "AI is probably NP-Complete." without any further explanation.

You've stated: "...according to this theory {ideal organization theory} artificial intelligence simply emerges from highly optimized organizational structure."

As far as I can tell, you've simply made up IOT without explaining what it is. Even if you had explained what it is, you haven't done anything to explain what you mean by "highly optimized" above.
posted by odinsdream at 10:48 AM on April 30, 2010 [2 favorites]


Ideal organization theory describes a classification system that structures the interaction of self-determining actors. It states that the actors need to be formally structured only to the extent that the internal interaction is finite, in other words there can't be too many cooks in the kitchen.

This is what oligopoly really means, a small set of actors, which each actor being able to understand the position of the other actors in the set and their behavior. If the set is too large by definition an actor cannot use the other actors behavior as part of their decision process.

Membership in the oligopoly is subject to terms and conditions, so an actor will only remain as part of that set in-so-much their actions are accepted by the definition of the structure. Even if the member actor is not intelligent, their self-determining actions (be it by randomness or an algorithm) will determine whether their membership is valid.

These organized sets then need to be able to interact within a free marketplace to participate in processes that are non-finite, that is the nature of the interaction makes oligopical interaction inefficient.

When these strategies are combined the resultant marketplace is optimal, that is it is the best combination of efficient and fair Arthur Okun's Equality and Efficiency: The Big Tradeoff for a great discussion of what optimal means in this context. He argues that Efficiency and Equality are on a Laffer Curve.

The resultant marketplace will be artificially intelligent to an outside observer as the structure results in information flow that would elicit a more intelligent response than any one actor or set of actors. Hence artificial intelligence simply emerges from highly optimal organizational structures.

Making more sense yet?
posted by jlind0 at 11:08 AM on April 30, 2010


Oh and if it isn't clear already, I'm using MetaFilter to collect my thoughts, so thank you for all those active on this board who have been instrumental in helping me do so. I hope that all of you are enjoying and learning from this discussion as much as I am.
posted by jlind0 at 11:28 AM on April 30, 2010


Making more sense yet?

No.

I hope that all of you are enjoying and learning from this discussion as much as I am.

No.

There's no there, there. There's nothing for us to learn because there's nothing there to learn. It reads like badly written hand-waving science fiction.

You're conflating (in the truest sense of the word) disparate concepts and terms, misusing those terms and otherwise misinterpreting or willfully misusing known terms and otherwise making shit up.

Are you familiar with Time Cube and Gene Ray? Because that's what you sound like.

MeFi Projects is not a place for you to hash out your ideas, collect notes or attempt to formalize that which likely cannot be formalized. It's not a classroom, a laboratory or a forum for debate.

"Applied theory" is apparently not your medium because you're failing to formalize anything. You're waving your hands a lot and attempting to conjure. Phrases like "simply emerges" don't really work. Theories don't actually work like this. Theories aren't just supposition or wild ideas. Theories require substance and rigidity.

Out of masochistic curiosity - what is your stance on hard AI vs. soft AI?
posted by loquacious at 3:47 PM on April 30, 2010


I would say the big difference that separates me from Gene Ray is I'm willing to take feedback and apply it to my theory, as I've demonstrated consistently throughout this threat. I strongly believe that my equation is correct, however I also acknowledge it is more than possible that it is not. The only way for me to determine one way or another is to engage in discussion with other knowledgeable and interested parties.

MeFi projects may not have been intended for this use, but it seems to working fairly well as it has enabled me to dramatically clean up and clarify my work. Just because something hasn't been used in such a way before doesn't mean it can't. As this thread has generated a ton of discussion points I'm sure if MeFi leadership felt this was inappropriate I would have been at least contacted by now.

IOT would suggest that hard-AI would emerge out of soft-AI: that individual actors with soft-intelligence, when placed in an optimized organization structure, would together be capable of solving the problems classified as AI-Complete.
posted by jlind0 at 4:42 PM on April 30, 2010


I meant thread there and not threat, blast you spell check.
posted by jlind0 at 5:45 PM on April 30, 2010


Just to repeat my earlier comment, you need a proper reference style. So rather than referring to Nash as "e.g. Nash's work" you should cite: Nash, 199X and then at the bottom list the paper or book that you are referring to, and or the equation or page number. Another term for this is 'author date' and a commonly used style is 'Harvard' (I mention this as you point out you are not an academic). I suggest not using wikipedia or web-references.

I find what you have written to be no more believable than someone telling me in a casual conversation about some important theory, and just referencing vague concepts. If it is not possible to reference the topics that are of fundamental importance to this work in such a way that they lend credibility to the argument, then I would immediately disregard it.

Secondly, I think the premise that people are unable to understand this because it is too complicated, to be insulting. As a graduate student, I interact with a lot of professors who are extremely clever, and occasionally they fail to understand something I describe. Usually that is because my description is not clear.

Finally, I agree with some of the commentators up-thread that this is not appropriate for 'Projects'. I don't think this is intentional on your part, but it is a poorly formed discussion/monologue about some ideas that are not very clear.
posted by a womble is an active kind of sloth at 7:43 AM on May 1, 2010


Well I think this poorly formed discussion has had value in enabling me to more clearly convey my ideas. That said I agree that it needs to be further clarified. I am not in a position to formalize this work with the kind of academic rigidly you are suggesting, I have neither the training nor the inclination to do so.

If there are specific terms or concepts I have no conveyed clearly or you disagree with I would be happy to defend my position, but the whole point of this structure is a way for me to get my ideas out there without the ridged academia demands.
posted by jlind0 at 12:40 PM on May 1, 2010


You lack a position to defend. You have not made any claims or assertions.
posted by odinsdream at 2:10 PM on May 1, 2010


but the whole point of this structure is a way for me to get my ideas out there without the ridged academia demands.

And yet you demand this rigidity from us and the scientific/academic community, even though you're unwilling to do it yourself.

Without that rigidity, you have no position to defend. What you have is utter nonsense. You might as well claim that the color "orange" is actually "rusty steam locomotive" or "shoe" - which is exactly what you're doing by haphazardly conflating disciplines and making shit up to force them to fit.

Intellectual rigor exists for a reason. To discard it and try to hand-wave it away is to throw the scientific method out the window and take a casual stroll back to alchemy and beyond to the dark ages.

If you think you actually have something that's this valuable and revolutionary then you need to back it up with a lot more study and formalization than this. I'm not saying that stumbling on a valid idea isn't possible - indeed, many of our scientific breakthroughs occur exactly this way - but none of them have ever gone anywhere without formal proof and a lot of hard work to get there. More often than not people spend their entire lives (and sometimes go mad) doing this kind of work. People went nuts working with infinity (and beyond), or trying to calculate Pi.

It's even worse when you keep coming back with what reads very strongly like condescension. You might not be aware of it, but you are and it's not helping. For example - when I brought up art and music you assume that that's all that I do and then claim to work in "applied theory", when your trade is software consulting. I, too, work in IT. Art and music is just one thing I do. I also do graphic design, photography, audio engineering and a whole lot of other stuff including writing. I'm a polymath. I do a little of everything. I'm interested in almost everything from cosmology to geography.

Am I then also an applied theorist? Hell no.

To assume and then imply that anyone here isn't getting it because we don't have the knowledge or background is insulting. No, I'm not a huge math nerd, but to counter your "I was doing differentials at 15" I was doing quadratics when I was 8. I actually kind of "invented" or "discovered" quadratics on my own because I wanted to be able to calculate parabolic arcs just for the purpose of writing a game that used parabolic trajectories on my old Apple 2. I didn't have a clue they even existed or how to do them. Yes, they were shoddy, circularly reasoned and inaccurate - but they worked well enough for the low pixel count games I was working on.

I don't bring this up to measure math-wangs or anything, because I suck at formal math, even if I don't suck at logic and abstract reasoning.

I bring it up because it's very likely someone has already explored the path you're looking at. These days it's very difficult to discover new, valid methods or realms of mathematics. It's just not likely that you're idea is actually new and unique, and it's more likely that it was already previously discarded because it's wrong.

And without that intellectual rigor and due diligence we will never know if you're right therefore you're wrong until you can prove otherwise. Conjecture is not and will never be proof.

Honestly, I don't even know what I'm doing here trying to point any of this out to you. I'm not trying to be mean or snarky, but the hubris involved is excruciating and a little embarrassing, so I'm likely just succumbing to "But, someone is wrong on the internet!"

Good luck. Work hard. Maybe in 10-20 years you'll prove me wrong and I can say "Wow, I was wrong about that" which I have no problem doing. But I'll believe it when I see it.

I'm certainly not here because I think your idea is even remotely plausible. I think I'm mainly here because your idea is so malformed yet vigorously defended that it's pushing my buttons, like I need to defend science from nonsense. Science doesn't need me to defend it, obviously, so with that I'll bow out and do my best to refrain from commenting further.
posted by loquacious at 3:39 PM on May 1, 2010 [5 favorites]


loquacious: a deep understanding of math probably is only going to help a bit here, the terms I use are primarily from economics and information theory. These terms have precise meaning in those fields and relate to various models with deep connections to mathematics. How could you possibly begin to understand what I'm talking about if you don't have a cursory background in those fields?

I'm doing my best to provide that, but that's not meant as an insult. It's like saying Newton would have been insulted by Cantor if he looked at Cantor's notation with no basis and not understood it. Well Newton was kind of a jackass so maybe he would have.

My thesis may look like plain English, but the point is I'm using words as concepts and to understand the thesis one needs to understand the concepts.

I am very confident that has not been attacked within economics as I discussed it in detail with a former professor of mine and he hadn't heard of this type of attack.

Does this need to be worked out further? YES. My goal was to peek some interest, generate discussion and use that to clarify my position.

Let me just say again, I'm not trying to insult anyone, I appreciate the candor and feedback of everyone on this chain. I contend that if I could provide proper background one would readily see that my thesis is indeed true, and I'm willing to work with anyone to clarify what I've wrote.
posted by jlind0 at 6:44 PM on May 1, 2010


Get more good sleep, dude.
posted by flabdablet at 10:59 PM on May 3, 2010 [1 favorite]


oquacious: a deep understanding of math probably is only going to help a bit here, the terms I use are primarily from economics and information theory. These terms have precise meaning in those fields and relate to various models with deep connections to mathematics. How could you possibly begin to understand what I'm talking about if you don't have a cursory background in those fields?

You sound like an ass here.

Have you read any papers in economics or information theory?

Whenever they use simple words which "have precise meaning" or which have "deep connections", they always, always provide a definition or a reference to a definition.

If you do that, then those words actually convey precise meaning and deep connections. If you do not, then you're just rambling.

For example, and we touched on this above, you've used the term "self-determining actors" but not defined it properly.

You said above "If the set is random, I would say by definition it is determining itself", but since you haven't given a definition, using the phrase "by definition" makes no sense.

In your page about IOT you say "This is what oligopoly really means, a small set of actors, which each actor being able to understand the position of the other actors in the set and their behavior."

How can Math.Random() "understand the position of the other actors". If it cannot, does that mean that it's not "self determining" for the purposes of this proof?

You say "these terms have precise meaning" but you're not even using the terms consistently.
posted by richb at 4:28 AM on May 4, 2010


I referenced using quotes, I know it's not accepted practice but I have neither the time nor inclination to take that any further right now.

Randomness = Determining Itself <>
Oligopoly means that the actors CAN use the positional information, obviously some actors would be incapable of doing that (even intelligent actors) correctly.

Anything else I can try to clear up?
posted by jlind0 at 4:06 PM on May 4, 2010


Randomness = Determining Itself : A definition of randomness (aparently mefi didn't like my syntax)
posted by jlind0 at 4:06 PM on May 4, 2010


The main trouble with what you've done here is that it doesn't mean anything to anybody except you.

This is because it's entirely expressed in terms belonging to your own private jargon. Nobody else attaches the same meanings to most of the terms you use as you do.

The consequence of this is that the entire intellectual structure becomes something which is intensely pleasing for you to contemplate, but remains inaccessible to anybody else.

In order to convince other people that your insight is worthwhile, which you will need to do if your aim is to share the pleasure you get from contemplating it, you will need to find ways to express it using terms with precise consensus meanings. Most people working in areas that seem likely to be related to yours find that mathematics is the only system within which the required degree of consensual conceptual precision is available.

This has both benefits and risks. The possible benefit is that if you have in fact understood something important, and you manage to communicate that understanding, good things may come of what you've discovered. The risk is that in the process of reformulating your insight in accessible terms, you will come to a realization that it is in fact either trivial or wrong.

So, whether you choose to do the further work necessary to unpack and explain your insight is up to you. But until you have done that work, dismissing people who fail to be interested in your stuff as intellectually beneath you is both unjustified and counterproductive.
posted by flabdablet at 6:01 PM on May 4, 2010 [8 favorites]


Randomness = Determining Itself

You are building meaning into something that does not have it. This statement is not true. Dice do not determine for themselves what number they will show when you roll them. Many millions of interactions of particles interact on the dice thrown until they're sufficiently at rest with regard to whatever reference surface a human has chosen. At that point, the face farthest from the reference surface is read by a human. The dice have nothing to do with it.

What you've done, though, is taken this literary shortcut of saying that randomness = determining and then further conflated it with things like "self determination" and gotten all the way to claiming that dice are intelligent.
posted by odinsdream at 7:50 PM on May 4, 2010


Also: anybody with a serious interest in communicating subtle ideas should have read The Tyranny of Words before attempting to do so. It remains every bit as fresh and relevant as when Stuart Chase wrote it in 1938.
posted by flabdablet at 10:09 PM on May 4, 2010 [2 favorites]


jlind0 wrote: How could you possibly begin to understand what I'm talking about if you don't have a cursory background in those fields?

Quoting richb for truth: You sound like an ass here.

Re-quoting myself here: To assume and then imply that anyone here isn't getting it because we don't have the knowledge or background is insulting.

I don't think I emphasized enough the fact that I'm very deeply interested in AI, TNT and the abstracts of math and I do have at least the cursory background you speak of.

If there was any one person reading your theory that could even attempt to get what you're trying to say by way of misappropriation and conflation I would have a pretty good crack at it, because I thrive on exactly these kinds of abstracts. I loved Godel, Escher and Bach. It's a lovely, dense tome filled to the brim with amazingly illustrative leaps of logic. I'm not saying I understand all of it, and I wouldn't trust my own interpretation of what the Eternal Golden Braid is beyond complexity theory's involvement with self-organizing, self-referring intelligences such as ourselves - IE, whatever the point to that maddening book actually is.

But what you're saying is all so much gibberish it makes me want to tear out my hair and eat it.
posted by loquacious at 12:05 AM on May 5, 2010


Anything else I can try to clear up?

lol, yes. Although I'm now fairly convinced that you can't and that this is just "mania" as you put it or as flabdablet puts it "it doesn't mean anything to anybody except you".

So you've said several times that Math.Random() is self determining and that it "CAN use the positional information" although it "would be incapable of doing that correctly".

So let's plug that example of an "actor" into your theory and see if "P=NP":

Your page says "demonstrating the [following] equation is true will not only demonstrate what real artificial intelligence is, but that the deep mathematical connections will prove that P=NP"

Where the equation is "Artificial Intelligence = Finite interaction is optimized through oligopical competition, whereas non-finite processes are optimized by the free market. Formal organizational group structure therefore must be oligopical, but their interaction must be free. The individual is a monopoly. Q.E.D."

So let's take the example of 5 dice being rolled. Do these self determining actors "optimise"? You seem to suggest that all self determining actors do. I asked above what you meant by "optimise" and you said: "So when I say classification of sets I refer to the definition of the structure of those sets, and optimizing that definition so that it is the structure."

I don't understand this. Am I to understand that rolling 5 dice results in them "optimising" their "definition" into a "structure" via the operation of some "oligopical" competiton?

And would that "opimisation" give insight into "what real artificial intelligence is"?



None of that makes any sense. Do you see why no-one here understands your private terminology?

Maybe if rolling 5 dice doesn't fit into your theory a concrete example that does fit would help.

Could you also show how each of the words in quotes above can be applied to your chosen example?
posted by richb at 3:24 AM on May 5, 2010


I realize it's a bit out there, but I am pretty confident it's not crack pot

Check.

I actually run a successful software consulting practice and haven’t been in school for some-time

Check.

I am merely attempting to promote what I feel is a valid and interesting approach with vast applications

Check.

I'm not completely sure what the ramifications of what I have posed are ... [specific topic] was not the goal of this work when I started but the connection seemed obvious

Check.

while I know exactly what it says I realize other people do not

Check.

I am not an academic, this is more or less a hobby, and I do not have the training nor inclination to turn this into accepted scholarly format myself ... I am merely attempting to promote what I feel is a valid and interesting approach with vast applications

Check.

This is deep stuff, and there are not many people with the backgrounds required to immediately grasp the substance of my attack, however if you sit back and attempt to understand what that equation really says I think that you will agree it is obviously true

Check.

Honestly the key components of this theory was written when I was in what’s considered a manic state, which I have been out of for about a month. So even I am trying to figure out the details of what I meant, I don't know if you've ever had a maddening moment of clarity

Check and BINGO.

How the story goes from this point forward:

The next few months were spent trying to dig value out of the massive flood of insight that occurred during the manic episode. It was very, very difficult to believe that something so crystalline, so obvious, so apparent, so strong, so huge - did not in fact connect meaningfully with anything at all outside itself; surely there must have been a nugget in there?

After a while, when I finally became able to accept that nothing I had discovered about the fundamental nature of reality was in fact startling, novel or correct, I reacted by distrusting my ability to understand anything abstract, and became quite afraid that attempting to do so would push me back into mania. I wouldn't have minded that were I the only person affected, but of course none of us are the only person affected by psychosis. Being mad can be tremendously pleasurable but it puts a hell of a load on those who care most about us.

I remained afraid to think about Certain Things for roughly two years. It felt like I was keeping my mind in a plaster cast, and it got very depressing. Zoloft helped limit the depth of that for a couple months, but I eventually just got burned out and had to give up making a living off project-based work.

Ten years on, I consider myself very very lucky to have experienced all of that, especially since I have not suffered lasting damage as a result. In fact I have learned nothing at all about the fundamental nature of reality per se, but I've learned a hell of a lot about my own reality-apprehending machinery. The key thing is that the moment of clarity is a feeling - an emotional experience - and that after enjoying the hell out of one, the right thing to do is reality check it. So I'm glad to see that this is, in fact, what you're doing here.

You're one month out. Take it easy. Get good sleep. Eat well. Walk a lot. Be as kind to yourself as you possibly can. Keep your friends and loved ones close.

Maybe one day you and I will meet and drink beer.
posted by flabdablet at 4:17 AM on May 5, 2010 [21 favorites]


as I've demonstrated consistently throughout this threat.

Don't be frightened! We're really trying to help.
posted by Obscure Reference at 4:43 AM on May 5, 2010


You are using the language of Information Theory and Cybernetics in seemingly solipsistic ways. You are mix levels (jumping between different systems) in a haphazard way, and your ideas are based on concepts that you don't seem to have stable (or in some cases even internally coherent) definitions for.

You display a good intuitive grasp of Cybernetic systems theory, but you make a huge number of classic beginner mistakes in the domain.
posted by idiopath at 9:47 AM on May 5, 2010 [1 favorite]


*you are mixing
posted by idiopath at 9:48 AM on May 5, 2010


flabdablet: I've been down this road before, you stare into infinity and infinity stares right back at you. After looking at this with "sane" vision I know that this is something significant, that I have framed artificial intelligence in a new framework.

Even the concept that oligopical competition has a role to play in marketplaces is highly controversial, many of my closest economist professors are deeply troubled by that notion (most of whom are of the Austrian School, which I whole heartedly believed in when I started this). However even they admit that I might be right.

This isn't my first time dealing with mania, I was hospitalized about three years ago due to a severe manic episode. However I think that extreme mania can be a very powerful tool, it is hard to channel right, you have to both recognize that your judgment is not perfectly sound, maintain a high standard of logic and manage your interaction to limit damage to those around you.

Next time I'm probably going to isolate myself from those around me, because through their well-intentioned concern they began to agitate me and it took me from an extreme manic state to a psychotic state that landed me in the hospital. Also I now have advanced directives in place that will hopefully keep me out of the hospital.

During my mania I actually was reality checking what I was working on with those in my inner circle, they didn't understand it, thought I'd lost my mind and then took steps to "make me sane" again. Was I acting weird? For sure, but not one decision I made has had any kind of negative impact, and it seems the reverse is true.

The biggest decisions I've made during my life have always put me at odd's with those closest to me: refusing to get confirmed, cutting back at work in high school so I could play Ultima Online where I was making more money, leaving high school without a diploma and going to college, leaving college to work full time, joining the military, starting my own consulting company, etc. When I have clarity of vision I really see things at a different level, and I can see what effect my actions have at a broad scale and I can make moves that I can't even begin to explain the logic behind them.

"Many sacrifices don't require concrete calculation at all. It is sufficient to only glance at the arising position to convince us all that the sacrifice is correct." - Mikhail Tal (if you don't know who that is I suggest you look him up)

I was able to design a system for developing an organizational marketplace that acts intelligently, and now I'm building my company around that theory. People are rallying around the idea, I have 20 people signed on as members who are actively working with me to found the company.

So if that’s madness, I can live with that.

Obscure: I really did mean thread there (I corrected it in a subsequent post)
posted by jlind0 at 4:06 PM on May 5, 2010


flabdablet (top comment): I don't see how I'm dismissing anyone, I have been open and available to everyone on this thread the best I possibly could be.

I have no better way to connect this to a wider audience without engaging them. Again this seems to be working as it is much clearer today than it was when this thread started. If there are any specific areas I can clear up please let me know.

Oh and I'll check out the book.

odinsdream: self-determination != intelligence. The set of dice would not be self determining, but a set that made decisions based on the role of the dice would be, regardless of if the set rolled the dice themselves.

True randomness actually only comes from phsyical systems, you may be aware that there are several products which generate true random data, the cheapest of which are probably basically giger counters reading a small amount of radiation, the levels of which at any given second are purly random.

loquacious: I really need to read that book, I had it on my bookshelf at one point, I don't know what happened to it. Again if you have a cursory understanding of the terms I'm using I'd expect you be able to ask me a direct question and not just assuming there's nothing here because the format is not up to your standards.

I can't improve the format unless people ask me questions that I can answer. I mean I don't think you'd continuing to write on this thread if you didn't have interest.
posted by jlind0 at 4:23 PM on May 5, 2010


richb: So as I mentioned above the dice themselves are clearly not the actors, since they cannot role themselves. Now an actor can have their decisions governed by the role of the dice.

So let’s say there's an actor making decisions randomly (you can have it be by dice roles if that’s what you like), the actors membership in groups would likely be very limited as membership in a group is generally defined by the actor's behavior. However in organizations there may be groups that want to rely on randomness, for example determining the order in a bidding system.

Even semi-intelligent actor's would likely not form organizations with random actors, and they would know this based on their oligopical interaction with that actor. Whereas in a large free market environment it is irrelevant who the actors are so if there are actors making random decisions it may even be optimal, there's a very good book called "Random Walk Down Wall Street" by Burton Malkiel which has a thesis that a "bare ass monkey throwing darts at the wall street journal has just a good a chance of picking a wining stock as a seasoned wall street analyst."

Competition is essentially making agreements, so the actors must also be able to define a contract to make those agreements or they are unable to interact with the other actors. A random actor for practical purposes must therefore be constructed so that it can randomly communicate with the other actors in ways that they would understand.

Organizations of actors would live and die by their ability to interact with the free market and among each other; the organization's reach would be a factor of its effectiveness. I contend that theoretically with the right initial organization structure purely random actors would sort themselves out in a way to achieve defined organizational goals. However the number of actors would need to be incomprehensibly large, probably something like the number of electrons in the moon.
To practically achieve AI the actors would require some level of intelligence, even a rudimentary equation that can attempt to evaluate the decisions it can make would be useful. The best actors will always be able to pass the Turing Test.

You may want to check out the other end of this, which is a high level look at a system using these principals assuming NI actors. http://www.lind-i.com/Docs/GoalStruct.pdf
posted by jlind0 at 5:44 PM on May 5, 2010


Again if you have a cursory understanding of the terms I'm using I'd expect you be able to ask me a direct question and not just assuming there's nothing here because the format is not up to your standards.

I did, and you replied with a weird sort of non-answer that leads me to believe you don't understand the difference between hard and soft AI. I wasn't asking how your ideas for this conjecture related to hard or soft AI, I was asking what your opinion was on hard VS soft AI in general, because they're rather incompatible ideas and it's a sort of gauge or litmus test of where people stand with AI.

I'm going to affirm what flabdablet said - a lot of us are alarmed because we've been down this road and we sympathize.

I would feel like an asshole if I somehow squashed what may be a valid idea (when properly formalized) and I would also feel like more of an asshole if my rather strident responses triggered you into doing something regretful.

Anyway, it's not the formatting we're having a problem with. It's the actual ideas. They don't make any sense to us no matter how much you insist that they make sense. This isn't a good sign.
posted by loquacious at 6:36 PM on May 5, 2010


I can summarize my feelings about all this a lot more succinctly, too.

I know you don't know me. I'm just some guy on the net. But ask around - I'm a genuine weirdo. Like gonzo strength weird. I try really hard not to be too weird and yet I'm still a USDA inspected grade A freak. Maybe not openly Crispin Glover strength weird but possibly weirder than he is on the inside.

If someone like me is scratching their head and deeply pondering "What the fuck?" - buddy, you need to watch out. The void doesn't just stare back - it consumes voraciously. It'll eat your head.
posted by loquacious at 6:43 PM on May 5, 2010 [5 favorites]


I think that extreme mania can be a very powerful tool, it is hard to channel right, you have to both recognize that your judgment is not perfectly sound, maintain a high standard of logic and manage your interaction to limit damage to those around you.

The trouble with mania is not that it renders judgment "not perfectly sound"; it scrambles it altogether while leaving the maniac feeling as if the damage is minor. If you're manic, you're simply not in a position to act in your own best interests.

A high standard of logic won't help you if the objects you're applying it to don't mean anything.

And one of the things that maniacs are particularly poor at is evaluating the effects of their own behaviors on the people around them.

Next time I'm probably going to isolate myself from those around me

I strongly recommend that you don't do that.

because through their well-intentioned concern they began to agitate me and it took me from an extreme manic state to a psychotic state that landed me in the hospital.

I think you'll find, on sober reflection several years on, that the progression from extreme mania to psychosis was all your own work, and that you would almost certainly have found a way to become sufficiently agitated even if those close to you had left you to your own devices. Of course I could be wrong, and of course I could simply be projecting, but I have met enough maniacs and enough people who care for maniacs to form an opinion that this is the usual way of these things.

Also I now have advanced directives in place that will hopefully keep me out of the hospital.

If you weird people out sufficiently, none of those will do you much good. If you attempt to avoid weirding other people out by cutting contact with them, then those who care about you are simply going to get extra concerned.

For the sake of those who care about you, please act to keep your mania in check. I know it feels wonderful, but it's really very, very selfish, and if you become addicted to it then your life and that of those close to you will turn to shit.

During my mania I actually was reality checking what I was working on with those in my inner circle, they didn't understand it, thought I'd lost my mind and then took steps to "make me sane" again.

Sounds very familiar. I have at times experienced a similar distortion in the meaning of the words "reality check".

Was I acting weird? For sure, but not one decision I made has had any kind of negative impact, and it seems the reverse is true.

How do the people who cared enough about you to attempt to restore your sanity feel about that? Have you asked them? Thanked them for their concern? Apologized? All are necessary.

The biggest decisions I've made during my life have always put me at odd's with those closest to me: refusing to get confirmed, cutting back at work in high school so I could play Ultima Online where I was making more money, leaving high school without a diploma and going to college, leaving college to work full time, joining the military, starting my own consulting company, etc.

Fair enough.

When I have clarity of vision I really see things at a different level, and I can see what effect my actions have at a broad scale and I can make moves that I can't even begin to explain the logic behind them.

By what process do you distinguish the outcome of moves made on the basis of a manic "broad scale vision" from moves made completely at random?

The reason you see things at a different level while manic, it seems to me, is that mania is all about ignoring or destroying information. The maniac inhabits a world increasingly untroubled by details, and this allows for a great many conceptual connections to be made - extremely pleasurably, and at an enormous rate - between things that would normally be considered quite at odds with one another.

"Many sacrifices don't require concrete calculation at all. It is sufficient to only glance at the arising position to convince us all that the sacrifice is correct." - Mikhail Tal

Intuition is very powerful stuff, if it's properly primed. The right way to prime your intuition is careful study. The "us all" that Tal refers to here is a bunch of people who have spent many, many hours playing good chess against skilled opponents.

If you spend a significant amount of your life operating in mania, you get to exercise your intuition a lot, but you don't give it much to work with. The risk is that you'll end up completely overdrawing your intuitive bank account: making so many connections between things that don't really justify those connections that your intuitions will become indistinguishable from noise, except that they will remain pleasurable. At that point, you are basically screwed.

Personally, I value my intuition enough that I try to avoid abusing it in such a fashion.

(if you don't know who that is I suggest you look him up)

You might want to remind yourself that you're on Metafilter now, not the YouTube comments section.

If you play chess, one useful reality check you could try is jumping on FICS while feeling manic, and seeing if your rating ends up higher or lower after twenty games. If you're feeling like a winner but playing like a fool, that's a Clue.

I was able to design a system for developing an organizational marketplace that acts intelligently, and now I'm building my company around that theory. People are rallying around the idea, I have 20 people signed on as members who are actively working with me to found the company.

In the early stages of my own mania, I was really very charismatic and very persuasive. Because I have always been bright, and my friends and colleagues know that I have always been bright, I was able to make a fairly persuasive case that the stuff I was understanding might in fact not be the random-sounding gibberish they at first took it to be.

It took quite some time, and some truly bizarre behavior on my part, for the last of them to give up and declare me officially bananas. I'm fairly sure that if my circle of acquaintances at the time had included a few venture capitalists, that I would have been able to extract quite large amounts of funds before tipping over the edge; I'm equally sure that if I had chosen to hone my skills at staying just this side of the edge, I could also have started up an org and kept it going for a while. But I'm now absolutely convinced that had I in fact done so, the outcome would have been loss and waste for all involved.

So I wish you well with your new venture, but have absolutely no interest in investing in it myself. I will be interested to see how it's shaped up in about five years.

So if that’s madness, I can live with that.

I truly hope that your friends and loved ones can live with it too. And I strongly, strongly encourage you to keep working on making your ideas comprehensible to those who have a background in related fields, and to listen to those people when they suggest that perhaps what you have is not quite all that.

If your ideas are good, then they will survive being transformed into understandable language. If they don't survive that process, or if after many attempts you find yourself simply unable to transform them into understandable language: please, please learn to take yourself and your manic vision less seriously. Your life will almost certainly be better for that.
posted by flabdablet at 9:30 PM on May 5, 2010 [3 favorites]


loquacious: I framed my answer in terms of my thesis since that's how I think about artificial intelligence now, it's impossible for me to disconnect my theories on artificial intelligence from the thesis since they are one in the same.

I consider myself a total eccentric, although I function at a high enough level where most people are willing to ignore my oddness since I produce great results. People, especially those closest to me, don't let me get away with what I used to, ten years ago when I was in college I would have been able to work in a manic state and left on my own until I came "back to reality".

I'm not too concerned when people don't understand this because of the nature of it, every professor I know so far had varying degrees of "I can't tell if this is madness or brilliance."
posted by jlind0 at 8:04 AM on May 6, 2010


My vote's on madness, obviously.
posted by flabdablet at 8:10 AM on May 6, 2010


flabdablet: I think history has shown that when I'm manic, not when I'm psychotic, I'm still acting in my own best interests. I'm able to make plays in life (not so much in chess, I've never been anything but a solid chess player) that result in better position for everyone.

When I was manic this time I was testing every thought I had against my understanding of the reality of the situation, I recognized my logic was flawed and made that a component of my decision making. I was running major decisions by those closest to me, my mistake was running them by people who had physical access to me (which is actually a small number) and my parents (who have always assumed I was insane when I didn't listen to them, and now that I have a diagnosis that’s just further proof).

Three years ago, the first time I ended up in the hospital, I was not applying that kind of rigor. I always had implicitly trusted my mind so when I thought something I assumed it was true, and didn't attempt to prove it to myself.

I had very little to apologize for, and I did thank my friends for their concern (both during and after the episode), in fact most of them were apologizing to me. And where I had something to apologize for I did.

I don't want to get into it but several of them, through their concern, exploited the trust I had with them to push me from the manic to delusional state so that I would be tricked to go into the hospital voluntarily. Once in the admit process of the hospital I got agitated some more and had a complete psychotic break. So no, I'm pretty confident very little of that was my own doing and if people had just let me alone to come down myself it would have been perfectly fine.

I take care of those around me really well, and always have. I recognize that I can be a bit much to deal with someone times, but I take massive risks all the time and i have a very good track record of being proven right over time.

I'm working very hard to get feedback on my ideas from any source I can, and it really seems like its coming together better than when it started, and hopefully that will continue to be the case.
posted by jlind0 at 8:22 AM on May 6, 2010


On further reflection, and after exchanging mefi mails with idiopath, I'm reminded that whether or not your ideas are derived from mania really has no bearing on their usefulness; they should stand or fall on their own merits. So I'm sorry for derailing the discussion down a mental health side track. Let's get back to the part where you work on rendering them comprehensible to we mere mortals and we'll all go from there.
posted by flabdablet at 8:27 AM on May 6, 2010


We're all mere mortals on this one :)
posted by jlind0 at 8:37 AM on May 6, 2010


In that case, your first job needs to be making the stuff comprehensible to you.
posted by flabdablet at 7:09 PM on May 6, 2010


It is, but I'm not going to pretend I know the conseuqences of everything that I've posed, as they are indeed vast.
posted by jlind0 at 9:58 AM on May 7, 2010


So vast.
posted by chunking express at 10:16 AM on May 7, 2010


I'm puzzled by your position that AI is NP. You say that the fact that we haven't produced true AI yet "...would indicate that the problem of artificial intelligence is in NP as a deterministic solution would likely have been found by now". But it usually isn't hard to find deterministic solutions to NP problems - it's just hard to find efficient solutions to them. The fact that we don't have working AI now says to me that it's not in NP at all (perhaps EXPTIME or perhaps it has to be more carefully defined before it even fits in the hierarchy at all).

I have issues with the definitions you have (or haven't) given, but those can wait (and others have already tackled some of them). I'm just not buying AI = NP in the first place. The rest of the argument doesn't matter until that gets resolved.
posted by It's Never Lurgi at 11:57 AM on May 7, 2010


You're coming off as condescending because in your responses, you use your own terms. The polite thing to do would be to use the terms of the person you are responding to.

If you truly cannot express your ideas without using your own terms, you need to define your terms using existing terms familiar to your audience--in this case the audience is another commenter, so you should use terms that the commenter used; if their terms are insufficient, you can try to use terms that a Metafilter user would know the meaning of.

If none of the language used by a Metafilter user is sufficient for your purposes, then you're in the wrong place. Go somewhere where there are people who have words for the concepts you're interested in. You claim the terms are from economics and information theory; there are web forums about those subjects, and perhaps they will understand you.

If they, like others in this thread, say you're using the terms wrong... you should listen.
posted by LogicalDash at 5:40 PM on May 7, 2010


It is, but I'm not going to pretend I know the conseuqences of everything that I've posed, as they are indeed vast.

If you don't understand the consequences, you can't possibly know how vast they are. This is another example of how you tend to contradict the definitions of the words you're using in your application of those very words. It is not a good sign for your theory.
posted by abc123xyzinfinity at 7:16 PM on May 10, 2010


Hate to beat a dead horse, but for what it's worth, for several months after my own manic episode I remained quite convinced that some of the stuff I'd "understood" while manic also had potentially vast consequences. It took quite a long time for that to wear off, and the appropriate restoration of humility involved was quite painfully ego-bruising.

For instance, one of the most promising of my Big Ideas was this: what if I built a potentially global filesystem that generated a 512-bit secure hash of each 4096-byte data block it was asked to write, and used those hashes in roughly the same way that a conventional filesystem used disk cluster or block numbers? There would need to be a key->value storage layer underneath the thing to make it feasible.

The main things that makes this proposal interesting are:

* Actual physical location of the data would cease to matter at all, as long as the key->value layer was able to find the value for a given stored key somewhere - perhaps in a local disk cache, perhaps on a network.

* Every write to the filesystem would involve regenerating a hierarchy of pointer blocks, all the way back to the superblock (since there's no way to overwrite a block whose address is determined solely by its content; closest you can get is redefining which blocks you care about) which would make the thing inherently self-snapshotting.

* The system is inherently a single-instance storage scheme - file copies would be very quick indeed, since they would devolve to the equivalent of a hard-link operation on a conventional filesystem.

* Networked implementations of the key->value layer could make good use of multicasting; any client waiting for filesystem block #X could generate hashes for each blocks it saw in the multicast stream while waiting its turn to submit a request, and skip the request if block #X just happened to arrive before then.

* Data spoofing is pretty much impossible. If you have a superblock, and you manage to retrieve a file by starting from that superblock, you can be absolutely certain that what you retrieved is exactly what was written.

Now, these are certainly interesting features, and the proposed filesystem would probably be quite a lot of fun to implement and play with, but for months after coming down from the acute manic phase I remained filled with the burning conviction that I had Solved The Internet Storage Problem and invented the One Universal File System To Rule Them All, which is clearly not the case.

Now I've gone and looked at your web site, and its vibe (AHA! I've Figured Out What's Gone Wrong Everywhere and have Worked Out How To Fix It All!) is really very familiar. But in fact all you've done is presented a bunch of half-baked organizational ideas and expressed the hope that things would all be fine if only everybody thought as you did.

There's really very little point in inviting people to reality-check you if you're going to persist in the belief that none of us know what we're on about. At some point, it's going to pay you to stop reacting to all this cold water by shouting about how waterproof you are, and start fixing your roof instead.
posted by flabdablet at 10:01 PM on May 10, 2010 [1 favorite]




Huh, I came here from flabdablet's linkin his AskMe post
I actually state that P = NP if Artificial Intelligence is NP-Complete.
There's no reason to think AI needs to be NP-Complete. My view is that AI is obviously possible, and therefore not NP complete. Usually modern machine learning involves lots of linear algebra, most of which can be done in O(n3) time for each step, with not too many steps in most cases. Linear algebra is also easy to parallelize across lots of machines.

And the thing about P = NP is that any proof of NP will be obviously true to anyone who understand what it means. You won't need any debate or discussion, in fact you will be able to verify it with a computer even.

(And actually, I also think biological intelligence is polynomial in computational complexity)

---

Anyway look: If P = NP, that means you have written a program that can solve 3SAT in polynomial time. If you have that program, then you have proven P = NP. If you don't have that program, then you haven't proven it. On the other hand proving P != NP would not be as straightforward.

Again, write a program that can solve 3SAT in polynomial time, and then you will have proven P = NP. Until then, you have not.

Finally, P=NP wouldn't matter much if the polynomial exponent was high. Like if you proved that you could solve an O(np) problem on an NP machine in O(np2+10) or even just O(np+10) or something on a P machine, it wouldn't change anything.
posted by delmoi at 4:39 AM on May 18, 2010


delmoi: P=NP wouldn't matter much if the polynomial exponent was high

A high exponent might mean that you couldn't crack SSL tomorrow, but it would still be a hugely important mathematical result if it were proven for any finite exponent.

Its Never Lurgi: I'm puzzled by your position that AI is NP. ... The fact that we don't have working AI now says to me that it's not in NP at all (perhaps EXPTIME or perhaps it has to be more carefully defined before it even fits in the hierarchy at all).

delmoi: There's no reason to think AI needs to be NP-Complete.

These are good points. You haven't defined "AI" in such a way that it can be classified into a complexity class.

The complexity classes (e.g. P, NP) refer to decision problems over a variable number of inputs. What specific devision problem do you mean by "AI"? You've made vague comments about "classifying sets" which I asked about above, but you didn't really give an answer.

I get the distinct impression that by "AI" you're talking about some kind of HAL9000 movie computer who can play chess and talk to you about baseball, and when you say that we haven't "solved" it yet, I take it you're talking about the Turing test?

Saying "if AI is NP-Complete" just doesn't make any sense with those definitions, and you won't or can't give any others.
posted by richb at 11:05 AM on May 20, 2010


What specific devision problem do you mean by "AI"?

Sorry, that should read "What specific decision problem do you mean by "AI""?
posted by richb at 11:07 AM on May 20, 2010


See also http://www.metafilter.com/94553/Complex-matters-for-the-millenium

(His proof does mention decision problems and you'll notice that each section starts with definitions of the terms being used ;-)
posted by richb at 2:46 AM on August 9, 2010


« Older Firescaping ...   |   Comment Macht Frei... Newer »


You are not currently logged in. Log in or create a new account to post comments.