Re: [singularity] pattern definition
On Mon, May 5, 2008 at 7:50 PM, [EMAIL PROTECTED] wrote: Hello I am writing a literature review on AGI and I am mentioning the definition of pattern as explained by Ben in his work. A pattern is a representation of an object on a simpler scale. For example, a pattern in a drawing of a mathematical curve could be a program that can compute the curve from a formula (Looks et al. 2004). My supervisor told me that she doesn?t see how this can be simpler than the actual drawing. Any other definition I could use in the same context to explain to a non-technical audience? Hi, See the Wikipedia article on Kolmogorov complexity: http://en.wikipedia.org/wiki/Kolmogorov_complexity It has a nice example with a Mandelbrot set: you get this arbitrarily detailed image from a one-line algorithm. -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=101816851-9a120b Powered by Listbox: http://www.listbox.com
Re: About the Nine Misunderstandings post [WAS Re: [singularity] I'm just not sure how well...]
On Fri, Apr 11, 2008 at 10:50 PM, Matt Mahoney [EMAIL PROTECTED] wrote: If the problem is so simple, why don't you just solve it? http://www.securitystats.com/ http://en.wikipedia.org/wiki/Storm_botnet There is a trend toward using (narrow) AI for security. It seems to be one of its biggest applications. Unfortunately, the knowledge needed to secure computers is almost exactly the same kind of knowledge needed to attack them. Matt, this issue was already raised a couple of times. It's a technical problem that can be solved perfectly, but isn't in practice, because it's too costly. Formal verification, specifically aided by languages with rich type systems that can express proofs of correctness for complex properties, can give you perfectly safe systems. It's just very difficult to specify all the details. These AIs for network security that you are talking about are a cost-effective hack that happens to work sometimes. It's not a low-budget vision of future super-hacks. -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=98631122-712fa4 Powered by Listbox: http://www.listbox.com
Re: About the Nine Misunderstandings post [WAS Re: [singularity] I'm just not sure how well...]
On Sat, Apr 12, 2008 at 12:34 AM, Matt Mahoney [EMAIL PROTECTED] wrote: Actually it cannot be solved even theoretically. A formal specification of a program is itself a program. It is undecidable whether two programs are equivalent. (It is equivalent to the halting problem). Converting natural language to a formal specification is AI-hard, or perhaps harder, because people can't get it right either. If we could write software without bugs, we would solve a big part of the security problem. You just evoked a Halting Problem Fallacy. You can't check if given arbitrary program terminates, but you can write a program that provably terminates. Likewise you can't check correctness of an arbitrary program, but you can write a provably correct program. Yes, it is possible to write software without bugs, starting from a formal definition of what 'bug' is (for example, 'it must never crash', 'it must never directly leak sensitive data' or 'it must continue to be able perform function A properly' are OK). The cause for this whole security problem thing is that presently it's very hard to write provably safe programs, so almost no one is doing it. Functional programming research community is working on this problem, but I doubt there will ever be tools that will enable average Joe the programmer to meaningfully write verified code. Understanding natural language specification and converting it to code is what programming is about. I certainly didn't imply that 'programming is unnecessary, perfect secure code can just write itself'. It won't happen until we have AGI. These AIs for network security that you are talking about are a cost-effective hack that happens to work sometimes. It's not a low-budget vision of future super-hacks. Not at present because we don't have AI. I responded assuming that you were talking about the following sort of thing, and its presumed further development to higher levels and subtler rules: http://en.wikipedia.org/wiki/Intrusion-prevention_system Rate based IPS (RBIPS) are primarily intended to prevent Denial of Service and Distributed Denial of Service attacks. They work by monitoring and learning normal network behaviors. Through real-time traffic monitoring and comparison with stored statistics, RBIPS can identify abnormal rates for certain types of traffic e.g. TCP, UDP or ARP packets, connections per second, packets per connection, packets to specific ports etc. Attacks are detected when thresholds are exceeded. The thresholds are dynamically adjusted based on time of day, day of the week etc., drawing on stored traffic statistics. We rely on humans to find vulnerabilities in software. We would like for machines to do this automatically. Unfortunately such machines would also be useful to hackers. Such double-edged tools already exist. For example, tools like SATAN, NESSES, and NMAP can quickly test a system by probing it to look for thousands of known or published vulnerabilities. Attackers use the same tools to break into systems. www.virustotal.com allows you to upload a file and scan it with 32 different virus detectors. This is a useful tool for virus writers who want to make sure their programs evade detection. I suggest it will be very difficult to develop any security tool that you could keep out of the hands of the bad guys. All automatic tools already work from formally specified things that they are trying to find in the system. If you write the code so that these things are not there, and ascertain it by using automatic verification based on e.g. sufficiently rich type system, these tools won't find anything either. And yes, if you don't make code clearer, it's a very difficult problem to find vulnerabilities in it, and the smarter/more resourceful you are, the more you'll be able to find. -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=98631122-712fa4 Powered by Listbox: http://www.listbox.com
Re: [singularity] Vista/AGI
On Mon, Mar 17, 2008 at 4:48 PM, John G. Rose [EMAIL PROTECTED] wrote: I think though that particular proof of concepts may not need more than a few people. Putting it all together would require more than a few. Then the resources needed to make it interact with various systems in the world would make the number of people needed grow exponentially. Then what's the point? We have this problem with existing software already, and it's precisely the magic bullet of AGI that should allow free lunch of automatic interfacing with real-world issues... -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=98631122-712fa4 Powered by Listbox: http://www.listbox.com
Re: [singularity] Re: Revised version of Jaron Lanier's thought experiment.
On Sun, Feb 24, 2008 at 2:51 AM, Stathis Papaioannou [EMAIL PROTECTED] wrote: Consider Arithmetical Functionalism: the theory that a calculation is multiply realisable, in any device that has the right functional organisation. But this might mean that somewhere in the vastness of the universe, a calculation such as 2 + 2 = 4 might be being implemented purely by chance: in the causal relationship between atoms in an interstellar gas cloud, for example. This is clearly ridiculous, so *either* Arithmetical Functionalism is false *or* it is impossible that a calculation will be implemented accidentally. Right? I feel a little uncomfortable when people say things like 'because Occam's razor is true' or 'otherwise computationalism is false' or 'consciousness doesn't exist'. As these notions are usually quite loaded and ambiguous, and main issues with them may revolve around the question of what they actually mean, it's far from clear what is being asserted when they are declared to be 'true' or 'false'. Does 2+2=4 make a sound when there is no one around? -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=96140713-a54b2b Powered by Listbox: http://www.listbox.com
Re: [singularity] Re: Revised version of Jaron Lanier's thought experiment.
On Sun, Feb 24, 2008 at 4:06 AM, Stathis Papaioannou [EMAIL PROTECTED] wrote: On 24/02/2008, Vladimir Nesov [EMAIL PROTECTED] wrote: Does 2+2=4 make a sound when there is no one around? Yes, but it is of no consequence since no one can hear it. However, if we believe that computation can result in consciousness, then by definition there *is* someone to hear it: itself. But it's still of no 'consequence', no? -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=96140713-a54b2b Powered by Listbox: http://www.listbox.com
Re: Infinitely Unlikely Coincidences [WAS Re: [singularity] AI critique by Jaron Lanier]
On Feb 20, 2008 6:13 AM, Stathis Papaioannou [EMAIL PROTECTED] wrote: The possibility of mind uploading to computers strictly depends on functionalism being true; if it isn't then you may as well shoot yourself in the head as undergo a destructive upload. Functionalism (invented, and later repudiated, by Hilary Putnam) is philosophy of mind if anything is philosophy of mind, and the majority of cognitive scientists are functionalists. Are you still happy asserting that it's all bunk? Philosophy is in most cases very inefficient, hence wasteful. It puts very much into building its theoretical constructions, few of which are useful for understainding reality. It might be fun for those who like this kind of thing, but it is a bad tool. -- Vladimir Nesov [EMAIL PROTECTED] --- singularity Archives: http://www.listbox.com/member/archive/11983/=now RSS Feed: http://www.listbox.com/member/archive/rss/11983/ Modify Your Subscription: http://www.listbox.com/member/?member_id=4007604id_secret=96140713-a54b2b Powered by Listbox: http://www.listbox.com
Re: [singularity] Multi-Multi-....-Multiverse
On Jan 29, 2008 11:49 AM, Ben Goertzel [EMAIL PROTECTED] wrote: OK, but why can't they all be dumped in a single 'normal' multiverse? If traveling between them is accommodated by 'decisions', there is a finite number of them for any given time, so it shouldn't pose structural problems. The whacko, speculative SF hypothesis is that lateral movement btw Yverses is conducted according to ordinary laws of physics, whereas vertical movement btw Yverses is conducted via extraphysical psychic actions ;-)' What differentiates psychic actions from non-psychic so that they can't be considered ordinary? If I can do both, why aren't they both equally ordinary to me (and everyone else)?.. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=91036630-4898ad
Re: [singularity] Multi-Multi-....-Multiverse
On Jan 28, 2008 2:17 PM, Ben Goertzel [EMAIL PROTECTED] wrote: Can you define what you mean by decision more precisely, please? That's difficult, I don't have it formalized. Something like application of knowledge about the world, it's likely to end up an intelligence-definition-complete problem... -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=90505077-ab77a2
Re: [singularity] Multi-Multi-....-Multiverse
On Jan 27, 2008 9:29 PM, John K Clark [EMAIL PROTECTED] wrote: Ben Goertzel [EMAIL PROTECTED] we can think about a multi-multiverse, i.e. a collection of multiverses, with a certain probability distribution over them. A probability distribution of what? Exactly. It needs stressing that probability is a tool for decision-making and it has no semantics when no decision enters the picture. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=90386232-2d2891
Re: [singularity] Wrong focus?
On Jan 26, 2008 8:57 PM, Bryan Bishop [EMAIL PROTECTED] wrote: On Saturday 26 January 2008, Mike Tintner wrote: Why does discussion never (unless I've missed something - in which case apologies) focus on the more realistic future threats/possibilities - future artificial species as opposed to future computer simulations? This is bias in the community. The majority of the information from SingInst, for example, focuses on digital ai and not the potential ai that we can get from the bio sector, like through synbio and gengineering and tissue engineering of neurons and brains. I guess limitation of biological substrate are too strict, and there is not much to hope for from this side. Maybe we'd be able to construct a genetically engineered scientist with huge brain that will develop AGI, before cracking this problem ourselves ;-) -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=90259355-e04be3
Re: [singularity] EidolonTLP
On Jan 23, 2008 1:06 AM, Daniel Allen [EMAIL PROTECTED] wrote: It is entertaining. I love the greeting -- Greetings, little people -- and the graphics along with the ambient and almost haunting background music. But speech is so boring that it must be a GOFAI... -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=88693780-508c2f
Re: [singularity] The Extropian Creed by Ben
On Jan 20, 2008 3:06 PM, Mike Tintner [EMAIL PROTECTED] wrote: Sorry if you've all read this: http://www.goertzel.org/benzine/extropians.htm But I found it a v. well written sympathetic critique of extropianism highly recommend it. What do people think of its call for a humanist transhumanism? Thanks Mike for highlighting this informative essay. I think that first and foremost we must not embrace mystery. Ben argues against oversimplifying, but are we honest in adding in details that we don't sufficiently understand? For each irresponsibly added detail brings us away from reality. Preferring a fabulous wrong impression over a simple speckle of truth is not virtuous. Humans don't have stable morality. They learn, they go mad. What is it about evolutionary preprogrammed reinforcers that makes them exceptional before other random concoctions? They have a good position of power, many people obey them. If one argues for personal moral freedom, it's not about enforcing freedom on others, it's about liberating oneself from influence of others. There is no reason in choosing a moral stance if you don't know what effect it will have. Seek understanding if you want to hold back an existing moral plague, including the part you embody yourself. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=87886040-d08b59
Re: [singularity] I feel your pain
Isn't empathy a failure to distinguish between yourself and the other? Deficiency, not strength? On 11/6/07, Don Detrich [EMAIL PROTECTED] wrote: Will an AGI with no bio-heritage be able to feel our pain, have empathy? If not, will that make it less conscious and more dangerous? http://www.salon.com/news/feature/2007/11/05/mirror_neurons/ Don Detrich This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?; -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=61462586-3970db
Re: [singularity] Uploaded p-zombies
Monday, September 10, 2007, Matt Mahoney wrote: MM --- Vladimir Nesov [EMAIL PROTECTED] wrote: I intentionally don't want to exactly define what S is as it describes vaguely-defined 'subjective experience generator'. I instead leave it at description level. MM If you can't define what subjective experience is, then how do you know it MM exists? It exists in the same sense anything else exists. All objective world theories can be regarded as invariants of subjective experience. Objective world theories are portable between agents of the same world. MM If it does exist, then is it a property of the computation, or does MM it depend on the physical implementation of the computer? How do you test for MM it? It certainly corresponds to physical implementation (brain) and it is a property of relations between its parts (atoms/neurons). If it's a property of computation is what I'm trying to find out. MM Do you claim that the human brain cannot be emulated by a Turing machine? Functionally equivalent implementation can be built. But physical world doesn't know system's design to find that certain relations between certain states in emulating computer correspond to relations between neurons in original brain. Main thesis is that subjective experience is a property of physical implementation, not of arbitrary mathematical model of that implementations. Two can be the same if that mathematical model is derivable purely from physical implementation, though. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=40046899-85db61
Re: [singularity] Uploaded p-zombies
Sunday, September 9, 2007, Matt Mahoney wrote: MM Also, Chalmers argues that a machine copy of your brain must be conscious. MM But he has the same instinct to believe in consciousness as everyone else. My MM claim is broader: that either a machine can be conscious or that consciousness MM does not exist. While I'm not yet ready to continue my discussion on essentially the same topic with Stathis on SL4, let me define this problem here. Let's replace discussion of consciousness with more simple of 'subjective experience'. So, there is a host universe in which there's an implementation of mind (a brain or any other such thing) which we as a starting point assume to have this subjective experience. Subjective experience exists as relations in mind's implementation in host universe (or process of their modification in time). From this it supposedly follows that subjective experience exists only as that relation and if that relation is instantiated in different implementation, the same subjective experience should also exist. Let X be original implementation of mind (X defines state of the matter in host universe that comprises the 'brain'), and S be the system of relations implemented by X (the mind). There is a simple correspondence between X and S, let's say S=F(X). As brain can be slightly modified without significantly affecting the mind (additional assumption), F can also be modification-tolerant, that is for example if you replace in X some components of neurons by constructs with different chemistry which still implement the same functions, F(X) will not change significantly. Now, let Z be an implementation of uploaded X. That is Z can as well be some network of future PCs plus required software and data extracted from X. Now, how does Z correspond to S? There clearly is some correspondence that was used in construction of Z. For example, let there be a certain feature of S that can be observed on X (say, feature is D and it can be extracted by procedure R, D=R(S)=R(F(X))=(RF)(X), D can be for example a certain word that S is saying right now). Implementation Z comes with a function L that enables to extract D, that is D=L(Z), or L(Z)=R(S). Presence of implementation Z and feature-extractor L only allow the observation of features of S. But to say that Z implements S in the sense defined above for X, there should be a correspondence S=F'(Z). This correspondence F' supposedly exists, but it is not implemented in any way, so there is nothing that makes it more appropriate for Z than other arbitrary correspondence F'' which results in a different mind F''(L)=S'S. F' is not a near-equivalence as F was. One can't say that implementation of uploaded mind simulates the same mind or even in any way similar mind. It observes behavious of original mind using feature-extractors and so is functionally equivalent, but it doesn't exclusively provides an implementation for the same subjective experience. So, here is a difference: simplicity of correspondence F between implementation and the mind. We know from experience that modifications which leave F a simple correspondence don't destroy subjective experience. But complex correspondences make it impossible to distinguish between possible subjective experiences implementation simulates, as correspondence function itself isn't implemented along with simulation. As a final paradoxical example, if implementation Z is nothing, that is it comprises no matter and information ar all, there still is a correspondence function F(Z)=S which supposedly asserts that Z is X's upload. There can even be a feature extractor (which will have to implement functional simulation of S) that works on an empty Z. What is the difference from subjective experience simulation point of view between this empty Z and a proper upload implementation? -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=39991599-a151a9
Re: [singularity] Uploaded p-zombies
Monday, September 10, 2007, Matt Mahoney wrote: MM Perhaps I misunderstand, but to make your argument more precise: MM X is an implementation of a mind, a Turing machine. No. The whole argument is about why turing machine-like implementation of uploaded brain doesn't seem to do the trick. X is an original meaty brain, say collection of locations of atoms, which is as starting point assumed to implement subjective experience. Point of discussion is to show that it's not as obvious as it seems from the first look that uploaded X will also implement subjective experience. I thought that too initially, before arriving at this argument. Now Z IS a turing machine-like thing. S is something that is an essense of subjective experience-generating structure, and in case of a brain it closely corresponds to its physical structure. It need not be the simpliest representation possible. MM S is the function computed by X, i.e. a canonical form of X, the smallest or MM first Turing machine in an enumeration of all machines equivalent to X. By MM equivalent, I mean that X(w) = S(w) for all input strings w in A* over some MM alphabet A. MM Define F: F(X) = S (canonical form of X), for all X. F is not computable, but MM that is not important for this discussion. I intentionally don't want to exactly define what S is as it describes vaguely-defined 'subjective experience generator'. I instead leave it at description level. MM An upload, Z, of X is defined as any Turing machine such that F(Z) = F(X) = S, MM i.e. Z and X are equivalent. F'(Z)=F(X)=S, Z is an upload of X. They are equivalent given F and F', and F' doesn't in itself correspond to Z, for example there is a F'' just as good, which results in a different F''(Z), so it's a big question if Z and X are equivalent, or rather Z is as equivalent to F'(Z)=S as it is equivalent to F''(Z)=S'S. Also bugfix in my previous message: it's F''(Z)=S'S, not F''(L)=S'S MM Then the paradox in your last example cannot exist because F(nothing) != S, That F was a different F, which is by definition equal to S. Say, F*(nothing)=S (by definition). I omitted details about time and I/O for simplicity, but they can be factored in with minor changes. MM The other problem is that you have not defined subjective experience. MM Presumably this is the input to a consciousness? If consciousness does not MM exist, then how can subjective experience exist? There is only input to the MM Turing machine that may or may not affect the output. A reasonable definition MM of subjective experience would be the subset of inputs that affect the output. It's more like I look for proper definition of subjective experience based on presented thought experiment. So I substitute unknown definition by associated symbol (S) and describe its properties. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=40011790-2fef3d
Re: [singularity] Current absence of activism
Sunday, August 19, 2007, Joshua Fox wrote: JF (but please, only well-confirmed reports rather than JF supposition). Maybe more general question that provides information about the same problem is [why don't you donate/work on it/spread the word, other than due to inability to do so]. So, framed that way: I'm not particularly interested in things other than AGI speculated around the topic, since I don't believe they can provide uploading/immortality in my lifetime. Even given nanotechnology with all perks singularitarians give it, control/engineering is going to be too complex to implement these things in resonable time. Race with existential risks doesn't help. By contrast, AGI is going to be relatively simple engineering project given a workable design. AGI doesn't work well with public awareness; only reason to spread the word on this topic is fundraising, for which there first should be identified a target. There is currently no framework for research framed as AGI development, no reasonable grants and institutions. Few people who hint on having a clue about what they are doing intentionally refrain from providing technical details, so one really couldn't tell. This secretive approach is also being rationalized as exclusively correct one. Since non of them has destroyed the world yet, I suppose they overdramatize importance of their findings, but at the same time stall incremental progress. Only thing that's left is to work on the problem yourself, but it falls out of this topic, since very few people have at the same time sufficient background and intuitive inclination telling problem just might be workable, and they are working on it anyway. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=33571207-85e887
Re: Symbol grounding (was [singularity] ESSAY: Why care about artificial intelligence?)
Thursday, July 12, 2007, Cenny Wenner wrote: CW There are certainly difficulties with natural language but I do not CW see how these empirical and practical difficulties can be called CW tautologies of languages in general. Problem here is not in certain deficiency of natural language or some mystical way goals always keep being misinterpreted. Problem is that if you want to formulate complex goal/value, it usually intrinsically employs informal everyday concepts, and these concepts exist within the system you want to instruct only as results of complex process of perception. If you want guarantees on overall result, you must include properties of this perceptual process in specification as well. It draws practical line between goals we can build systems to reliably achieve and those we can't. There's obviously a technical problem of building a system powerful enough to be able to recognize these concepts, but it doesn't help in verification of whether these cancepts are really recognized in those situations we want them to or not. Problem is twofold: you should be able to guarantee properties of this perceptual process (so it can't be an arbitrary emergent one), and you should be able to figure out the essense of your own perception of these concepts. Middle ground is to make simplier formulation of goals not involving too much hairy perception. Meta ground is to include human perception itself in the loop, thus tasking the system to figure out perceptual details of humans as a subtask. -- Vladimir Nesovmailto:[EMAIL PROTECTED] - This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=4007604id_secret=20587889-3be28f