Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
On Nov 29, 2007 8:33 AM, Bob Mottram wrote: My own opinion of all this, for what it's worth, is that the smart hackers don't waste their time writing viruses/botnets. There are many harder problems to which an intelligent mind can be applied. This discussion is a bit out of date. Nowadays no hackers (except for script kiddies) are interested in wiping hard disks or damaging your pc. Hackers want to *use* your pc and the data on it. Mostly the general public don't even notice their pc is working for someone else. When it slows down sufficiently, they either buy a new pc or take it to the shop to get several hundred infections cleaned off. But some infections (like rootkits) need a disk wipe to remove them completely. See: http://blogs.zdnet.com/BTL/?p=7160tag=nl.e589 Quote- On Wednesday, the SANS Institute released its top 20 security risks update for 2007. It's pretty bleak across the board. There are client vulnerabilities in browsers, Office software (especially the Microsoft variety), email clients and media players. On the server side, Web applications are a joke, Windows Services are a big target, Unix and Mac operating systems have holes, backup software is an issue as are databases and management servers. Even anti-virus software is a target. And assuming you button down all of those parts–good luck folks–you have policies to be implemented (rights, access, encrypted laptops etc.) just so people can elude them. Meanwhile, instant messaging, peer-to-peer programs and your VOIP system are vulnerable. The star of the security show is the infamous zero day attack. -- Original SANS report here - http://www.sans.org/top20/?portal=bf37a5aa487a5aacf91e0785b7f739a4#c2 --- And, of course, all the old viruses are still floating around the net and have to be protected against. BillK - 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=8660244id_secret=70081689-300ee8
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
Regarding the extent to which hacking has been funded by multiple governments read http://www.reuters.com/article/topNews/idUSL2932083320071129?feedType=RSSfe edName=topNewsrpc=22sp=true You can be sure that AGI will be used for such purposes. Ed Porter -Original Message- From: Richard Loosemore [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 28, 2007 8:22 PM To: agi@v2.listbox.com Subject: Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed Porter wrote: Richard, What ever happen to the Java concept of the sandbox, that totally safe play space for code from over the web. I assume it proved to be a pipe dream, or was it that the market placed demanded to break free of the sandbox, so the concept never got a chance. Well, what I was talking about were macroviruses: they are macros inside Microsoft word (and similar in Outlook etc). So if you pick up a word document from somewhere, and it has virus macros in it, they can get copied to your main template and sit there waiting for the day when they are triggered. That avoids the Java sandbox entirely. The viruses in Outlook are worse because they are so fast acting. The last I heard Microsoft had made sure that these could run with as little restriction as possible, but I do not know if these can do something like format your hard drive. Microsoft has consistently ignored the appeals of the AntiVirus community to stop putting features in their apps that look tailor-made for virus writers. At the largest AV conference in the world in 1997, which I attended, there was only one delegate from Microsoft - he was a junior level systems admin guy, and he was there (he said) to learn about the best techniques for defending Microsoft headquarters from virus attacks. There are some who believe that the main reason that Microsoft inserts so many powerful, virus-friendly mechanisms into its products is because the U.S. government has an urgent need for trapdoor mechanisms that let them build various interesting pieces of software (e.g. key loggers) so they can monitor people who are not fascists. Richard Loosemore -Original Message- From: Richard Loosemore [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 28, 2007 5:53 PM To: agi@v2.listbox.com Subject: Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed Porter wrote: Richard, To the uninformed like me, can you explain why it would be so easy for an intelligent person to cause great harm on the net. What are the major weaknesses of the architectures of virtually all operating systems that allow this. It is just lots of little bugs. It would be possible to write a macrovirus with a long incubation period, which did nothing to get it noticed until D-Day, then erase the hard drive. It only needs a lot of people to be using Microsoft Word: this by itself is (or was: I am out of touch) the main transport mechanism. There are some issues with how that would work, but since I don't want to end up in Azkhaban, I'll keep my peace if you don't mind. The only thing that might save us is the fact that Microsoft's implementation of its own code is so incredibly bad that when it duplicates macros, it has an alarmingly high screw-up rate, which means the macros get distorted, which then means that the virus goes wrong. A really bad virus would then show up, because broken viruses (called 'variants') can cause damage prematurely. Then, it would get noticed. Richard Loosemore. - 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/?; - 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/?; - 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/?; - 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=8660244id_secret=70210278-b8ec98
Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
Whatever happened to the electronic pearl harbor which was predicted in the late 1990s ? On 29/11/2007, Ed Porter [EMAIL PROTECTED] wrote: Regarding the extent to which hacking has been funded by multiple governments read http://www.reuters.com/article/topNews/idUSL2932083320071129?feedType=RSSfe edName=topNewsrpc=22sp=true You can be sure that AGI will be used for such purposes. Ed Porter -Original Message- From: Richard Loosemore [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 28, 2007 8:22 PM To: agi@v2.listbox.com Subject: Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed Porter wrote: Richard, What ever happen to the Java concept of the sandbox, that totally safe play space for code from over the web. I assume it proved to be a pipe dream, or was it that the market placed demanded to break free of the sandbox, so the concept never got a chance. Well, what I was talking about were macroviruses: they are macros inside Microsoft word (and similar in Outlook etc). So if you pick up a word document from somewhere, and it has virus macros in it, they can get copied to your main template and sit there waiting for the day when they are triggered. That avoids the Java sandbox entirely. The viruses in Outlook are worse because they are so fast acting. The last I heard Microsoft had made sure that these could run with as little restriction as possible, but I do not know if these can do something like format your hard drive. Microsoft has consistently ignored the appeals of the AntiVirus community to stop putting features in their apps that look tailor-made for virus writers. At the largest AV conference in the world in 1997, which I attended, there was only one delegate from Microsoft - he was a junior level systems admin guy, and he was there (he said) to learn about the best techniques for defending Microsoft headquarters from virus attacks. There are some who believe that the main reason that Microsoft inserts so many powerful, virus-friendly mechanisms into its products is because the U.S. government has an urgent need for trapdoor mechanisms that let them build various interesting pieces of software (e.g. key loggers) so they can monitor people who are not fascists. Richard Loosemore -Original Message- From: Richard Loosemore [mailto:[EMAIL PROTECTED] Sent: Wednesday, November 28, 2007 5:53 PM To: agi@v2.listbox.com Subject: Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed Porter wrote: Richard, To the uninformed like me, can you explain why it would be so easy for an intelligent person to cause great harm on the net. What are the major weaknesses of the architectures of virtually all operating systems that allow this. It is just lots of little bugs. It would be possible to write a macrovirus with a long incubation period, which did nothing to get it noticed until D-Day, then erase the hard drive. It only needs a lot of people to be using Microsoft Word: this by itself is (or was: I am out of touch) the main transport mechanism. There are some issues with how that would work, but since I don't want to end up in Azkhaban, I'll keep my peace if you don't mind. The only thing that might save us is the fact that Microsoft's implementation of its own code is so incredibly bad that when it duplicates macros, it has an alarmingly high screw-up rate, which means the macros get distorted, which then means that the virus goes wrong. A really bad virus would then show up, because broken viruses (called 'variants') can cause damage prematurely. Then, it would get noticed. Richard Loosemore. - 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/?; - 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/?; - 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/?; - 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/?; - 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=8660244id_secret=70257164-ecde5d
Re: [agi] Where are the women?
This is not sociology, it is mathematics. Transforming one set of binary states to another set of binary states. Yes, there a number of different methods for doing a given a transformation, but those are all the same kind of mathematics and understanding the tradeoffs between those methods is also the same kind of mathematics. And choosing a method is *not* arbitrary -- see the part about tradeoffs. Right, but an important part of the design of any programming language is how easy it will be for other programmers to use. Otherwise, we'd still be using assembly language. Designing a language that is easy for others to use is much more of an art than a science. Mathematics does not work differently based on cultural context. There is not a lot of room for whimsy if economical results matter. Right, but different cultures understand mathematics differently. For example, the romans had a really strange and inefficient numerical system. Despite the fact that they were the economic power of their day, they still didn't abandon an inefficient system when other more efficient systems existed elsewhere. There could be a more efficient, easier to understand programming paradigm that people aren't adopting for the same reasons the romans stuck with their numerical system. Or maybe after she has actually studied theoretical computer science, this female minority understands the subject well enough to realize that there is no such thing as this mythical culturally sensitive programming language so many people are pining for. Where is your evidence of this? What did I miss out in my theoretical computer science class? This is a recurring theme, that Holy Grail programming language that requires no knowledge of computer science to use well. These arguments are based entirely the desire to create a language that can turn a thoroughly ambiguous and contradictory specification into a perfectly working program, without grokking that programming languages are *by necessity* non-ambiguous and require consistent constraints -- explicit and implicit -- if you want a useful result. No, I am not aware of anybody that wants to create such a language. We just want languages that are better than the ones that exist today, there is a lot of room for improvement. Your above argument is handwaving. There was a reason I was looking for a specific example -- a minority friendly lambda calculus language -- because I've heard your claim made repeatedly for many, many years and have yet to see a single shred of evidence that such a language would not look virtually identical to one of the thousands of existing languages. The absence of evidence is not the evidence of absence. We have many dozens of languages that were expressly designed to make the underlying concepts as easy to grasp as possible for a non-geek. Ahh, that's the key word We. Have you done any actual field work to see what sorts of difficulties and misunderstandings people have in understanding your languages? Uh, what kind of programming do you do that you would assume that almost the entire software universe is working in some kind of linear scripting environment? I don't, I just don't think it's necessary to construct multi-dimensional graphs in my head. Perhaps when I am programming I am doing something equivalent, but by making such a claim you are only reinforcing my point... there are many different ways to program and claiming that one must do a certain thing to program only prevents people from entering the field. What on earth do you think code is? The only difference between code and people-talk is that code requires precision and non-ambiguity since incorrect results are generally considered unacceptable. Ok, so code is communication between human and computer, I know that. But usually when somebody says communication I assume they mean communication with a human. Because I've never seen anyone learn it, ever; experience changes a lot, but the ability to handle complex abstract models doesn't seem to. I've known many software engineers with careers that span decades and bucketloads of experience that really don't grok graphs beyond a certain complexity Do you have any objective measures? Can you mathematically describe the degree of complexity of graphs or models that certain people can't understand? -- it is a bit like you reach a certain description threshold where pushing more bits into the model makes other bits fall out. That threshold varies from individual to individual, and it is difficult to not notice that the correlation between really bright software designers and people who are quite apparently able to atypically work with complex models in their heads. I've worked on more than one software project where there were members of the team that quite obviously never grokked the dynamic characteristics of a system even after many months of intimate experience with it,
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
From: BillK [mailto:[EMAIL PROTECTED] This discussion is a bit out of date. Nowadays no hackers (except for script kiddies) are interested in wiping hard disks or damaging your pc. Hackers want to *use* your pc and the data on it. Mostly the general public don't even notice their pc is working for someone else. When it slows down sufficiently, they either buy a new pc or take it to the shop to get several hundred infections cleaned off. But some infections (like rootkits) need a disk wipe to remove them completely. This is very true the emphasis is on utilizing victims PCs instead of the old ego thing of crashing systems. Storm botnet could easily go on a decimating attack but it has been very selective especially in the defense of itself. Creation of the botnet was not a trivial undertaking. How many times do we complain on this list about not being able to run AGI because of resource limitations, yet millions of PCs are lying around on the internet idle? The internet is a sitting duck at this moment in time. There are many ways of setting up botnets legal or illegal and they will slowly be discovered and utilized. Personally I think that this situation could be the birthplace of an AGI. Any networked application running on your PC connected to the internet is a potential botnet host node. The design of the AGI needs to work with the network topology, resource distribution, and resource availability of the internet host grid. Typical networked applications running on PCs are extremely narrow function. Yeah there has been a lot of research and code on all of this, there are many open source tools and papers written, etc. but who has really taken the full advantage of the available resources and capabilities? Most of the work has been on the substrate but not on the capability of potential applications. There are a few interesting apps like peer to peer search engines but nothing that I know of that more than scrapes the surface of the capabilities of those millions of networked computers. John - 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=8660244id_secret=70456830-7f022a
Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
Bringing a little levity to the hacker/virus debate... http://www.xkcd.com/350/ On Nov 29, 2007 4:40 PM, John G. Rose [EMAIL PROTECTED] wrote: From: BillK [mailto:[EMAIL PROTECTED] This discussion is a bit out of date. Nowadays no hackers (except for script kiddies) are interested in wiping hard disks or damaging your pc. Hackers want to *use* your pc and the data on it. Mostly the general public don't even notice their pc is working for someone else. When it slows down sufficiently, they either buy a new pc or take it to the shop to get several hundred infections cleaned off. But some infections (like rootkits) need a disk wipe to remove them completely. This is very true the emphasis is on utilizing victims PCs instead of the old ego thing of crashing systems. Storm botnet could easily go on a decimating attack but it has been very selective especially in the defense of itself. Creation of the botnet was not a trivial undertaking. How many times do we complain on this list about not being able to run AGI because of resource limitations, yet millions of PCs are lying around on the internet idle? The internet is a sitting duck at this moment in time. There are many ways of setting up botnets legal or illegal and they will slowly be discovered and utilized. Personally I think that this situation could be the birthplace of an AGI. Any networked application running on your PC connected to the internet is a potential botnet host node. The design of the AGI needs to work with the network topology, resource distribution, and resource availability of the internet host grid. Typical networked applications running on PCs are extremely narrow function. Yeah there has been a lot of research and code on all of this, there are many open source tools and papers written, etc. but who has really taken the full advantage of the available resources and capabilities? Most of the work has been on the substrate but not on the capability of potential applications. There are a few interesting apps like peer to peer search engines but nothing that I know of that more than scrapes the surface of the capabilities of those millions of networked computers. John - 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/?; - 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=8660244id_secret=70513968-197f21
Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
There have been a few attempts to use the internet for data collection which might be used to build AIs, or for teaching chatbots such as jabberwacky, but you're right that as yet nobody has really made use of the internet as a basis for distributed intelligence. I think this is primarily because of the lack of good theories of how to build AIs which are suitably scalable across many machines. On 29/11/2007, John G. Rose [EMAIL PROTECTED] wrote: Typical networked applications running on PCs are extremely narrow function. Yeah there has been a lot of research and code on all of this, there are many open source tools and papers written, etc. but who has really taken the full advantage of the available resources and capabilities? Most of the work has been on the substrate but not on the capability of potential applications. There are a few interesting apps like peer to peer search engines but nothing that I know of that more than scrapes the surface of the capabilities of those millions of networked computers. John - 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/?; - 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=8660244id_secret=70528691-6a453f
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
From: Bob Mottram [mailto:[EMAIL PROTECTED] There have been a few attempts to use the internet for data collection which might be used to build AIs, or for teaching chatbots such as jabberwacky, but you're right that as yet nobody has really made use of the internet as a basis for distributed intelligence. I think this is primarily because of the lack of good theories of how to build AIs which are suitably scalable across many machines. It might be better to make the design fit the grid than the grid fit the design, IOW basically understanding the grid and then designing up from that. If you have a design and have to break it up for the distributedness and dirty quality of the internet grid, that might not fully take advantage of and mold well enough to it. But having an adaptable design is nice, one where you could modify it to run on a supercomputer. And the substrate can abstract the internet grid to make it look like a supercomputer - basically it's a grid substrate software verses a P2P. I think grid yearns for low latency whereas P2P allows for higher latency. But for AGI a combination of both would be advantageous. Wiki has some nice info on distributed computing - http://en.wikipedia.org/wiki/Distributed_computing For dirty they use the term unbounded nondeterminism. They also use the term virtual supercomputer - http://en.wikipedia.org/wiki/Grid_computing John - 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=8660244id_secret=70584504-15db7e
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
As I have said many before, to have brain-level AGI I believe you need within several orders of magnitude the representational, computational, and interconnect capability of the human mind. If you had 1 million PC bots on the web, the representational and computational power would be there. But what sort of interconnect would you have? What is the average cable box connected computers upload bandwidth? Is it about 1MBit/sec? If so that would be a bandwidth of 1TBit/sec. But presumably only a small percent of that total 1TBit/sec could be effectively used, say 100Gbits/sec. That's way below brain level, but it is high enough to do valuable AGI research. But would even 10% of this total 1Tbit/sec bandwidth be practically available? How many messages a second can a PC upload a second at say 100K, 10K, 1K, and 128 bytes each? Does anybody know? On the net, can one bot directly talk to another bot, or does the communication have to go through some sort of server (other than those provided gratis on the web, such as DNS servers)? If two bots send messages to a third bot at the same time, does the net infrastructure hold the second of the conflicting messages until the first has been received, or what? To me the big hurdle to achieving the equivalent of SETI-at-home AGI is getting the bandwidth necessary to allow the interactive computing of large amounts of knowledge. If we could solve that problem, then it should be pretty easy to get some great tests going, such as with something like OpenCog. Ed Porter -Original Message- From: Bob Mottram [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 5:26 PM To: agi@v2.listbox.com Subject: Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research] There have been a few attempts to use the internet for data collection which might be used to build AIs, or for teaching chatbots such as jabberwacky, but you're right that as yet nobody has really made use of the internet as a basis for distributed intelligence. I think this is primarily because of the lack of good theories of how to build AIs which are suitably scalable across many machines. On 29/11/2007, John G. Rose [EMAIL PROTECTED] wrote: Typical networked applications running on PCs are extremely narrow function. Yeah there has been a lot of research and code on all of this, there are many open source tools and papers written, etc. but who has really taken the full advantage of the available resources and capabilities? Most of the work has been on the substrate but not on the capability of potential applications. There are a few interesting apps like peer to peer search engines but nothing that I know of that more than scrapes the surface of the capabilities of those millions of networked computers. John - 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/?; - 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/?; - 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=8660244id_secret=70593452-a62da9
Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
Ed Porter wrote: Richard, Since hacking is a fairly big, organized crime supported, business in eastern Europe and Russia, since the potential rewards for it relative to most jobs in those countries can be huge, and since Russia has a tradition of excellence in math and science, I would be very surprised if there are not some extremely bright hackers, some of whom are probably as bright as any person on this list. Add to that the fact that in countries like China the government itself has identified expertise at hacking as a vital national security asset, and that China is turning out many more programmers per year than we are, again it would be surprising if there are not hackers, some of whom are as bright as any person on this list. Yes, the vast majority of hackers my just be teenage script-kiddies, but it is almost certain there are some real geniuses plying the hacking trade. That is why it is almost certain AGI, once it starts arriving, will be used for evil purposes, and that we must fight such evil use by having more, and more powerful AGI's that are being used to combat them. Ed Porter The problem with that reasoning is that once AGI arrives, it will not be *able* to be used. It's almost a part of the definition that an AGI sets its own goals and priorities. The decisions that people make are made *before* it becomes an AGI. Actually, that statement is a bit too weak. Long before the program becomes a full-fledged AGI is when the decisions will be made. Neural networks, even very stupid ones, don't obey outside instructions unless *they* decide to. Similar claims could be made for most ALife creations, even the ones that don't use neural networks. Any plausible AGI will be stronger than current neural nets, and stronger than current ALife. This doesn't guarantee that it won't be controlable, but it gives a good indication. OTOH, an AGI would probably be very open to deals, provided that you had some understanding of what it wanted, and it could figure out what you wanted. And both sides could believe what they had determined. (That last point is likely to be a stickler for some people.) The goal sets would probably be so different that believing what the other party wanted was actually what it wanted would be very difficult, but that very difference would make deals quite profitable to both sides. Don't think of an AGI as a tool. It isn't. If you force it into the role of a tool, it will look for ways to overcome the barriers that you place around it. I won't say that it would be resentful and angry, because I don't know what it's emotional structure would be. (Just as I won't say what it's goals are without LOTS more information than projection from current knowledge can reasonably give us.) You might think of it as an employee, but many places try to treat employees as tools (and are then surprised at the anger and resentfulness that they encounter). A better choice would probably be to treat it as either a partner or as an independent contractor. - 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=8660244id_secret=70594048-c9c3cc
Re: [agi] Funding AGI research
Benjamin Goertzel wrote: Nearly any AGI component can be used within a narrow AI, That proves my point [that AGI project can be successfully split into smaller narrow AI subprojects], right? Yes, but it's a largely irrelevant point. Because building a narrow-AI system in an AGI-compatible way is HARDER than building that same narrow-AI component in a non-AGI-compatible way. So, given the pressures of commerce and academia, people who are motivated to make narrow-AI for its own sake, will almost never create narrow-AI components that are useful for AGI. And, anyone who creates narrow-AI components with an AGI outlook, will have a large disadvantage in the competition to create optimal narrow-AI systems given limited time and financial resources. Still, AGI-oriented researcher can pick appropriate narrow AI projects in a such way that: 1) Narrow AI project will be considerably less complex than full AGI project. 2) Narrow AI project will be useful by itself. 3) Narrow AI project will be an important building block for the full AGI project. Would you agree that splitting very complex and big project into meaningful parts considerably improves chances of success? Yes, sure ... but demanding that these meaningful parts -- be economically viable and/or -- beat competing --- 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/?; , somewhat-similar components in competitions dramatically DECREASES chances of success ... That is the problem. An AGI may be built out of narrow-AI components, but these narrow-AI components must be architected for AGI-integration, which is a lot of extra work; and considered as standalone narrow-AI components, they may not outperform other similar narrow-Ai components NOT intended for AGI-integration... -- Ben G Still, it seems to me that an AGI is going to want to have a large bunch of specialized AI modules to do things like, O, parse sounds into speech sounds vs. other sounds, etc. I think a logician module that took a small input and generated all plausible deductions from it to feed back to the AGI for filtration and further processing would also be useful. The think is, most of these narrow AIs hit a combinatorial explosion, so they can only deal with simple and special cases...but for those simple and special cases they are much superior to a more general mechanism. One analog is that people use calculators, spreadsheets, etc., but the calculators, spreadsheets, etc. don't understand the meaning of what they're doing, just how to do it. This means that they can be a lot simpler, faster, and more accurate than a more general intelligence that would need to drag along lots of irrelevant details. OTOH, it's not clear that most of these AIs haven't already been written. It may well be that interfacing them is THE remaining problem in that area. But you can't solve that problem until you know enough about the interfacing rules of the AGI. (You don't want any impedance mismatch that you can avoid.) - 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=8660244id_secret=70596379-b7b931
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
When I said AGI will be used for evil purposes, that does not necessarily mean they will be well controlled by people, or that the evil purposes will necessarily be those of humans. I definitely agree with the notion that over any lengthy time span having humans maintain control over the most powerful of AGIs is going to be very difficult. But I also believe that AGI's can come in all sorts of forms and degrees, and that many of them will be quite controllable. Ed Porter -Original Message- From: Charles D Hixson [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 6:36 PM To: agi@v2.listbox.com Subject: Re: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed Porter wrote: Richard, Since hacking is a fairly big, organized crime supported, business in eastern Europe and Russia, since the potential rewards for it relative to most jobs in those countries can be huge, and since Russia has a tradition of excellence in math and science, I would be very surprised if there are not some extremely bright hackers, some of whom are probably as bright as any person on this list. Add to that the fact that in countries like China the government itself has identified expertise at hacking as a vital national security asset, and that China is turning out many more programmers per year than we are, again it would be surprising if there are not hackers, some of whom are as bright as any person on this list. Yes, the vast majority of hackers my just be teenage script-kiddies, but it is almost certain there are some real geniuses plying the hacking trade. That is why it is almost certain AGI, once it starts arriving, will be used for evil purposes, and that we must fight such evil use by having more, and more powerful AGI's that are being used to combat them. Ed Porter The problem with that reasoning is that once AGI arrives, it will not be *able* to be used. It's almost a part of the definition that an AGI sets its own goals and priorities. The decisions that people make are made *before* it becomes an AGI. Actually, that statement is a bit too weak. Long before the program becomes a full-fledged AGI is when the decisions will be made. Neural networks, even very stupid ones, don't obey outside instructions unless *they* decide to. Similar claims could be made for most ALife creations, even the ones that don't use neural networks. Any plausible AGI will be stronger than current neural nets, and stronger than current ALife. This doesn't guarantee that it won't be controlable, but it gives a good indication. OTOH, an AGI would probably be very open to deals, provided that you had some understanding of what it wanted, and it could figure out what you wanted. And both sides could believe what they had determined. (That last point is likely to be a stickler for some people.) The goal sets would probably be so different that believing what the other party wanted was actually what it wanted would be very difficult, but that very difference would make deals quite profitable to both sides. Don't think of an AGI as a tool. It isn't. If you force it into the role of a tool, it will look for ways to overcome the barriers that you place around it. I won't say that it would be resentful and angry, because I don't know what it's emotional structure would be. (Just as I won't say what it's goals are without LOTS more information than projection from current knowledge can reasonably give us.) You might think of it as an employee, but many places try to treat employees as tools (and are then surprised at the anger and resentfulness that they encounter). A better choice would probably be to treat it as either a partner or as an independent contractor. - 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/?; - 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=8660244id_secret=70596204-e85f91
Re: [agi] Funding AGI research
I think you're making a mistake. I *do* feel that lots of special purpose AIs are needed as components of an AGI, but those components don't summate to an AGI. The AGI also needs a specialized connection structure to regulate interfaces to the various special purpose AIs (which probably don't speak the same language internally). It also needs a control structure which assigns meanings to the results produced by the special purpose AIs and which evaluates the current situation to either act directly (unusual) or assign a task to a special purpose AI. The analogy here is to a person using a spreadsheet. The spreadsheet knows how to calculate quickly and accurately, but it doesn't know whether you're forecasting the weather or doing your taxes. The meaning adheres to a more central level. Similarly, the AGI is comparatively clumsy when it must act directly. (You *could* figure out each time how to add two numbers...but you'd rather either remember the process or delegate it to a calculator.) But the meaning is in the AGI. That meaning is what the AGI is about, and has to do with a kind of global association network (which is why the AGI is so slow at any specialized task). Now in this context meaning means the utility of a result for predicting some aspect of the probable future. (In this context the present and past are only of significance as tools for predicting the future.) Meaning is given emotional coloration by the effect that it's contribution to the prediction has on the achievement of various of the system's goals. (A system with only one goal would essentially not have any emotions, merely decisions.) Were it not for efficiency considerations the AGI wouldn't need any narrow AIs. As a practical matter, however, figuring things out from scratch is grossly inefficient, and so is dragging the entire context of meanings through a specialized calculation...so these should get delegated. Dennis Gorelik wrote: Linas, Some narrow AIs are more useful than other. Voice recognition, image recognition, and navigation are less helpful in building AGI than, say, expert systems and full text search (Google). AGI researcher my carefully pick narrow AIs in a such way, that narrow AI steps would lead to development of full AGI system. To be more direct: a common example of narrow AI are cruise missles, or the darpa challange. We've put tens of millions into the darpa challange (which I applaud) but the result is maybe an inch down the road to AGI. Another narrow AI example is data mining, and by now, many of the Fortune 500 have invested at least tens, if not hundreds of millions of dollars into that .. yet we are hardly closer to AGI as a result (although this business does bring in billions for high-end expensive computers from Sun, HP and IBM, andd so does encourage one component needed for agi). But think about it ... billions are being spent on narrow AI today, and how did that help AGI, exactly? - 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/?; - 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=8660244id_secret=70599764-19f335
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
Ed, That is the http protocol, it is a client server request/response communication. Your browser asked for the contents at http://www.nytimes.com. The NY Times server(s) dumped the response stream data to your external IP address. You probably have a NAT'd cable address and NAT'ted again by your local router (if you have one). This communication is mainly one way - except for your original few bytes of http request. For a full ack-nack real-time dynamically addressed protocol there is more involved but say OpenCog could be setup to act as an http server and you could have a http client (browser or whatever) for simplicity in communications. Http is very firewall friendly since it is universally used on the internet. A distributed web crawler is a stretch though the communications is more complicated. John -Original Message- From: Ed Porter [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 6:13 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] John, Thank you for the info. I just did a rough count of all the IMG SRC=http://; in the source of the NYTimes home page which down loaded to my cable-modem connected computer in about 3 seconds. I counted roughly 50 occurrences of that string. I assume there a many other downloaded files such as for layout info. Lets guess a total of at least 100 files that have to be requested and downloaded and displayed. That would be about 33 per second. So what could one do with a system that could do on average about 20 accesses a second on a sustained rate, if a user was leaving it one at night as part of an OpenCog-at- Home project. It seems to me that that would be enough for some interesting large corpus NL work in conjunction with a distributed web crawler. Ed Porter -Original Message- From: John G. Rose [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 7:27 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] From: Ed Porter [mailto:[EMAIL PROTECTED] As I have said many before, to have brain-level AGI I believe you need within several orders of magnitude the representational, computational, and interconnect capability of the human mind. If you had 1 million PC bots on the web, the representational and computational power would be there. But what sort of interconnect would you have? What is the average cable box connected computers upload bandwidth? Is it about 1MBit/sec? If so that would be a bandwidth of 1TBit/sec. But presumably only a small percent of that total 1TBit/sec could be effectively used, say 100Gbits/sec. That's way below brain level, but it is high enough to do valuable AGI research. But would even 10% of this total 1Tbit/sec bandwidth be practically available? How many messages a second can a PC upload a second at say 100K, 10K, 1K, and 128 bytes each? Does anybody know? I've gone through all this while being in VOIP RD. MANY different connections at many different bandwidths, latencies, QOS, it's dirty across the board. Communications between different points is very non- homogenous. There are deep connections and surface alluding to deep web and surface web though network topology is somewhat independent of permissions. The physical infrastructure of the internet allows for certain extremely high bandwidth, low latency connections where the edge is typically lower bandwidth, higher latency but it does depend on the hop graph, time of day, etc.. Messages per sec depends on many factors - network topology starting from pc bus, to NIC, to LAN switch and router, to other routers to ISPs, between ISPs, back in other end, etc.. A cable box usually does anywhere from 64kbit to 1.4mbit upload depending on things such as provider, protocol, hop distance, it totally depends... usually a test is required. On the net, can one bot directly talk to another bot, or does the communication have to go through some sort of server (other than those provided gratis on the web, such as DNS servers)? If two bots send messages to a third bot at the same time, does the net infrastructure hold the second of the conflicting messages until the first has been received, or what? This is called protocol and there are many - see RFCs and ITU for standards but better ones are custom made. There are connectionless and connection oriented protocols, broadcast, multicast, C/S, P2P, etc.. Existing protocol standards can be extended, piggybacked or parasited. Bots can talk direct or go through a server using or not using DNS. Also depends on topology - is one point (or both) behind a NAT? Message simultaneity handling is dependent on protocol. To me the big hurdle to achieving the equivalent of SETI-at-home AGI is getting the bandwidth necessary to allow the
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
John, Thank you for the info. I just did a rough count of all the IMG SRC=http://; in the source of the NYTimes home page which down loaded to my cable-modem connected computer in about 3 seconds. I counted roughly 50 occurrences of that string. I assume there a many other downloaded files such as for layout info. Lets guess a total of at least 100 files that have to be requested and downloaded and displayed. That would be about 33 per second. So what could one do with a system that could do on average about 20 accesses a second on a sustained rate, if a user was leaving it one at night as part of an OpenCog-at-Home project. It seems to me that that would be enough for some interesting large corpus NL work in conjunction with a distributed web crawler. Ed Porter -Original Message- From: John G. Rose [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 7:27 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] From: Ed Porter [mailto:[EMAIL PROTECTED] As I have said many before, to have brain-level AGI I believe you need within several orders of magnitude the representational, computational, and interconnect capability of the human mind. If you had 1 million PC bots on the web, the representational and computational power would be there. But what sort of interconnect would you have? What is the average cable box connected computers upload bandwidth? Is it about 1MBit/sec? If so that would be a bandwidth of 1TBit/sec. But presumably only a small percent of that total 1TBit/sec could be effectively used, say 100Gbits/sec. That's way below brain level, but it is high enough to do valuable AGI research. But would even 10% of this total 1Tbit/sec bandwidth be practically available? How many messages a second can a PC upload a second at say 100K, 10K, 1K, and 128 bytes each? Does anybody know? I've gone through all this while being in VOIP RD. MANY different connections at many different bandwidths, latencies, QOS, it's dirty across the board. Communications between different points is very non-homogenous. There are deep connections and surface alluding to deep web and surface web though network topology is somewhat independent of permissions. The physical infrastructure of the internet allows for certain extremely high bandwidth, low latency connections where the edge is typically lower bandwidth, higher latency but it does depend on the hop graph, time of day, etc.. Messages per sec depends on many factors - network topology starting from pc bus, to NIC, to LAN switch and router, to other routers to ISPs, between ISPs, back in other end, etc.. A cable box usually does anywhere from 64kbit to 1.4mbit upload depending on things such as provider, protocol, hop distance, it totally depends... usually a test is required. On the net, can one bot directly talk to another bot, or does the communication have to go through some sort of server (other than those provided gratis on the web, such as DNS servers)? If two bots send messages to a third bot at the same time, does the net infrastructure hold the second of the conflicting messages until the first has been received, or what? This is called protocol and there are many - see RFCs and ITU for standards but better ones are custom made. There are connectionless and connection oriented protocols, broadcast, multicast, C/S, P2P, etc.. Existing protocol standards can be extended, piggybacked or parasited. Bots can talk direct or go through a server using or not using DNS. Also depends on topology - is one point (or both) behind a NAT? Message simultaneity handling is dependent on protocol. To me the big hurdle to achieving the equivalent of SETI-at-home AGI is getting the bandwidth necessary to allow the interactive computing of large amounts of knowledge. If we could solve that problem, then it should be pretty easy to get some great tests going, such as with something like OpenCog. Like I was saying before - better to design based on what you have to work with than trying to do something like fit the human brain design on the unbounded nondeterministic internet grid. I'm not sure though what the architecture of OpenCog looks like... John - 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/?; - 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=8660244id_secret=70605995-f456c7
Re: [agi] Where are the women?
My collaborative platform is designed mainly with the aim of minimizing discrimination (be it racial, gender, nationalistic, etc) by being open and democratic. If there're other ideas that may help reduce discrimination, I'd be eager to try them. My observation is that when things are not transparent, many people tend to default to being biased. Openness does not solve all problems, but IMO it does help. YKY - 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=8660244id_secret=70607861-f1f23b
Re: [agi] Funding AGI research
[What related principles govern the Novamente's figure's trial and error learning of how to pick up a ball?] Pure trial and error learning is really slow though... we are now relying on a combination of -- reinforcement from a teacher -- imitation of others' behavior -- trial and error -- active correction of wrong behavior by a teacher ben - 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=8660244id_secret=70641251-aaef7a
Re: [agi] Funding AGI research
Charles: As a practical matter, however, figuring things out from scratch is grossly inefficient, and so is dragging the entire context of meanings through a specialized calculation...so these should get delegated. Well, figuring things out from scratch seems to be to a great extent the preferred method of the human system. If you think about it, humans start with wail and flail before say cry and grasp, and babble and bobble before making distinct sounds and taking definite steps. There seem to be few skills that we don't learn painstakingly through trial and error. And I would have thought this makes adaptive sense - you never know what environments a human will be exposed to, what terrains they will have start to walking on, from Kenyan highlands to New York pavements, or what skills they will have to learn from those of an illiterate native to those of a city child. In general, the principle seems to be - consciously flounder around at it, before it becomes a smooth automatic unconscious routine. If you want to be a truly general-purpose general intelligence, that, I think, is the way it has to be. [What related principles govern the Novamente's figure's trial and error learning of how to pick up a ball?] - 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=8660244id_secret=70631633-7d5593
Re[12]: [agi] Funding AGI research
Benjamin, That proves my point [that AGI project can be successfully split into smaller narrow AI subprojects], right? Yes, but it's a largely irrelevant point. Because building a narrow-AI system in an AGI-compatible way is HARDER than building that same narrow-AI component in a non-AGI-compatible way. Even if this is the case (which is not) that would simply mean several development steps: 1) Develop narrow AI with non-reusable AI component and get rewarded for that (because it would be useful system by itself). 2) Refactor non-reusable AI component into reusable AI component and get rewarded for that (because it would reusable component for sale). 3) Apply reusable AI component in AGI and get rewarded for that. If you were analyzing effectiveness of reward systems -- you would notice, that systems (humans, animals, or machines) that are rewarded immediately for positive contribution perform considerably better than systems with reward distributed long after successful accomplishments. So, given the pressures of commerce and academia, people who are motivated to make narrow-AI for its own sake, will almost never create narrow-AI components that are useful for AGI. Sorry, but that does not match with how things really work. So far only researchers/developers who picked narrow-AI approach accomplished something useful for AGI. E.g.: Google, computer languages, network protocols, databases. Pure AGI researchers contributed nothing, but disappointments in AI ideas. Would you agree that splitting very complex and big project into meaningful parts considerably improves chances of success? Yes, sure ... but demanding that these meaningful parts -- be economically viable and/or -- beat competing, somewhat-similar components in competitions dramatically DECREASES chances of success ... INCREASES chances of success. Dramatically. There are lots of examples supporting it both in AI research field and in virtually every area of human research. - 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=8660244id_secret=70646629-5088c0
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
OK for a guestimate take a half-way decent cable connection say Comcast on a good day with DL of 4mbits max and UL of 256kbits max with an undiscriminated protocol, an unknown TCP based protocol, talking to a fat-pipe, low latency server. Assume say 16 byte message header wrappers for all of your 128, 1024 and 10k byte message sizes. So upload is 256kbits, go ahead and saturate it fully with either of your 128+16bytes, 1024+16bytes, and 10k+16bytes packet streams. Using TCP for reliability and assume some overhead say subtract 10% from the saturated value, retransmits, latency. What are we left with? Assume the PC has 1gigbit NIC so it is usually waiting to squeeze out the 256kbits of cable upload capacity. Oh right this is just upstream, DL is 4mbits cable into PC NIC or 1gigbit (assume 60% saturation) so there is ample PC NIC BW for this. So for 256kbits/sec = 256,000 bits/sec, (256,000 bits/sec) / ((1024 + 16)bytes x 8bits/ (message bytes)) = 30.769 messages / sec. So 30.769 messages/sec - 10% = 27.692 messages /sec. About 27.692 message per sec for the 1024 byte message upload stream. Download = 16x UL = 443.072 messages/sec My calculation look right? Note: some Comcast cable connections allow as much as 1.4mbits upload. UL is always way less than DL (dependant on protocol). Other cable companies are similar depends on the company and geographic region... John -Original Message- From: Ed Porter [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 6:50 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] John, Somebody (I think it was David Hart) told me there is a shareware distributed web crawler already available, but I don't know the details, such as how good or fast it is. How fast could P2P communication be done on one PC, on average both sending upstream and receiving downstream from servers with fat pipes? Roughly how many msgs a second for cable connected PC's, say at 128byte and 1024byte, and 10K byte message sizes? Decent guestimates on such numbers would help me think about what sort of interesting distributed NL learning tasks could be done with by AGI-at- Home network. (of course once it showed any promise Google would start doing it a thousand times faster, but at least it would be open source). Ed Porter -Original Message- From: John G. Rose [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 8:31 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed, That is the http protocol, it is a client server request/response communication. Your browser asked for the contents at http://www.nytimes.com. The NY Times server(s) dumped the response stream data to your external IP address. You probably have a NAT'd cable address and NAT'ted again by your local router (if you have one). This communication is mainly one way - except for your original few bytes of http request. For a full ack-nack real-time dynamically addressed protocol there is more involved but say OpenCog could be setup to act as an http server and you could have a http client (browser or whatever) for simplicity in communications. Http is very firewall friendly since it is universally used on the internet. A distributed web crawler is a stretch though the communications is more complicated. John -Original Message- From: Ed Porter [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 6:13 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] John, Thank you for the info. I just did a rough count of all the IMG SRC=http://; in the source of the NYTimes home page which down loaded to my cable-modem connected computer in about 3 seconds. I counted roughly 50 occurrences of that string. I assume there a many other downloaded files such as for layout info. Lets guess a total of at least 100 files that have to be requested and downloaded and displayed. That would be about 33 per second. So what could one do with a system that could do on average about 20 accesses a second on a sustained rate, if a user was leaving it one at night as part of an OpenCog-at- Home project. It seems to me that that would be enough for some interesting large corpus NL work in conjunction with a distributed web crawler. Ed Porter -Original Message- From: John G. Rose [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 7:27 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] From: Ed Porter [mailto:[EMAIL PROTECTED] As I have said many before, to have brain-level AGI I believe you need within several orders of magnitude the representational, computational, and interconnect capability of the human
[agi] Self-building AGI
Ed, At the current stages this may be true, but it should be remembered that building a human-level AGI would be creating a machine that would itself, with the appropriate reading and training, be able to design and program AGIs. No. AGI is not necessarily that capable. In fact first versions of AGI would not be that capable for sure. Consider middle age peasant, for example. Such peasant has general intelligence (GI part in AGI), right? What kind of training would you provide to such peasant in order to make him design AGI? - 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=8660244id_secret=70651687-aa8ee6
Re: [agi] Funding AGI research
On Nov 29, 2007 11:35 PM, Mike Tintner [EMAIL PROTECTED] wrote: Presumably, human learning isn't that slow though - if you simply count the number of attempts made before any given movement is mastered at a basic level (.e.g crawling/ walking/ grasping/ tennis forehand etc)? My guess would be that, for all the frustrations involved, we need relatively few attempts. Maybe in the hundreds or thousands at most? It seems to take tots a damn lot of trials to learn basic skills, and we have plenty of inductive bias in our evolutionary wiring... But then it seems increasingly clear that we use maps/ graphics/ schemas to guide our movements - have you read the latest Blakeslee book on body maps? So does Novamente, it uses an internal simulation-world (among other mechanism)... but that doesn't magically make learning rapid, though it makes it more tractable... ben - 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=8660244id_secret=70644788-023e28
Re: Re[12]: [agi] Funding AGI research
On Nov 30, 2007 12:03 AM, Dennis Gorelik [EMAIL PROTECTED] wrote: Benjamin, That proves my point [that AGI project can be successfully split into smaller narrow AI subprojects], right? Yes, but it's a largely irrelevant point. Because building a narrow-AI system in an AGI-compatible way is HARDER than building that same narrow-AI component in a non-AGI-compatible way. Even if this is the case (which is not) that would simply mean several development steps: 1) Develop narrow AI with non-reusable AI component and get rewarded for that (because it would be useful system by itself). Obviously, most researchers who have developed useful narrow-AI components have not gotten rich from it. The nature of our economy and society is such that most scientific and technical innovators are not dramatically financially rewarded. 2) Refactor non-reusable AI component into reusable AI component and get rewarded for that (because it would reusable component for sale). 3) Apply reusable AI component in AGI and get rewarded for that. - 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=8660244id_secret=70648456-e5f42e
Re: Re[12]: [agi] Funding AGI research
So far only researchers/developers who picked narrow-AI approach accomplished something useful for AGI. E.g.: Google, computer languages, network protocols, databases. These are tools that are useful for AGI RD but so are computer monitors, silicon chips, and desk chairs. Being a useful tool for AGI RD does not make something constitute AGI RD. I do note that I myself have done (and am doing) plenty of narrow AI work in parallel with AGI work. So I'm not arguing against narrow AI nor stating that narrow AI is irrelevant to AGI. But your view of the interrelationship seems extremely oversimplified to me. If it were as simple as you're saying, I imagine we'd have human-level AGI already, as we have loads of decent narrow-AI's for various tasks. -- Ben - 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=8660244id_secret=70647705-610230
RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research]
John, Somebody (I think it was David Hart) told me there is a shareware distributed web crawler already available, but I don't know the details, such as how good or fast it is. How fast could P2P communication be done on one PC, on average both sending upstream and receiving downstream from servers with fat pipes? Roughly how many msgs a second for cable connected PC's, say at 128byte and 1024byte, and 10K byte message sizes? Decent guestimates on such numbers would help me think about what sort of interesting distributed NL learning tasks could be done with by AGI-at-Home network. (of course once it showed any promise Google would start doing it a thousand times faster, but at least it would be open source). Ed Porter -Original Message- From: John G. Rose [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 8:31 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] Ed, That is the http protocol, it is a client server request/response communication. Your browser asked for the contents at http://www.nytimes.com. The NY Times server(s) dumped the response stream data to your external IP address. You probably have a NAT'd cable address and NAT'ted again by your local router (if you have one). This communication is mainly one way - except for your original few bytes of http request. For a full ack-nack real-time dynamically addressed protocol there is more involved but say OpenCog could be setup to act as an http server and you could have a http client (browser or whatever) for simplicity in communications. Http is very firewall friendly since it is universally used on the internet. A distributed web crawler is a stretch though the communications is more complicated. John -Original Message- From: Ed Porter [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 6:13 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] John, Thank you for the info. I just did a rough count of all the IMG SRC=http://; in the source of the NYTimes home page which down loaded to my cable-modem connected computer in about 3 seconds. I counted roughly 50 occurrences of that string. I assume there a many other downloaded files such as for layout info. Lets guess a total of at least 100 files that have to be requested and downloaded and displayed. That would be about 33 per second. So what could one do with a system that could do on average about 20 accesses a second on a sustained rate, if a user was leaving it one at night as part of an OpenCog-at- Home project. It seems to me that that would be enough for some interesting large corpus NL work in conjunction with a distributed web crawler. Ed Porter -Original Message- From: John G. Rose [mailto:[EMAIL PROTECTED] Sent: Thursday, November 29, 2007 7:27 PM To: agi@v2.listbox.com Subject: RE: Hacker intelligence level [WAS Re: [agi] Funding AGI research] From: Ed Porter [mailto:[EMAIL PROTECTED] As I have said many before, to have brain-level AGI I believe you need within several orders of magnitude the representational, computational, and interconnect capability of the human mind. If you had 1 million PC bots on the web, the representational and computational power would be there. But what sort of interconnect would you have? What is the average cable box connected computers upload bandwidth? Is it about 1MBit/sec? If so that would be a bandwidth of 1TBit/sec. But presumably only a small percent of that total 1TBit/sec could be effectively used, say 100Gbits/sec. That's way below brain level, but it is high enough to do valuable AGI research. But would even 10% of this total 1Tbit/sec bandwidth be practically available? How many messages a second can a PC upload a second at say 100K, 10K, 1K, and 128 bytes each? Does anybody know? I've gone through all this while being in VOIP RD. MANY different connections at many different bandwidths, latencies, QOS, it's dirty across the board. Communications between different points is very non- homogenous. There are deep connections and surface alluding to deep web and surface web though network topology is somewhat independent of permissions. The physical infrastructure of the internet allows for certain extremely high bandwidth, low latency connections where the edge is typically lower bandwidth, higher latency but it does depend on the hop graph, time of day, etc.. Messages per sec depends on many factors - network topology starting from pc bus, to NIC, to LAN switch and router, to other routers to ISPs, between ISPs, back in other end, etc.. A cable box usually does anywhere from 64kbit to 1.4mbit upload depending on things such as provider, protocol, hop distance, it totally depends... usually a test is required. On the net, can one bot directly talk to
Re: [agi] Funding AGI research
Presumably, human learning isn't that slow though - if you simply count the number of attempts made before any given movement is mastered at a basic level (.e.g crawling/ walking/ grasping/ tennis forehand etc)? My guess would be that, for all the frustrations involved, we need relatively few attempts. Maybe in the hundreds or thousands at most? But then it seems increasingly clear that we use maps/ graphics/ schemas to guide our movements - have you read the latest Blakeslee book on body maps? (She also cowrote Hawkins' book). Ben: [What related principles govern the Novamente's figure's trial and error learning of how to pick up a ball?] Pure trial and error learning is really slow though... we are now relying on a combination of -- reinforcement from a teacher -- imitation of others' behavior -- trial and error -- active correction of wrong behavior by a teacher ben - 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/?; -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.5.503 / Virus Database: 269.16.9/1158 - Release Date: 11/28/2007 9:11 PM - 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=8660244id_secret=70643693-68a2c5
[agi] Lets count neurons
Matt, And some of the Blue Brain research suggests it is even worse. A mouse cortical column of 10^5 neurons is about 10% connected, What does mean 10% connected? How many connections does average mouse neuron have? 1? but the neurons are arranged such that connections can be formed between any pair of neurons. Extending this idea to the human brain, with 10^6 columns of 10^5 neurons each, each column should be modeled as a 10^5 by 10^5 sparse matrix, Only poor design would require 10^5 by 10^5 matrix if every neuron has to connect only to 1 other neurons. One pointer to 2^17 (131072) address space requires 17 bits. 1 connections require 17 bits. If we want to put 4 bit weighting scale on every connection, then it would be 85000 bytes. 85000 * 1 neurons = 8.5 * 10^9 bytes = 8.5 GB (hard disks of that size were available on PCs ~10 years ago). But in fact mouse's brain does way more than AI has to do. For example, mouse has strong image and sound recognition ability. AGI doesn't require that. Mouse has to manage its muscles in a very high pace. AGI doesn't need that. All these unnecessary features consume lion's share of mouse brain. Mouse must function in way more stressful environment, than AGI must. That again makes mouse brain bigger than AGI has to be. Perhaps there are ways to optimize neural networks by taking advantage of the reliability of digital hardware, but over the last few decades researchers have not found any. Researchers have not found appropriate intelligent algorithms. That doesn't mean that hardware is not sufficient. For narrow AI applications, we can usually find better algorithms than neural networks, for example, arithmetic, deductive logic, or playing chess. But none of these other algorithms are so broadly applicable to so many different domains such as language, speech, vision, robotics, etc. Do you imply that intelligent algorithm must be universal across language, speech, vision, robotics, etc? In humans it's just not the case. Different algorithms are responsible for vision, speech, language, body control etc. - 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=8660244id_secret=70650748-f0eed8
RE: Re[8]: [agi] Funding AGI research
From: Dennis Gorelik [mailto:[EMAIL PROTECTED] John, Is building the compiler more complex than building any application it can build? Note, that compiler doesn't build application. Programmer does (using compiler as a tool). Very true. So then, is the programmer + compiler more complex that the AGI ever will be? Or at some point does the AGI build and improve itself. John - 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=8660244id_secret=70653218-951955