RE: [agi] Epiphany - Statements of Stupidity
Well, these artificial identities need to complete a loop. Say the artificial identity acquires an email address, phone#, a physical address, a bank account, logs onto Amazon and purchases stuff automatically it needs to be able to put money into its bank account. So let's say it has a low profit scheme to scalp day trading profits with its stock trading account. That's the loop, it has to be able to make money to make purchases. And then automatically file its taxes with the IRS. Then it's really starting to look like a full legally functioning identity. It could persist in this fashion for years. I would bet that these identities already exist. What happens when there are many, many of them? Would we even know? John From: Steve Richfield [mailto:steve.richfi...@gmail.com] Sent: Saturday, August 07, 2010 8:17 PM To: agi Subject: Re: [agi] Epiphany - Statements of Stupidity Ian, I recall several years ago that a group in Britain was operating just such a chatterbox as you explained, but did so on numerous sex-related sites, all running simultaneously. The chatterbox emulated young girls looking for sex. The program just sat there doing its thing on numerous sites, and whenever a meeting was set up, it would issue a message to its human owners to alert the police to go and arrest the pedophiles at the arranged time and place. No human interaction was needed between arrests. I can imagine an adaptation, wherein a program claims to be manufacturing explosives, and is looking for other people to deliver those explosives. With such a story line, there should be no problem arranging deliveries, at which time you would arrest the would-be bombers. I wish I could tell you more about the British project, but they were VERY secretive. I suspect that some serious Googling would yield much more. Hopefully you will find this helpful. Steve = On Sat, Aug 7, 2010 at 1:16 PM, Ian Parker ianpark...@gmail.com wrote: I wanted to see what other people's views were.My own view of the risks is as follows. If the Turing Machine is built to be as isomorphic with humans as possible, it would be incredibly dangerous. Indeed I feel that the biological model is far more dangerous than the mathematical. If on the other hand the TM was not isomorphic and made no attempt to be, the dangers would be a lot less. Most Turing/Löbner entries are chatterboxes that work on databases. The database being filled as you chat. Clearly the system cannot go outside its database and is safe. There is in fact some use for such a chatterbox. Clearly a Turing machine would be able to infiltrate militant groups however it was constructed. As for it pretending to be stupid, it would have to know in what direction it had to be stupid. Hence it would have to be a good psychologist. Suppose it logged onto a jihardist website, as well as being able to pass itself off as a true adherent, it could also look at the other members and assess their level of commitment and knowledge. I think that the true Turing/Löbner test is not working in a laboratory environment but they should log onto jihardist sites and see how well they can pass themselves off. If it could do that it really would have arrived. Eventually it could pass itself off as a peniti to use the Mafia term and produce arguments from the Qur'an against the militant position. There would be quite a lot of contracts to be had if there were a realistic prospect of doing this. - Ian Parker On 7 August 2010 06:50, John G. Rose johnr...@polyplexic.com wrote: Philosophical question 2 - Would passing the TT assume human stupidity and if so would a Turing machine be dangerous? Not necessarily, the Turing machine could talk about things like jihad without ultimately identifying with it. Humans without augmentation are only so intelligent. A Turing machine would be potentially dangerous, a really well built one. At some point we'd need to see some DNA as ID of another extended TT. Philosophical question 3 :- Would a TM be a psychologist? I think it would have to be. Could a TM become part of a population simulation that would give us political insights. You can have a relatively stupid TM or a sophisticated one just like humans. It might be easier to pass the TT by not exposing too much intelligence. John These 3 questions seem to me to be the really interesting ones. - Ian Parker --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/? https://www.listbox.com/member/?; Powered by Listbox: http://www.listbox.com agi | https://www.listbox.com/member/archive/303/=now Archives https://www.listbox.com/member/archive/rss/303/ | https://www.listbox.com/member/?; Modify Your Subscription http://www.listbox.com agi | https://www.listbox.com/member
Re: [agi] Epiphany - Statements of Stupidity
If you have a *physical* address an avatar needs to *physically* be there. - Roxxy lives here with her friend Miss Al-Fasaq the belly dancer. Chat lines as Steve describes are not too difficult. In fact the girls (real) on a chat site have a sheet in front of them that gives the appropriate response to a variety of questions. The WI (Women's Institute) did an investigation of the sex industry, and one volunteer actually became a *chatterbox*. Do such entities exist? Probably not in the sex industry, at least not yet. Why do I believe this? Basically because if the sex industry were moving in this direction it would without a doubt be looking at some metric of brain activity to give the customer the best erotic experience. You don't ask Are you gay? You have men making love to men, women-men and women-women. Fing out what gives the customer the biggest kick. You set the story of the porm video. In terms of security I am impressed by the fact that large numbers of bombs have been constructed that don't work and could not work. Hydrogen Peroxidehttp://en.wikipedia.org/wiki/Hydrogen_peroxide can only be prepared in the pure state by chemical reactions. It is unlikely (see notes on vapour pressure at 50C) that anything viable could be produced by distillation on a kitchen stove. Is this due to deliberately misleading information? Have I given the game away? Certainly misleading information is being sent out. However it is probably not being sent out by robotic entities. After all nothing has yet achieved Turing status. In the case of sex it may not be necessary for the client to believe that he is confronted by a *real woman*. A top of the range masturbator/sex aid may not have to pretend to be anything else. - Ian Parker On 8 August 2010 07:30, John G. Rose johnr...@polyplexic.com wrote: Well, these artificial identities need to complete a loop. Say the artificial identity acquires an email address, phone#, a physical address, a bank account, logs onto Amazon and purchases stuff automatically it needs to be able to put money into its bank account. So let's say it has a low profit scheme to scalp day trading profits with its stock trading account. That's the loop, it has to be able to make money to make purchases. And then automatically file its taxes with the IRS. Then it's really starting to look like a full legally functioning identity. It could persist in this fashion for years. I would bet that these identities already exist. What happens when there are many, many of them? Would we even know? John *From:* Steve Richfield [mailto:steve.richfi...@gmail.com] *Sent:* Saturday, August 07, 2010 8:17 PM *To:* agi *Subject:* Re: [agi] Epiphany - Statements of Stupidity Ian, I recall several years ago that a group in Britain was operating just such a chatterbox as you explained, but did so on numerous sex-related sites, all running simultaneously. The chatterbox emulated young girls looking for sex. The program just sat there doing its thing on numerous sites, and whenever a meeting was set up, it would issue a message to its human owners to alert the police to go and arrest the pedophiles at the arranged time and place. No human interaction was needed between arrests. I can imagine an adaptation, wherein a program claims to be manufacturing explosives, and is looking for other people to deliver those explosives. With such a story line, there should be no problem arranging deliveries, at which time you would arrest the would-be bombers. I wish I could tell you more about the British project, but they were VERY secretive. I suspect that some serious Googling would yield much more. Hopefully you will find this helpful. Steve = On Sat, Aug 7, 2010 at 1:16 PM, Ian Parker ianpark...@gmail.com wrote: I wanted to see what other people's views were.My own view of the risks is as follows. If the Turing Machine is built to be as isomorphic with humans as possible, it would be incredibly dangerous. Indeed I feel that the biological model is far more dangerous than the mathematical. If on the other hand the TM was *not* isomorphic and made no attempt to be, the dangers would be a lot less. Most Turing/Löbner entries are chatterboxes that work on databases. The database being filled as you chat. Clearly the system cannot go outside its database and is safe. There is in fact some use for such a chatterbox. Clearly a Turing machine would be able to infiltrate militant groups however it was constructed. As for it pretending to be stupid, it would have to know in what direction it had to be stupid. Hence it would have to be a good psychologist. Suppose it logged onto a jihardist website, as well as being able to pass itself off as a true adherent, it could also look at the other members and assess their level of commitment and knowledge. I think that the true Turing/Löbner test is not working in a laboratory environment
Re: [agi] Epiphany - Statements of Stupidity
John, You brought up some interesting points... On Fri, Aug 6, 2010 at 10:54 PM, John G. Rose johnr...@polyplexic.comwrote: -Original Message- From: Steve Richfield [mailto:steve.richfi...@gmail.com] On Fri, Aug 6, 2010 at 10:09 AM, John G. Rose johnr...@polyplexic.com wrote: statements of stupidity - some of these are examples of cramming sophisticated thoughts into simplistic compressed text. Definitely, as even the thoughts of stupid people transcends our (present) ability to state what is happening behind their eyeballs. Most stupidity is probably beyond simple recognition. For the initial moment, I was just looking at the linguistic low hanging fruit. You are talking about, those phrases, some are clichés, There seems to be no clear boundary between clichés and other stupid statements, except maybe that clichés are exactly quoted like that's just your opinion while other statements are grammatically adapted to fit the sentences and paragraphs that they inhabit. Dr. Eliza already translates idioms before processing. I could add clichés without changing a line of code, e.g. that's just your opinion might translate into something like I am too stupid to to understand your explanation. Dr. Eliza has an extensive wildcard handler, so it should be able to handle the majority of grammatically adapted statements in the same way, by simply including appropriate wildcards in the pattern. are like local K complexity minima, in a knowledge graph of partial linguistic structure, where neural computational energy is preserved, and the statements are patterns with isomorphisms to other experiential knowledge intra and inter agent. That is, other illogical misunderstanding of the real world, which are probably NOT shared with more intelligent agents. This present a serious problem with understanding by more intelligent agents. More intelligent agents have ways of working more optimally with the neural computational energy, perhaps by using other more efficient patterns thus avoiding those particular detrimental pattern/statements. ... and this present a communications problem with agents with radically different intelligences, both greater and lesser. But the statements are catchy because they are common and allow some minimization of computational energy as well as they are like objects in a higher level communication protocol. To store them is less bits and transfer is less bits per second. However, they have negative information content - if that is possible, because they require a false model of the world to process, and produce completely erroneous results. Of course, despite these problems, they DO somewhat accurately communicate the erroneous nature of the thinking, so there IS some value there. Their impact is maximal since they are isomorphic across knowledge and experience. ... the ultimate being: Do, or do not. There is no try. At some point they may just become symbols due to their pre-calculated commonness. Egad, symbols to display stupidity. Could linguistics have anything that is WORSE?! Language is both intelligence enhancing and limiting. Human language is a protocol between agents. So there is minimalist data transfer, I had no choice but to ... is a compressed summary of potentially vastly complex issues. My point is that they could have left the country, killed their adversaries, taken on a new ID, or done any number of radical things that they probably never considered, other than taking whatever action they chose to take. A more accurate statement might be I had no apparent rational choice but to The other low probability choices are lossily compressed out of the expressed statement pattern. It's assumed that there were other choices, usually factored in during the communicational complexity related decompression, being situational. The onus at times is on the person listening to the stupid statement. I see. This example was in reality a gapped or ellipsis, where reasonably presumed words were omitted. These are always a challenge, except in common places like clichés where the missing words can be automatically inserted. Thanks again for your thoughts. Steve = The mind gets hung-up sometimes on this language of ours. Better off at times to think less using English language and express oneself with a wider spectrum communiqué. Doing a dance and throwing paint in the air for example, as some *primitive* cultures actually do, conveys information also and is medium of expression rather than using a restrictive human chat protocol. You are saying that the problem is that our present communication permits statements of stupidity, so we shouldn't have our present system of communication? Scrap English?!!! I consider statements of stupidity as a sort of communications checksum, to see if real interchange of ideas is even possible. Often, it is
Re: [agi] Epiphany - Statements of Stupidity
I wanted to see what other people's views were.My own view of the risks is as follows. If the Turing Machine is built to be as isomorphic with humans as possible, it would be incredibly dangerous. Indeed I feel that the biological model is far more dangerous than the mathematical. If on the other hand the TM was *not* isomorphic and made no attempt to be, the dangers would be a lot less. Most Turing/Löbner entries are chatterboxes that work on databases. The database being filled as you chat. Clearly the system cannot go outside its database and is safe. There is in fact some use for such a chatterbox. Clearly a Turing machine would be able to infiltrate militant groups however it was constructed. As for it pretending to be stupid, it would have to know in what direction it had to be stupid. Hence it would have to be a good psychologist. Suppose it logged onto a jihardist website, as well as being able to pass itself off as a true adherent, it could also look at the other members and assess their level of commitment and knowledge. I think that the true Turing/Löbner test is not working in a laboratory environment but they should log onto jihardist sites and see how well they can pass themselves off. If it could do that it really would have arrived. Eventually it could pass itself off as a *peniti* to use the Mafia term and produce arguments from the Qur'an against the militant position. There would be quite a lot of contracts to be had if there were a realistic prospect of doing this. - Ian Parker On 7 August 2010 06:50, John G. Rose johnr...@polyplexic.com wrote: Philosophical question 2 - Would passing the TT assume human stupidity and if so would a Turing machine be dangerous? Not necessarily, the Turing machine could talk about things like jihad without ultimately identifying with it. Humans without augmentation are only so intelligent. A Turing machine would be potentially dangerous, a really well built one. At some point we'd need to see some DNA as ID of another extended TT. Philosophical question 3 :- Would a TM be a psychologist? I think it would have to be. Could a TM become part of a population simulation that would give us political insights. You can have a relatively stupid TM or a sophisticated one just like humans. It might be easier to pass the TT by not exposing too much intelligence. John These 3 questions seem to me to be the really interesting ones. - Ian Parker --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?; Powered by Listbox: http://www.listbox.com --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
Re: [agi] Epiphany - Statements of Stupidity
Ian, I recall several years ago that a group in Britain was operating just such a chatterbox as you explained, but did so on numerous sex-related sites, all running simultaneously. The chatterbox emulated young girls looking for sex. The program just sat there doing its thing on numerous sites, and whenever a meeting was set up, it would issue a message to its human owners to alert the police to go and arrest the pedophiles at the arranged time and place. No human interaction was needed between arrests. I can imagine an adaptation, wherein a program claims to be manufacturing explosives, and is looking for other people to deliver those explosives. With such a story line, there should be no problem arranging deliveries, at which time you would arrest the would-be bombers. I wish I could tell you more about the British project, but they were VERY secretive. I suspect that some serious Googling would yield much more. Hopefully you will find this helpful. Steve = On Sat, Aug 7, 2010 at 1:16 PM, Ian Parker ianpark...@gmail.com wrote: I wanted to see what other people's views were.My own view of the risks is as follows. If the Turing Machine is built to be as isomorphic with humans as possible, it would be incredibly dangerous. Indeed I feel that the biological model is far more dangerous than the mathematical. If on the other hand the TM was *not* isomorphic and made no attempt to be, the dangers would be a lot less. Most Turing/Löbner entries are chatterboxes that work on databases. The database being filled as you chat. Clearly the system cannot go outside its database and is safe. There is in fact some use for such a chatterbox. Clearly a Turing machine would be able to infiltrate militant groups however it was constructed. As for it pretending to be stupid, it would have to know in what direction it had to be stupid. Hence it would have to be a good psychologist. Suppose it logged onto a jihardist website, as well as being able to pass itself off as a true adherent, it could also look at the other members and assess their level of commitment and knowledge. I think that the true Turing/Löbner test is not working in a laboratory environment but they should log onto jihardist sites and see how well they can pass themselves off. If it could do that it really would have arrived. Eventually it could pass itself off as a *peniti* to use the Mafia term and produce arguments from the Qur'an against the militant position. There would be quite a lot of contracts to be had if there were a realistic prospect of doing this. - Ian Parker On 7 August 2010 06:50, John G. Rose johnr...@polyplexic.com wrote: Philosophical question 2 - Would passing the TT assume human stupidity and if so would a Turing machine be dangerous? Not necessarily, the Turing machine could talk about things like jihad without ultimately identifying with it. Humans without augmentation are only so intelligent. A Turing machine would be potentially dangerous, a really well built one. At some point we'd need to see some DNA as ID of another extended TT. Philosophical question 3 :- Would a TM be a psychologist? I think it would have to be. Could a TM become part of a population simulation that would give us political insights. You can have a relatively stupid TM or a sophisticated one just like humans. It might be easier to pass the TT by not exposing too much intelligence. John These 3 questions seem to me to be the really interesting ones. - Ian Parker --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?; Powered by Listbox: http://www.listbox.com *agi* | Archives https://www.listbox.com/member/archive/303/=now https://www.listbox.com/member/archive/rss/303/ | Modifyhttps://www.listbox.com/member/?;Your Subscription http://www.listbox.com --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
Re: [agi] Epiphany - Statements of Stupidity
sTEVE:I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, What will be the SIMPLEST thing that will mark the first sign of AGI ? - Given that there are zero but zero examples of AGI. Don't you think it would be a good idea to begin at the beginning? With initial AGI? Rather than advanced AGI? --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
Re: [agi] Epiphany - Statements of Stupidity
Mike, Your reply flies in the face of two obvious facts: 1. I have little interest in what is called AGI here. My interests lie elsewhere, e.g. uploading, Dr. Eliza, etc. I posted this piece for several reasons, as it is directly applicable to Dr. Eliza, and because it casts a shadow on future dreams of AGI. I was hoping that those people who have thought things through regarding AGIs might have some thoughts here. Maybe these people don't (yet) exist?! 2. You seem to think that a walk before you run approach, basically a bottom-up approach to AGI, is the obvious one. It sure isn't obvious to me. Besides, if my statements of stupidity theory is true, then why even bother building AGIs, because we won't even be able to meaningfully discuss things with them. Steve == On Fri, Aug 6, 2010 at 2:57 AM, Mike Tintner tint...@blueyonder.co.ukwrote: sTEVE:I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, What will be the SIMPLEST thing that will mark the first sign of AGI ? - Given that there are zero but zero examples of AGI. Don't you think it would be a good idea to begin at the beginning? With initial AGI? Rather than advanced AGI? *agi* | Archives https://www.listbox.com/member/archive/303/=now https://www.listbox.com/member/archive/rss/303/ | Modifyhttps://www.listbox.com/member/?;Your Subscription http://www.listbox.com --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
Re: [agi] Epiphany - Statements of Stupidity
Mike Tintner wrote: What will be the SIMPLEST thing that will mark the first sign of AGI ? - Given that there are zero but zero examples of AGI. Machines have already surpassed human intelligence. If you don't think so, try this IQ test. http://mattmahoney.net/iq/ Or do you prefer to define intelligence as more like a human? In that case I agree that AGI will never happen. No machine will ever be more like a human than a human. I really don't care how you define it. Either way, computers are profoundly affecting the way people interact with each other and with the world. Where is the threshold when machines do most of our thinking for us? Who cares as long as the machines still give us the feeling that we are in charge. -- Matt Mahoney, matmaho...@yahoo.com From: Mike Tintner tint...@blueyonder.co.uk To: agi agi@v2.listbox.com Sent: Fri, August 6, 2010 5:57:33 AM Subject: Re: [agi] Epiphany - Statements of Stupidity sTEVE:I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, What will be the SIMPLEST thing that will mark the first sign of AGI ? - Given that there are zero but zero examples of AGI. Don't you think it would be a good idea to begin at the beginning? With initial AGI? Rather than advanced AGI? agi | Archives | Modify Your Subscription --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
RE: [agi] Epiphany - Statements of Stupidity
statements of stupidity - some of these are examples of cramming sophisticated thoughts into simplistic compressed text. Language is both intelligence enhancing and limiting. Human language is a protocol between agents. So there is minimalist data transfer, I had no choice but to ... is a compressed summary of potentially vastly complex issues. The mind gets hung-up sometimes on this language of ours. Better off at times to think less using English language and express oneself with a wider spectrum communiqué. Doing a dance and throwing paint in the air for example, as some *primitive* cultures actually do, conveys information also and is medium of expression rather than using a restrictive human chat protocol. BTW the rules of etiquette of the human language protocol are even more potentially restricting though necessary for efficient and standardized data transfer to occur. Like, TCP/IP for example. The Etiquette in TCP/IP is like an OSI layer, akin to human language etiquette. John From: Steve Richfield [mailto:steve.richfi...@gmail.com] To All, I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, heidenbugs, etc. I am now wrestling with a new (to me) concept that hopefully others here can shed some light on. People often say things that indicate their limited mental capacity, or at least their inability to comprehend specific situations. 1) One of my favorites are people who say I had no choice but to ..., which of course indicates that they are clearly intellectually challenged because there are ALWAYS other choices, though it may be difficult to find one that is in all respects superior. While theoretically this statement could possibly be correct, in practice I have never found this to be the case. 2) Another one recently from this very forum was If it sounds too good to be true, it probably is. This may be theoretically true, but in fact was, as usual, made as a statement as to why the author was summarily dismissing an apparent opportunity of GREAT value. This dismissal of something BECAUSE of its great value would seem to severely limit the authors prospects for success in life, which probably explains why he spends so much time here challenging others who ARE doing something with their lives. 3) I used to evaluate inventions for some venture capitalists. Sometimes I would find that some basic law of physics, e.g. conservation of energy, would have to be violated for the thing to work. When I explained this to the inventors, their inevitable reply was Yea, and they also said that the Wright Brothers' plane would never fly. To this, I explained that the Wright Brothers had invested ~200 hours of effort working with their crude homemade wind tunnel, and ask what the inventors have done to prove that their own invention would work. 4) One old stupid standby, spoken when you have make a clear point that shows that their argument is full of holes That is just your opinion. No, it is a proven fact for you to accept or refute. 5) Perhaps you have your own pet statements of stupidity? I suspect that there may be enough of these to dismiss some significant fraction of prospective users of beyond-human-capability (I just hate the word intelligence) programs. In short, semantic analysis of these statements typically would NOT find them to be conspicuously false, and hence even an AGI would be tempted to accept them. However, their use almost universally indicates some short-circuit in thinking. The present Dr. Eliza program could easily recognize such statements. OK, so what? What should an AI program do when it encounters a stupid user? Should some attempt be made to explain stupidity to someone who is almost certainly incapable of comprehending their own stupidity? Stupidity is forever is probably true, especially when expressed by an adult. Note my own dismissal of a some past posters for insufficient mental ability to understand certain subjects, whereupon they invariably come back repeating the SAME flawed logic, after I carefully explained the breaks in their logic. Clearly, I was just wasting my effort by continuing to interact with these people. Note that providing a stupid user with ANY output is probably a mistake, because they will almost certainly misconstrue it in some way. Perhaps it might be possible to dumb down the output to preschool-level, at least that (small) part of the output that can be accurately stated in preschool terms. Eventually as computers continue to self-evolve, we will ALL be categorized as some sort of stupid, and receive stupid-adapted output. I wonder whether, ultimately, computers will have ANYTHING to say to us, like any more than we now say to our dogs. Perhaps the final winner of the Reverse Turing Test will remain completely silent?! You don't explain to your dog why you can't pay the rent from The Fall of Colossus. Any thoughts? Steve
Re: [agi] Epiphany - Statements of Stupidity
Maybe you could give me one example from the history of technology where machines ran before they could walk? Where they started complex rather than simple? Or indeed from evolution of any kind? Or indeed from human development? Where children started doing complex mental operations like logic, say, or maths or the equivalent before they could speak? Or started running before they could control their arms, roll over, crawl, sit up, haul themselves up, stand up, totter - just went straight to running?** A bottom-up approach, I would have to agree, clearly isn't obvious to AGI-ers. But then there are v. few AGI-ers who have much sense of history or evolution. It's so much easier to engage in sci-fi fantasies about future, topdown AGI's. It's HARDER to think about where AGI starts - requires serious application to the problem. And frankly, until you or anyone else has a halfway viable of where AGI will or can start, and what uses it will serve, speculation about whether it's worth building complex, sci-fi AGI's is a waste of your valuable time. **PS Note BTW - a distinction that eludes most AGI-ers - a present computer program doing logic or maths or chess, is a fundamentally and massively different thing from a human or AGI doing the same, just as a current program doing NLP is totally different from a human using language. IN all these cases, humans (and real AGIs to come) don't merely manipulate meaningless patterns of numbers, they relate the symbols first to concepts and then to real world referents - massively complex operations totally beyond current computers. The whole history of AI/would-be AGI shows the terrible price of starting complex - with logic/maths/chess programs for example - and not having a clue about how intelligence has to be developed from v. simple origins, step by step, in order to actually understand these activities. From: Steve Richfield Sent: Friday, August 06, 2010 4:52 PM To: agi Subject: Re: [agi] Epiphany - Statements of Stupidity Mike, Your reply flies in the face of two obvious facts: 1. I have little interest in what is called AGI here. My interests lie elsewhere, e.g. uploading, Dr. Eliza, etc. I posted this piece for several reasons, as it is directly applicable to Dr. Eliza, and because it casts a shadow on future dreams of AGI. I was hoping that those people who have thought things through regarding AGIs might have some thoughts here. Maybe these people don't (yet) exist?! 2. You seem to think that a walk before you run approach, basically a bottom-up approach to AGI, is the obvious one. It sure isn't obvious to me. Besides, if my statements of stupidity theory is true, then why even bother building AGIs, because we won't even be able to meaningfully discuss things with them. Steve == On Fri, Aug 6, 2010 at 2:57 AM, Mike Tintner tint...@blueyonder.co.uk wrote: sTEVE:I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, What will be the SIMPLEST thing that will mark the first sign of AGI ? - Given that there are zero but zero examples of AGI. Don't you think it would be a good idea to begin at the beginning? With initial AGI? Rather than advanced AGI? agi | Archives | Modify Your Subscription agi | Archives | Modify Your Subscription --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
Re: [agi] Epiphany - Statements of Stupidity
I think that some quite important philosofical questions are raised by Steve's posting. I don't know BTW how you got it. I monitor all correspondence to the group, and I did not see it. The Turing test is not in fact a test of intelligence, it is a test of similarity with the human. Hence for a machine to be truly Turing it would have to make mistakes. Now any *useful* system will be made as intelligent as we can make it. The TT will be seen to be an irrelevancy. Philosophical question no 1 :- How useful is the TT. As I said in my correspondence With Jan Klouk, the human being is stupid, often dangerously stupid. Philosophical question 2 - Would passing the TT assume human stupidity and if so would a Turing machine be dangerous? Not necessarily, the Turing machine could talk about things like jihad without ultimately identifying with it. Philosophical question 3 :- Would a TM be a psychologist? I think it would have to be. Could a TM become part of a population simulation that would give us political insights. These 3 questions seem to me to be the really interesting ones. - Ian Parker On 6 August 2010 18:09, John G. Rose johnr...@polyplexic.com wrote: statements of stupidity - some of these are examples of cramming sophisticated thoughts into simplistic compressed text. Language is both intelligence enhancing and limiting. Human language is a protocol between agents. So there is minimalist data transfer, I had no choice but to ... is a compressed summary of potentially vastly complex issues. The mind gets hung-up sometimes on this language of ours. Better off at times to think less using English language and express oneself with a wider spectrum communiqué. Doing a dance and throwing paint in the air for example, as some **primitive** cultures actually do, conveys information also and is medium of expression rather than using a restrictive human chat protocol. BTW the rules of etiquette of the human language protocol are even more potentially restricting though necessary for efficient and standardized data transfer to occur. Like, TCP/IP for example. The Etiquette in TCP/IP is like an OSI layer, akin to human language etiquette. John *From:* Steve Richfield [mailto:steve.richfi...@gmail.com] To All, I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, heidenbugs, etc. I am now wrestling with a new (to me) concept that hopefully others here can shed some light on. People often say things that indicate their limited mental capacity, or at least their inability to comprehend specific situations. 1) One of my favorites are people who say I had no choice but to ..., which of course indicates that they are clearly intellectually challenged because there are ALWAYS other choices, though it may be difficult to find one that is in all respects superior. While theoretically this statement could possibly be correct, in practice I have never found this to be the case. 2) Another one recently from this very forum was If it sounds too good to be true, it probably is. This may be theoretically true, but in fact was, as usual, made as a statement as to why the author was summarily dismissing an apparent opportunity of GREAT value. This dismissal of something BECAUSE of its great value would seem to severely limit the authors prospects for success in life, which probably explains why he spends so much time here challenging others who ARE doing something with their lives. 3) I used to evaluate inventions for some venture capitalists. Sometimes I would find that some basic law of physics, e.g. conservation of energy, would have to be violated for the thing to work. When I explained this to the inventors, their inevitable reply was Yea, and they also said that the Wright Brothers' plane would never fly. To this, I explained that the Wright Brothers had invested ~200 hours of effort working with their crude homemade wind tunnel, and ask what the inventors have done to prove that their own invention would work. 4) One old stupid standby, spoken when you have make a clear point that shows that their argument is full of holes That is just your opinion. No, it is a proven fact for you to accept or refute. 5) Perhaps you have your own pet statements of stupidity? I suspect that there may be enough of these to dismiss some significant fraction of prospective users of beyond-human-capability (I just hate the word intelligence) programs. In short, semantic analysis of these statements typically would NOT find them to be conspicuously false, and hence even an AGI would be tempted to accept them. However, their use almost universally indicates some short-circuit in thinking. The present Dr. Eliza program could easily recognize such statements. OK, so what? What should an AI program do when it encounters a stupid user? Should some attempt be made to explain
Re: [agi] Epiphany - Statements of Stupidity
John, Congratulations, as your response was the only one that was on topic!!! On Fri, Aug 6, 2010 at 10:09 AM, John G. Rose johnr...@polyplexic.comwrote: statements of stupidity - some of these are examples of cramming sophisticated thoughts into simplistic compressed text. Definitely, as even the thoughts of stupid people transcends our (present) ability to state what is happening behind their eyeballs. Most stupidity is probably beyond simple recognition. For the initial moment, I was just looking at the linguistic low hanging fruit. Language is both intelligence enhancing and limiting. Human language is a protocol between agents. So there is minimalist data transfer, I had no choice but to ... is a compressed summary of potentially vastly complex issues. My point is that they could have left the country, killed their adversaries, taken on a new ID, or done any number of radical things that they probably never considered, other than taking whatever action they chose to take. A more accurate statement might be I had no apparent rational choice but to The mind gets hung-up sometimes on this language of ours. Better off at times to think less using English language and express oneself with a wider spectrum communiqué. Doing a dance and throwing paint in the air for example, as some **primitive** cultures actually do, conveys information also and is medium of expression rather than using a restrictive human chat protocol. You are saying that the problem is that our present communication permits statements of stupidity, so we shouldn't have our present system of communication? Scrap English?!!! I consider statements of stupidity as a sort of communications checksum, to see if real interchange of ideas is even possible. Often, it is quite impossible to communicate new ideas to inflexible-minded people. BTW the rules of etiquette of the human language protocol are even more potentially restricting though necessary for efficient and standardized data transfer to occur. Like, TCP/IP for example. The Etiquette in TCP/IP is like an OSI layer, akin to human language etiquette. I'm not sure how this relates, other than possibly identifying people who don't honor linguistic etiquette as being (potentially) stupid. Was that your point? Steve == *From:* Steve Richfield [mailto:steve.richfi...@gmail.com] To All, I have posted plenty about statements of ignorance, our probable inability to comprehend what an advanced intelligence might be thinking, heidenbugs, etc. I am now wrestling with a new (to me) concept that hopefully others here can shed some light on. People often say things that indicate their limited mental capacity, or at least their inability to comprehend specific situations. 1) One of my favorites are people who say I had no choice but to ..., which of course indicates that they are clearly intellectually challenged because there are ALWAYS other choices, though it may be difficult to find one that is in all respects superior. While theoretically this statement could possibly be correct, in practice I have never found this to be the case. 2) Another one recently from this very forum was If it sounds too good to be true, it probably is. This may be theoretically true, but in fact was, as usual, made as a statement as to why the author was summarily dismissing an apparent opportunity of GREAT value. This dismissal of something BECAUSE of its great value would seem to severely limit the authors prospects for success in life, which probably explains why he spends so much time here challenging others who ARE doing something with their lives. 3) I used to evaluate inventions for some venture capitalists. Sometimes I would find that some basic law of physics, e.g. conservation of energy, would have to be violated for the thing to work. When I explained this to the inventors, their inevitable reply was Yea, and they also said that the Wright Brothers' plane would never fly. To this, I explained that the Wright Brothers had invested ~200 hours of effort working with their crude homemade wind tunnel, and ask what the inventors have done to prove that their own invention would work. 4) One old stupid standby, spoken when you have make a clear point that shows that their argument is full of holes That is just your opinion. No, it is a proven fact for you to accept or refute. 5) Perhaps you have your own pet statements of stupidity? I suspect that there may be enough of these to dismiss some significant fraction of prospective users of beyond-human-capability (I just hate the word intelligence) programs. In short, semantic analysis of these statements typically would NOT find them to be conspicuously false, and hence even an AGI would be tempted to accept them. However, their use almost universally indicates some short-circuit in thinking. The present Dr. Eliza program could easily recognize such statements. OK,
RE: [agi] Epiphany - Statements of Stupidity
-Original Message- From: Ian Parker [mailto:ianpark...@gmail.com] The Turing test is not in fact a test of intelligence, it is a test of similarity with the human. Hence for a machine to be truly Turing it would have to make mistakes. Now any useful system will be made as intelligent as we can make it. The TT will be seen to be an irrelevancy. Philosophical question no 1 :- How useful is the TT. TT in its basic form is rather simplistic. It's thought of usually in its ideal form, the determination of an AI or a human. I look at it more of analogue verses discrete boolean. Much of what is out there is human with computer augmentation and echoes of human interaction. It's blurry in reality and the TT has been passed in some ways but not in its most ideal way. As I said in my correspondence With Jan Klouk, the human being is stupid, often dangerously stupid. Philosophical question 2 - Would passing the TT assume human stupidity and if so would a Turing machine be dangerous? Not necessarily, the Turing machine could talk about things like jihad without ultimately identifying with it. Humans without augmentation are only so intelligent. A Turing machine would be potentially dangerous, a really well built one. At some point we'd need to see some DNA as ID of another extended TT. Philosophical question 3 :- Would a TM be a psychologist? I think it would have to be. Could a TM become part of a population simulation that would give us political insights. You can have a relatively stupid TM or a sophisticated one just like humans. It might be easier to pass the TT by not exposing too much intelligence. John These 3 questions seem to me to be the really interesting ones. - Ian Parker --- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244id_secret=8660244-6e7fb59c Powered by Listbox: http://www.listbox.com
