rlm@57: #+TITLE:Transcript of Aaron Sloman - Artificial Intelligence - Psychology - Oxford Interview
rlm@57: #+AUTHOR:Dylan Holmes
rlm@57: #+EMAIL:
rlm@57: #+STYLE:
rlm@57:
rlm@57:
rlm@57: #+BEGIN_QUOTE
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57:
rlm@57: *Editor's note:* This is a working draft transcript which I made of
rlm@57: [[http://www.youtube.com/watch?feature=player_detailpage&v=iuH8dC7Snno][this nice interview]] of Aaron Sloman. Having just finished one
rlm@57: iteration of transcription, I still need to go in and clean up the
rlm@57: formatting and fix the parts that I misheard, so you can expect the
rlm@57: text to improve significantly in the near future.
rlm@57:
rlm@57: To the extent that this is my work, you have my permission to make
rlm@57: copies of this transcript for your own purposes. Also, feel free to
rlm@57: e-mail me with comments or corrections.
rlm@57:
rlm@57: You can send mail to =transcript@aurellem.org=.
rlm@57:
rlm@57: Cheers,
rlm@57:
rlm@57: ---Dylan
rlm@57: #+END_QUOTE
rlm@57:
rlm@57:
rlm@57:
rlm@57: * Introduction
rlm@57:
rlm@57: ** Aaron Sloman evolves into a philosopher of AI
rlm@57: [0:09] My name is Aaron Sloman. My first degree many years ago in
rlm@57: Capetown University was in Physics and Mathematics, and I intended to
rlm@57: go and be a mathematician. I came to Oxford and encountered
rlm@57: philosophers --- I had started reading philosophy and discussing
rlm@57: philosophy before then, and then I found that there were philosophers
rlm@57: who said things about mathematics that I thought were wrong, so
rlm@57: gradually got more and more involved in [philosophy] discussions and
rlm@57: switched to doing philosophy DPhil. Then I became a philosophy
rlm@57: lecturer and about six years later, I was introduced to artificial
rlm@57: intelligence when I was a lecturer at Sussex University in philosophy
rlm@57: and I very soon became convinced that the best way to make progress in
rlm@57: both areas of philosophy (including philosophy of mathematics which I
rlm@57: felt i hadn't dealt with adequately in my DPhil) about the philosophy
rlm@57: of mathematics, philosophy of mind, philsophy of language and all
rlm@57: those things---the best way was to try to design and test working
rlm@57: fragments of mind and maybe eventually put them all together but
rlm@57: initially just working fragments that would do various things.
rlm@57:
rlm@57: [1:12] And I learned to program and ~ with various other people
rlm@57: including ~Margaret Boden whom you've interviewed, developed---helped
rlm@57: develop an undergraduate degree in AI and other things and also began
rlm@57: to do research in AI and so on which I thought of as doing philosophy,
rlm@57: primarily.
rlm@57:
rlm@57: [1:29] And then I later moved to the University of Birmingham and I
rlm@57: was there --- I came in 1991 --- and I've been retired for a while but
rlm@57: I'm not interested in golf or gardening so I just go on doing full
rlm@57: time research and my department is happy to keep me on without paying
rlm@57: me and provide space and resources and I come, meeting bright people
rlm@57: at conferences and try to learn and make progress if I can.
rlm@57:
rlm@57: ** AI is hard, in part because there are tempting non-problems.
rlm@57:
rlm@57: One of the things I learnt and understood more and more over the many
rlm@57: years --- forty years or so since I first encountered AI --- is how
rlm@57: hard the problems are, and in part that's because it's very often
rlm@57: tempting to /think/ the problem is something different from what it
rlm@57: actually is, and then people design solutions to the non-problems, and
rlm@57: I think of most of my work now as just helping to clarify what the
rlm@57: problems are: what is it that we're trying to explain --- and maybe
rlm@57: this is leading into what you wanted to talk about:
rlm@57:
rlm@57: I now think that one of the ways of getting a deep understanding of
rlm@57: that is to find out what were the problems that biological evolution
rlm@57: solved, because we are a product of /many/ solutions to /many/
rlm@57: problems, and if we just try to go in and work out what the whole
rlm@57: system is doing, we may get it all wrong, or badly wrong.
rlm@57:
rlm@57:
rlm@57: * What problems of intelligence did evolution solve?
rlm@57:
rlm@57: ** Intelligence consists of solutions to many evolutionary problems; no single development (e.g. communication) was key to human-level intelligence.
rlm@57:
rlm@57: [2:57] Well, first I would challenge that we are the dominant
rlm@57: species. I know it looks like that but actually if you count biomass,
rlm@57: if you count number of species, if you count number of individuals,
rlm@57: the dominant species are microbes --- maybe not one of them but anyway
rlm@57: they're the ones who dominate in that sense, and furthermore we are
rlm@57: mostly --- we are largely composed of microbes, without which we
rlm@57: wouldn't survive.
rlm@57:
rlm@57:
rlm@57: # ** Many nonlinguistic competences require sophisticated internal representations
rlm@57: [3:27] But there are things that make humans (you could say) best at
rlm@57: those things, or worst at those things, but it's a combination. And I
rlm@57: think it was a collection of developments of which there isn't any
rlm@57: single one. [] there might be, some people say, human language which
rlm@57: changed everything. By our human language, they mean human
rlm@57: communication in words, but I think that was a later development from
rlm@57: what must have started as the use of /internal/ forms of
rlm@57: representation --- which are there in nest-building birds, in
rlm@57: pre-verbal children, in hunting mammals --- because you can't take in
rlm@57: information about a complex structured environment in which things can
rlm@57: change and you may have to be able to work out what's possible and
rlm@57: what isn't possible, without having some way of representing the
rlm@57: components of the environment, their relationships, the kinds of
rlm@57: things they can and can't do, the kinds of things you might or might
rlm@57: not be able to do --- and /that/ kind of capability needs internal
rlm@57: languages, and I and colleagues [at Birmingham] have been referring to
rlm@57: them as generalized languages because some people object to
rlm@57: referring...to using language to refer to something that isn't used
rlm@57: for communication. But from that viewpoint, not only humans but many
rlm@57: other animals developed abilities to do things to their environment to
rlm@57: make them more friendly to themselves, which depended on being able to
rlm@57: represent possible futures, possible actions, and work out what's the
rlm@57: best thing to do.
rlm@57:
rlm@57: [5:13] And nest-building in corvids for instance---crows, magpies,
rlm@57: [hawks], and so on --- are way beyond what current robots can do, and
rlm@57: in fact I think most humans would be challenged if they had to go and
rlm@57: find a collection of twigs, one at a time, maybe bring them with just
rlm@57: one hand --- or with your mouth --- and assemble them into a
rlm@57: structure that, you know, is shaped like a nest, and is fairly rigid,
rlm@57: and you could trust your eggs in them when wind blows. But they're
rlm@57: doing it, and so ... they're not our evolutionary ancestors, but
rlm@57: they're an indication --- and that example is an indication --- of
rlm@57: what must have evolved in order to provide control over the
rlm@57: environment in /that/ species.
rlm@57:
rlm@57: ** Speculation about how communication might have evolved from internal lanagues.
rlm@57: [5:56] And I think hunting mammals, fruit-picking mammals, mammals
rlm@57: that can rearrange parts of the environment, provide shelters, needed
rlm@57: to have .... also needed to have ways of representing possible
rlm@57: futures, not just what's there in the environment. I think at a later
rlm@57: stage, that developed into a form of communication, or rather the
rlm@57: /internal/ forms of representation became usable as a basis for
rlm@57: providing [context] to be communicated. And that happened, I think,
rlm@57: initially through performing actions that expressed intentions, and
rlm@57: probably led to situtations where an action (for instance, moving some
rlm@57: large object) was performed more easily, or more successfully, or more
rlm@57: accurately if it was done collaboratively. So someone who had worked
rlm@57: out what to do might start doing it, and then a conspecific might be
rlm@57: able to work out what the intention is, because that person has the
rlm@57: /same/ forms of representation and can build theories about what's
rlm@57: going on, and might then be able to help.
rlm@57:
rlm@57: [7:11] You can imagine that if that started happening more (a lot of
rlm@57: collaboration based on inferred intentions and plans) then sometimes
rlm@57: the inferences might be obscure and difficult, so the /actions/ might
rlm@57: be enhanced to provide signals as to what the intention is, and what
rlm@57: the best way is to help, and so on.
rlm@57:
rlm@57: [7:35] So, this is all handwaving and wild speculation, but I think
rlm@57: it's consistent with a large collection of facts which one can look at
rlm@57: --- and find if one looks for them, but one won't know if [some]one
rlm@57: doesn't look for them --- about the way children, for instance, who
rlm@57: can't yet talk, communicate, and the things they'll do, like going to
rlm@57: the mother and turning the face to point in the direction where the
rlm@57: child wants it to look and so on; that's an extreme version of action
rlm@57: indicating intention.
rlm@57:
rlm@57: [8:03] Anyway. That's a very long roundabout answer to one conjecture
rlm@57: that the use of communicative language is what gave humans their
rlm@57: unique power to create and destroy and whatever, and I'm saying that
rlm@57: if by that you mean /communicative/ language, then I'm saying there
rlm@57: was something before that which was /non/-communicative language, and I
rlm@57: suspect that noncommunicative language continues to play a deep role
rlm@57: in /all/ human perception ---in mathematical and scientific reasoning, in
rlm@57: problem solving --- and we don't understand very much about it.
rlm@57:
rlm@57: [8:48]
rlm@57: I'm sure there's a lot more to be said about the development of
rlm@57: different kinds of senses, the development of brain structures and
rlm@57: mechanisms is above all that, but perhaps I've droned on long enough
rlm@57: on that question.
rlm@57:
rlm@57:
rlm@57: * How do language and internal states relate to AI?
rlm@57:
rlm@57: [9:09] Well, I think most of the human and animal capabilities that
rlm@57: I've been referring to are not yet to be found in current robots or
rlm@57: [computing] systems, and I think there are two reasons for that: one
rlm@57: is that it's intrinsically very difficult; I think that in particular
rlm@57: it may turn out that the forms of information processing that one can
rlm@57: implement on digital computers as we currently know them may not be as
rlm@57: well suited to performing some of these tasks as other kinds of
rlm@57: computing about which we don't know so much --- for example, I think
rlm@57: there may be important special features about /chemical/ computers
rlm@57: which we might [talk about in a little bit? find out about].
rlm@57:
rlm@57: ** In AI, false assumptions can lead investigators astray.
rlm@57: [9:57] So, one of the problems then is that the tasks are hard ... but
rlm@57: there's a deeper problem as to why AI hasn't made a great deal of
rlm@57: progress on these problems that I'm talking about, and that is that
rlm@57: most AI researchers assume things---and this is not just AI
rlm@57: researchers, but [also] philsophers, and psychologists, and people
rlm@57: studying animal behavior---make assumptions about what it is that
rlm@57: animals or humans do, for instance make assumptions about what vision
rlm@57: is for, or assumptions about what motivation is and how motivation
rlm@57: works, or assumptions about how learning works, and then they try ---
rlm@57: the AI people try --- to model [or] build systems that perform those
rlm@57: assumed functions. So if you get the /functions/ wrong, then even if
rlm@57: you implement some of the functions that you're trying to implement,
rlm@57: they won't necessarily perform the tasks that the initial objective
rlm@57: was to imitate, for instance the tasks that humans, and nest-building
rlm@57: birds, and monkeys and so on can perform.
rlm@57:
rlm@57: ** Example: Vision is not just about finding surfaces, but about finding affordances.
rlm@57: [11:09] I'll give you a simple example --- well, maybe not so simple,
rlm@57: but --- It's often assumed that the function of vision in humans (and
rlm@57: in other animals with good eyesight and so on) is to take in optical
rlm@57: information that hits the retina, and form into the (maybe changing
rlm@57: --- or, really, in our case definitely changing) patterns of
rlm@57: illumination where there are sensory receptors that detect those
rlm@57: patterns, and then somehow from that information (plus maybe other
rlm@57: information gained from head movement or from comparisons between two
rlm@57: eyes) to work out what there was in the environment that produced
rlm@57: those patterns, and that is often taken to mean \ldquo{}where were the
rlm@57: surfaces off which the light bounced before it came to me\rdquo{}. So
rlm@57: you essentially think of the task of the visual system as being to
rlm@57: reverse the image formation process: so the 3D structure's there, the
rlm@57: lens causes the image to form in the retina, and then the brain goes
rlm@57: back to a model of that 3D structure there. That's a very plausible
rlm@57: theory about vision, and it may be that that's a /subset/ of what
rlm@57: human vision does, but I think James Gibson pointed out that that kind
rlm@57: of thing is not necessarily going to be very useful for an organism,
rlm@57: and it's very unlikely that that's the main function of perception in
rlm@57: general, namely to produce some physical description of what's out
rlm@57: there.
rlm@57:
rlm@57: [12:37] What does an animal /need/? It needs to know what it can do,
rlm@57: what it can't do, what the consequences of its actions will be
rlm@57: .... so, he introduced the word /affordance/, so from his point of
rlm@57: view, the function of vision, perception, are to inform the organism
rlm@57: of what the /affordances/ are for action, where that would mean what
rlm@57: the animal, /given/ its morphology (what it can do with its mouth, its
rlm@57: limbs, and so on, and the ways it can move) what it can do, what its
rlm@57: needs are, what the obstacles are, and how the environment supports or
rlm@57: obstructs those possible actions.
rlm@57:
rlm@57: [13:15] And that's a very different collection of information
rlm@57: structures that you need from, say, \ldquo{}where are all the
rlm@57: surfaces?\rdquo{}: if you've got all the surfaces, /deriving/ the
rlm@57: affordances would still be a major task. So, if you think of the
rlm@57: perceptual system as primarily (for biological organisms) being
rlm@57: devices that provide information about affordances and so on, then the
rlm@57: tasks look very different. And most of the people working, doing
rlm@57: research on computer vision in robots, I think haven't taken all that
rlm@57: on board, so they're trying to get machines to do things which, even
rlm@57: if they were successful, would not make the robots very intelligent
rlm@57: (and in fact, even the ones they're trying to do are not really easy
rlm@57: to do, and they don't succeed very well--- although, there's progress;
rlm@57: I shouldn't disparage it too much.)
rlm@57:
rlm@57: ** Online and offline intelligence
rlm@57:
rlm@57: [14:10] It gets more complex as animals get more sophisticated. So, I
rlm@57: like to make a distinction between online intelligence and offline
rlm@57: intelligence. So, for example, if I want to pick something up --- like
rlm@57: this leaf --- I was able to select
rlm@57: it from all the others in there, and while moving my hand towards it,
rlm@57: I was able to guide its trajectory, making sure it was going roughly
rlm@57: in the right direction --- as opposed to going out there, which
rlm@57: wouldn't have been able to pick it up --- and these two fingers ended
rlm@57: up with a portion of the leaf between them, so that I was able to tell
rlm@57: when I'm ready to do that
rlm@57: and at that point, I clamped my fingers and then I could pick up the
rlm@57: leaf.
rlm@57:
rlm@57: [14:54] Whereas, --- and that's an example of online intelligence:
rlm@57: during the performance of an action (both from the stage where it's
rlm@57: initiated, and during the intermediate stages, and where it's
rlm@57: completed) I'm taking in information relevant to controlling all those
rlm@57: stages, and that relevant information keeps changing. That means I
rlm@57: need stores of transient information which gets discarded almost
rlm@57: immediately and replaced or something. That's online intelligence. And
rlm@57: there are many forms; that's just one example, and Gibson discussed
rlm@57: quite a lot of examples which I won't try to replicate now.
rlm@57:
rlm@57: [15:30] But in offline intelligence, you're not necessarily actually
rlm@57: /performing/ the actions when you're using your intelligence; you're
rlm@57: thinking about /possible/ actions. So, for instance, I could think
rlm@57: about how fast or by what route I would get back to the lecture room
rlm@57: if I wanted to [get to the next talk] or something. And I know where
rlm@57: the door is, roughly speaking, and I know roughly which route I would
rlm@57: take, when I go out, I should go to the left or to the right, because
rlm@57: I've stored information about where the spaces are, where the
rlm@57: buildings are, where the door was that we came out --- but in using
rlm@57: that information to think about that route, I'm not actually
rlm@57: performing the action. I'm not even /simulating/ it in detail: the
rlm@57: precise details of direction and speed and when to clamp my fingers,
rlm@57: or when to contract my leg muscles when walking, are all irrelevant to
rlm@57: thinking about a good route, or thinking about the potential things
rlm@57: that might happen on the way. Or what would be a good place to meet
rlm@57: someone who I think [for an acquaintance in particular] --- [barber]
rlm@57: or something --- I don't necessarily have to work out exactly /where/
rlm@57: the person's going to stand, or from what angle I would recognize
rlm@57: them, and so on.
rlm@57:
rlm@57: [16:46] So, offline intelligence --- which I think became not just a
rlm@57: human competence; I think there are other animals that have aspects of
rlm@57: it: Squirrels are very impressive as you watch them. Gray squirrels at
rlm@57: any rate, as you watch them defeating squirrel-proof birdfeeders, seem
rlm@57: to have a lot of that [offline intelligence], as well as the online
rlm@57: intelligence when they eventually perform the action they've worked
rlm@57: out [] that will get them to the nuts.
rlm@57:
rlm@57: [17:16] And I think that what happened during our evolution is that
rlm@57: mechanisms for acquiring and processing and storing and manipulating
rlm@57: information that is more and more remote from the performance of
rlm@57: actions developed. An example is taking in information about where
rlm@57: locations are that you might need to go to infrequently: There's a
rlm@57: store of a particular type of material that's good for building on
rlm@57: roofs of houses or something out around there in some
rlm@57: direction. There's a good place to get water somewhere in another
rlm@57: direction. There are people that you'd like to go and visit in
rlm@57: another place, and so on.
rlm@57:
rlm@57: [17:59] So taking in information about an extended environment and
rlm@57: building it into a structure that you can make use of for different
rlm@57: purposes is another example of offline intelligence. And when we do
rlm@57: that, we sometimes use only our brains, but in modern times, we also
rlm@57: learned how to make maps on paper and walls and so on. And it's not
rlm@57: clear whether the stuff inside our heads has the same structures as
rlm@57: the maps we make on paper: the maps on paper have a different
rlm@57: function; they may be used to communicate with others, or meant for
rlm@57: /looking/ at, whereas the stuff in your head you don't /look/ at; you
rlm@57: use it in some other way.
rlm@57:
rlm@57: [18:46] So, what I'm getting at is that there's a great deal of human
rlm@57: intelligence (and animal intelligence) which is involved in what's
rlm@57: possible in the future, what exists in distant places, what might have
rlm@57: happened in the past (sometimes you need to know why something is as
rlm@57: it is, because that might be relevant to what you should or shouldn't
rlm@57: do in the future, and so on), and I think there was something about
rlm@57: human evolution that extended that offline intelligence way beyond
rlm@57: that of animals. And I don't think it was /just/ human language, (but
rlm@57: human language had something to do with it) but I think there was
rlm@57: something else that came earlier than language which involves the
rlm@57: ability to use your offline intelligence to discover something that
rlm@57: has a rich mathematical structure.
rlm@57:
rlm@57: ** Example: Even toddlers use sophisticated geometric knowledge
rlm@57: #+<>
rlm@57: [19:44] I'll give you a simple example: if you look through a gap, you
rlm@57: can see something that's on the other side of the gap. Now, you
rlm@57: /might/ see what you want to see, or you might see only part of it. If
rlm@57: you want to see more of it, which way would you move? Well, you could
rlm@57: either move /sideways/, and see through the gap---and see it roughly
rlm@57: the same amount but a different part of it [if it's a ????], or you
rlm@57: could move /towards/ the gap and then your view will widen as you
rlm@57: approach the gap. Now, there's a bit of mathematics in there, insofar
rlm@57: as you are implicitly assuming that information travels in straight
rlm@57: lines, and as you go closer to a gap, the straight lines that you can
rlm@57: draw from where you are through the gap, widen as you approach that
rlm@57: gap. Now, there's a kind of theorem of Euclidean geometry in there
rlm@57: which I'm not going to try to state very precisely (and as far as I
rlm@57: know, wasn't stated explicitly in Euclidean geometry) but it's
rlm@57: something every toddler--- human toddler---learns. (Maybe other
rlm@57: animals also know it, I don't know.) But there are many more things,
rlm@57: actions to perform, to get you more information about things, actions
rlm@57: to perform to conceal information from other people, actions that will
rlm@57: enable you to operate, to act on a rigid object in one place in order
rlm@57: to produce an effect on another place. So, there's a lot of stuff that
rlm@57: involves lines and rotations and angles and speeds and so on that I
rlm@57: think humans (maybe, to a lesser extent, other animals) develop the
rlm@57: ability to think about in a generic way. That means that you could
rlm@57: take out the generalizations from the particular contexts and then
rlm@57: re-use them in a new contexts in ways that I think are not yet
rlm@57: represented at all in AI and in theories of human learning in any []
rlm@57: way --- although some people are trying to study learning of mathematics.
rlm@57:
rlm@57: * Animal intelligence
rlm@57:
rlm@57: ** The priority is /cataloguing/ what competences have evolved, not ranking them.
rlm@57: [22:03] I wasn't going to challenge the claim that humans can do more
rlm@57: sophisticated forms of [tracking], just to mention that there are some
rlm@57: things that other animals can do which are in some ways comparable,
rlm@57: and some ways superior to [things] that humans can do. In particular,
rlm@57: there are species of birds and also, I think, some rodents ---
rlm@57: squirrels, or something --- I don't know enough about the variety ---
rlm@57: that can hide nuts and remember where they've hidden them, and go back
rlm@57: to them. And there have been tests which show that some birds are able
rlm@57: to hide tens --- you know, [eighteen] or something nuts --- and to
rlm@57: remember which ones have been taken, which ones haven't, and so
rlm@57: on. And I suspect most humans can't do that. I wouldn't want to say
rlm@57: categorically that maybe we couldn't, because humans are very
rlm@57: [varied], and also [a few] people can develop particular competences
rlm@57: through training. But it's certainly not something I can do.
rlm@57:
rlm@57:
rlm@57: ** AI can be used to test philosophical theories
rlm@57: [23:01] But I also would like to say that I am not myself particularly
rlm@57: interested in trying to align animal intelligences according to any
rlm@57: kind of scale of superiority; I'm just trying to understand what it
rlm@57: was that biological evolution produced, and how it works, and I'm
rlm@57: interested in AI /mainly/ because I think that when one comes up with
rlm@57: theories about how these things work, one needs to have some way of
rlm@57: testing the theory. And AI provides ways of implementing and testing
rlm@57: theories that were not previously available: Immanuel Kant was trying
rlm@57: to come up with theories about how minds work, but he didn't have any
rlm@57: kind of a mechanism that he could build to test his theory about the
rlm@57: nature of mathematical knowledge, for instance, or how concepts were
rlm@57: developed from babyhood onward. Whereas now, if we do develop a
rlm@57: theory, we have a criterion of adequacy, namely it should be precise
rlm@57: enough and rich enough and detailed to enable a model to be
rlm@57: built. And then we can see if it works.
rlm@57:
rlm@57: [24:07] If it works, it doesn't mean we've proved that the theory is
rlm@57: correct; it just shows it's a candidate. And if it doesn't work, then
rlm@57: it's not a candidate as it stands; it would need to be modified in
rlm@57: some way.
rlm@57:
rlm@57: * Is abstract general intelligence feasible?
rlm@57:
rlm@57: ** It's misleading to compare the brain and its neurons to a computer made of transistors
rlm@57: [24:27] I think there's a lot of optimism based on false clues:
rlm@57: the...for example, one of the false clues is to count the number of
rlm@57: neurons in the brain, and then talk about the number of transistors
rlm@57: you can fit into a computer or something, and then compare them. It
rlm@57: might turn out that the study of the way synapses work (which leads
rlm@57: some people to say that a typical synapse [] in the human brain has
rlm@57: computational power comparable to the Internet a few years ago,
rlm@57: because of the number of different molecules that are doing things,
rlm@57: the variety of types of things that are being done in those molecular
rlm@57: interactions, and the speed at which they happen, if you somehow count
rlm@57: up the number of operations per second or something, then you get
rlm@57: these comparable figures).
rlm@57:
rlm@57: ** For example, brains may rely heavily on chemical information processing
rlm@57: Now even if the details aren't right, there may just be a lot of
rlm@57: information processing that...going on in brains at the /molecular/
rlm@57: level, not the neural level. Then, if that's the case, the processing
rlm@57: units will be orders of magnitude larger in number than the number of
rlm@57: neurons. And it's certainly the case that all the original biological
rlm@57: forms of information processing were chemical; there weren't brains
rlm@57: around, and still aren't in most microbes. And even when humans grow
rlm@57: their brains, the process of starting from a fertilized egg and
rlm@57: producing this rich and complex structure is, for much of the time,
rlm@57: under the control of chemical computations, chemical information
rlm@57: processing---of course combined with physical sorts of materials and
rlm@57: energy and so on as well.
rlm@57:
rlm@57: [26:25] So it would seem very strange if all that capability was
rlm@57: something thrown away when you've got a brain and all the information
rlm@57: processing, the [challenges that were handled in making a brain],
rlm@57: ... This is handwaving on my part; I'm just saying that we /might/
rlm@57: learn that what brains do is not what we think they do, and that
rlm@57: problems of replicating them are not what we think they are, solely in
rlm@57: terms of numerical estimate of time scales, the number of components,
rlm@57: and so on.
rlm@57:
rlm@57: ** Brain algorithms may simply be optimized for certain kinds of information processing other than bit manipulations
rlm@57: [26:56] But apart from that, the other basis of skepticism concerns
rlm@57: how well we understand what the problems are. I think there are many
rlm@57: people who try to formalize the problems of designing an intelligent
rlm@57: system in terms of streams of information thought of as bit streams or
rlm@57: collections of bit streams, and they think of as the problems of
rlm@57: intelligence as being the construction or detection of patterns in
rlm@57: those, and perhaps not just detection of patterns, but detection of
rlm@57: patterns that are useable for sending /out/ streams to control motors
rlm@57: and so on in order to []. And that way of conceptualizing the problem
rlm@57: may lead on the one hand to oversimplification, so that the things
rlm@57: that /would/ be achieved, if those goals were achieved, maybe much
rlm@57: simpler, in some ways inadequate. Or the replication of human
rlm@57: intelligence, or the matching of human intelligence---or for that
rlm@57: matter, squirrel intelligence---but in another way, it may also make
rlm@57: the problem harder: it may be that some of the kinds of things that
rlm@57: biological evolution has achieved can't be done that way. And one of
rlm@57: the ways that might turn out to be the case is not because it's not
rlm@57: impossible in principle to do some of the information processing on
rlm@57: artificial computers-based-on-transistors and other bit-manipulating
rlm@57: []---but it may just be that the computational complexity of solving
rlm@57: problems, processes, or finding solutions to complex problems, are
rlm@57: much greater and therefore you might need a much larger universe than
rlm@57: we have available in order to do things.
rlm@57:
rlm@57: ** Example: find the shortest path by dangling strings
rlm@57: [28:55] Then if the underlying mechanisms were different, the
rlm@57: information processing mechanisms, they might be better tailored to
rlm@57: particular sorts of computation. There's a [] example, which is
rlm@57: finding the shortest route if you've got a collection of roads, and
rlm@57: they may be curved roads, and lots of tangled routes from A to B to C,
rlm@57: and so on. And if you start at A and you want to get to Z --- a place
rlm@57: somewhere on that map --- the process of finding the shortest route
rlm@57: will involve searching through all these different possibilities and
rlm@57: rejecting some that are longer than others and so on. But if you make
rlm@57: a model of that map out of string, where these strings are all laid
rlm@57: out on the maps and so have the lengths of the routes. Then if you
rlm@57: hold the two knots in the string -- it's a network of string --- which
rlm@57: correspond to the start point and end point, then /pull/, then the
rlm@57: bits of string that you're left with in a straight line will give you
rlm@57: the shortest route, and that process of pulling just gets you the
rlm@57: solution very rapidly in a parallel computation, where all the others
rlm@57: just hang by the wayside, so to speak.
rlm@57:
rlm@57: ** In sum, we know surprisingly little about the kinds of problems that evolution solved, and the manner in which they were solved.
rlm@57: [30:15] Now, I'm not saying brains can build networks of string and
rlm@57: pull them or anything like that; that's just an illustration of how if
rlm@57: you have the right representation, correctly implemented---or suitably
rlm@57: implemented---for a problem, then you can avoid very combinatorially
rlm@57: complex searches, which will maybe grow exponentially with the number
rlm@57: of components in your map, whereas with this thing, the time it takes
rlm@57: won't depend on how many strings you've [got on the map]; you just
rlm@57: pull, and it will depend only on the shortest route that exists in
rlm@57: there. Even if that shortest route wasn't obvious on the original map.
rlm@57:
rlm@57:
rlm@57: [30:59] So that's a rather long-winded way of formulating the
rlm@57: conjecture which---of supporting, a roundabout way of supporting the
rlm@57: conjecture that there may be something about the way molecules perform
rlm@57: computations where they have the combination of continuous change as
rlm@57: things move through space and come together and move apart, and
rlm@57: whatever --- and also snap into states that then persist, so [as you
rlm@57: learn from] quantum mechanics, you can have stable molecular
rlm@57: structures which are quite hard to separate, and then in catalytic
rlm@57: processes you can separate them, or extreme temperatures, or strong
rlm@57: forces, but they may nevertheless be able to move very rapidly in some
rlm@57: conditions in order to perform computations.
rlm@57:
rlm@57: [31:49] Now there may be things about that kind of structure that
rlm@57: enable searching for solutions to /certain/ classes of problems to be
rlm@57: done much more efficiently (by brain) than anything we could do with
rlm@57: computers. It's just an open question.
rlm@57:
rlm@57: [32:04] So it /might/ turn out that we need new kinds of technology
rlm@57: that aren't on the horizon in order to replicate the functions that
rlm@57: animal brains perform ---or, it might not. I just don't know. I'm not
rlm@57: claiming that there's strong evidence for that; I'm just saying that
rlm@57: it might turn out that way, partly because I think we know less than
rlm@57: many people think we know about what biological evolution achieved.
rlm@57:
rlm@57: [32:28] There are some other possibilities: we may just find out that
rlm@57: there are shortcuts no one ever thought of, and it will all happen
rlm@57: much more quickly---I have an open mind; I'd be surprised, but it
rlm@57: could turn up. There /is/ something that worries me much more than the
rlm@57: singularity that most people talk about, which is machines achieving
rlm@57: human-level intelligence and perhaps taking over [the] planet or
rlm@57: something. There's what I call the /singularity of cognitive
rlm@57: catch-up/ ...
rlm@57:
rlm@57: * A singularity of cognitive catch-up
rlm@57:
rlm@57: ** What if it will take a lifetime to learn enough to make something new?
rlm@57: ... SCC, singularity of cognitive catch-up, which I think we're close
rlm@57: to, or maybe have already reached---I'll explain what I mean by
rlm@57: that. One of the products of biological evolution---and this is one of
rlm@57: the answers to your earlier questions which I didn't get on to---is
rlm@57: that humans have not only the ability to make discoveries that none of
rlm@57: their ancestors have ever made, but to shorten the time required for
rlm@57: similar achievements to be reached by their offspring and their
rlm@57: descendants. So once we, for instance, worked out ways of complex
rlm@57: computations, or ways of building houses, or ways of finding our way
rlm@57: around, we don't need...our children don't need to work it out for
rlm@57: themselves by the same lengthy trial and error procedure; we can help
rlm@57: them get there much faster.
rlm@57:
rlm@57: Okay, well, what I've been referring to as the singularity of
rlm@57: cognitive catch-up depends on the fact that---fairly obvious, and it's
rlm@57: often been commented on---that in case of humans, it's not necessary
rlm@57: for each generation to learn what previous generations learned /in the
rlm@57: same way/. And we can speed up learning once something has been
rlm@57: learned, [it is able to] be learned by new people. And that has meant
rlm@57: that the social processes that support that kind of education of the
rlm@57: young can enormously accelerate what would have taken...perhaps
rlm@57: thousands [or] millions of years for evolution to produce, can happen in
rlm@57: a much shorter time.
rlm@57:
rlm@57:
rlm@57: [34:54] But here's the catch: in order for a new advance to happen ---
rlm@57: so for something new to be discovered that wasn't there before, like
rlm@57: Newtonian mechanics, or the theory of relativity, or Beethoven's music
rlm@57: or [style] or whatever --- the individuals have to have traversed a
rlm@57: significant amount of what their ancestors have learned, even if they
rlm@57: do it much faster than their ancestors, to get to the point where they
rlm@57: can see the gaps, the possibilities for going further than their
rlm@57: ancestors, or their parents or whatever, have done.
rlm@57:
rlm@57: [35:27] Now in the case of knowledge of science, mathematics,
rlm@57: philosophy, engineering and so on, there's been a lot of accumulated
rlm@57: knowledge. And humans are living a /bit/ longer than they used to, but
rlm@57: they're still living for [whatever it is], a hundred years, or for
rlm@57: most people, less than that. So you can imagine that there might come
rlm@57: a time when in a normal human lifespan, it's not possible for anyone
rlm@57: to learn enough to understand the scope and limits of what's already
rlm@57: been achieved in order to see the potential for going beyond it and to
rlm@57: build on what's already been done to make that...those future steps.
rlm@57:
rlm@57: [36:10] So if we reach that stage, we will have reached the
rlm@57: singularity of cognitive catch-up because the process of education
rlm@57: that enables individuals to learn faster than their ancestors did is
rlm@57: the catching-up process, and it may just be that we at some point
rlm@57: reach a point where catching up can only happen within a lifetime of
rlm@57: an individual, and after that they're dead and they can't go
rlm@57: beyond. And I have some evidence that there's a lot of that around
rlm@57: because I see a lot of people coming up with what /they/ think of as
rlm@57: new ideas which they've struggled to come up with, but actually they
rlm@57: just haven't taken in some of what was...some of what was done [] by
rlm@57: other people, in other places before them. And I think that despite
rlm@57: the availability of search engines which make it /easier/ for people
rlm@57: to get the information---for instance, when I was a student, if I
rlm@57: wanted to find out what other people had done in the field, it was a
rlm@57: laborious process---going to the library, getting books, and
rlm@57: ---whereas now, I can often do things in seconds that would have taken
rlm@57: hours. So that means that if seconds [are needed] for that kind of
rlm@57: work, my lifespan has been extended by a factor of ten or
rlm@57: something. So maybe that /delays/ the singularity, but it may not
rlm@57: delay it enough. But that's an open question; I don't know. And it may
rlm@57: just be that in some areas, this is more of a problem than others. For
rlm@57: instance, it may be that in some kinds of engineering, we're handing
rlm@57: over more and more of the work to machines anyways and they can go on
rlm@57: doing it. So for instance, most of the production of computers now is
rlm@57: done by a computer-controlled machine---although some of the design
rlm@57: work is done by humans--- a lot of /detail/ of the design is done by
rlm@57: computers, and they produce the next generation, which then produces
rlm@57: the next generation, and so on.
rlm@57:
rlm@57: [37:57] I don't know if humans can go on having major advances, so
rlm@57: it'll be kind of sad if we can't.
rlm@57:
rlm@57: * Spatial reasoning: a difficult problem
rlm@57:
rlm@57: [38:15] Okay, well, there are different problems [ ] mathematics, and
rlm@57: they have to do with properties. So for instance a lot of mathematics
rlm@57: that can be expressed in terms of logical structures or algebraic
rlm@57: structures and those are pretty well suited for manipulation and...on
rlm@57: computers, and if a problem can be specified using the
rlm@57: logical/algebraic notation, and the solution method requires creating
rlm@57: something in that sort of notation, then computers are pretty good,
rlm@57: and there are lots of mathematical tools around---there are theorem
rlm@57: provers and theorem checkers, and all kinds of things, which couldn't
rlm@57: have existed fifty, sixty years ago, and they will continue getting
rlm@57: better.
rlm@57:
rlm@57:
rlm@57: But there was something that I was [[example-gap][alluding to earlier]] when I gave the
rlm@57: example of how you can reason about what you will see by changing your
rlm@57: position in relation to a door, where what you are doing is using your
rlm@57: grasp of spatial structures and how as one spatial relationship
rlm@57: changes namely you come closer to the door or move sideways and
rlm@57: parallel to the wall or whatever, other spatial relationships change
rlm@57: in parallel, so the lines from your eyes through to other parts of
rlm@57: the...parts of the room on the other side of the doorway change,
rlm@57: spread out more as you go towards the doorway, and as you move
rlm@57: sideways, they don't spread out differently, but focus on different
rlm@57: parts of the internal ... that they access different parts of the
rlm@57: ... of the room.
rlm@57:
rlm@57: Now, those are examples of ways of thinking about relationships and
rlm@57: changing relationships which are not the same as thinking about what
rlm@57: happens if I replace this symbol with that symbol, or if I substitute
rlm@57: this expression in that expression in a logical formula. And at the
rlm@57: moment, I do not believe that there is anything in AI amongst the
rlm@57: mathematical reasoning community, the theorem-proving community, that
rlm@57: can model the processes that go on when a young child starts learning
rlm@57: to do Euclidean geometry and is taught things about---for instance, I
rlm@57: can give you a proof that the angles of any triangle add up to a
rlm@57: straight line, 180 degrees.
rlm@57:
rlm@57: ** Example: Spatial proof that the angles of any triangle add up to a half-circle
rlm@57: There are standard proofs which involves starting with one triangle,
rlm@57: then adding a line parallel to the base one of my former students,
rlm@57: Mary Pardoe, came up with which I will demonstrate with this --- can you see it? If I have a triangle here that's got
rlm@57: three sides, if I put this thing on it, on one side --- let's say the
rlm@57: bottom---I can rotate it until it lies along the second...another
rlm@57: side, and then maybe move it up to the other end ~. Then I can rotate
rlm@57: it again, until it lies on the third side, and move it back to the
rlm@57: other end. And then I'll rotate it again and it'll eventually end up
rlm@57: on the original side, but it will have changed the direction it's
rlm@57: pointing in --- and it won't have crossed over itself so it will have
rlm@57: gone through a half-circle, and that says that the three angles of a
rlm@57: triangle add up to the rotations of half a circle, which is a
rlm@57: beautiful kind of proof and almost anyone can understand it. Some
rlm@57: mathematicians don't like it, because they say it hides some of the
rlm@57: assumptions, but nevertheless, as far as I'm concerned, it's an
rlm@57: example of a human ability to do reasoning which, once you've
rlm@57: understood it, you can see will apply to any triangle --- it's got to
rlm@57: be a planar triangle --- not a triangle on a globe, because then the
rlm@57: angles can add up to more than ... you can have three /right/ angles
rlm@57: if you have an equator...a line on the equator, and a line going up to
rlm@57: to the north pole of the earth, and then you have a right angle and
rlm@57: then another line going down to the equator, and you have a right
rlm@57: angle, right angle, right angle, and they add up to more than a
rlm@57: straight line. But that's because the triangle isn't in the plane,
rlm@57: it's on a curved surface. In fact, that's one of the
rlm@57: differences...definitional differences you can take between planar and
rlm@57: curved surfaces: how much the angles of a triangle add up to. But our
rlm@57: ability to /visualize/ and notice the generality in that process, and
rlm@57: see that you're going to be able to do the same thing using triangles
rlm@57: that stretch in all sorts of ways, or if it's a million times as
rlm@57: large, or if it's made...you know, written on, on...if it's drawn in
rlm@57: different colors or whatever --- none of that's going to make any
rlm@57: difference to the essence of that process. And that ability to see
rlm@57: the commonality in a spatial structure which enables you to draw some
rlm@57: conclusions with complete certainty---subject to the possibility that
rlm@57: sometimes you make mistakes, but when you make mistakes, you can
rlm@57: discover them, as has happened in the history of geometrical theorem
rlm@57: proving. Imre Lakatos had a wonderful book called [[http://en.wikipedia.org/wiki/Proofs_and_Refutations][/Proofs and
rlm@57: Refutations/]] --- which I won't try to summarize --- but he has
rlm@57: examples: mistakes were made; that was because people didn't always
rlm@57: realize there were subtle subcases which had slightly different
rlm@57: properties, and they didn't take account of that. But once they're
rlm@57: noticed, you rectify that.
rlm@57:
rlm@57: ** Geometric results are fundamentally different than experimental results in chemistry or physics.
rlm@57: [43:28] But it's not the same as doing experiments in chemistry and
rlm@57: physics, where you can't be sure it'll be the same on [] or at a high
rlm@57: temperature, or in a very strong magnetic field --- with geometric
rlm@57: reasoning, in some sense you've got the full information in front of
rlm@57: you; even if you don't always notice an important part of it. So, that
rlm@57: kind of reasoning (as far as I know) is not implemented anywhere in a
rlm@57: computer. And most people who do research on trying to model
rlm@57: mathematical reasoning, don't pay any attention to that, because of
rlm@57: ... they just don't think about it. They start from somewhere else,
rlm@57: maybe because of how they were educated. I was taught Euclidean
rlm@57: geometry at school. Were you?
rlm@57:
rlm@57: (Adam ford: Yeah)
rlm@57:
rlm@57: Many people are not now. Instead they're taught set theory, and
rlm@57: logic, and arithmetic, and [algebra], and so on. And so they don't use
rlm@57: that bit of their brains, without which we wouldn't have built any of
rlm@57: the cathedrals, and all sorts of things we now depend on.
rlm@57:
rlm@57: * Is near-term artificial general intelligence likely?
rlm@57:
rlm@57: ** Two interpretations: a single mechanism for all problems, or many mechanisms unified in one program.
rlm@57:
rlm@57: [44:35] Well, this relates to what's meant by general. And when I
rlm@57: first encountered the AGI community, I thought that what they all
rlm@57: meant by general intelligence was /uniform/ intelligence ---
rlm@57: intelligence based on some common simple (maybe not so simple, but)
rlm@57: single powerful mechanism or principle of inference. And there are
rlm@57: some people in the community who are trying to produce things like
rlm@57: that, often in connection with algorithmic information theory and
rlm@57: computability of information, and so on. But there's another sense of
rlm@57: general which means that the system of general intelligence can do
rlm@57: lots of different things, like perceive things, understand language,
rlm@57: move around, make things, and so on --- perhaps even enjoy a joke;
rlm@57: that's something that's not nearly on the horizon, as far as I
rlm@57: know. Enjoying a joke isn't the same as being able to make laughing
rlm@57: noises.
rlm@57:
rlm@57: Given, then, that there are these two notions of general
rlm@57: intelligence---there's one that looks for one uniform, possibly
rlm@57: simple, mechanism or collection of ideas and notations and algorithms,
rlm@57: that will deal with any problem that's solvable --- and the other
rlm@57: that's general in the sense that it can do lots of different things
rlm@57: that are combined into an integrated architecture (which raises lots
rlm@57: of questions about how you combine these things and make them work
rlm@57: together) and we humans, certainly, are of the second kind: we do all
rlm@57: sorts of different things, and other animals also seem to be of the
rlm@57: second kind, perhaps not as general as humans. Now, it may turn out
rlm@57: that in some near future time, who knows---decades, a few
rlm@57: decades---you'll be able to get machines that are capable of solving
rlm@57: in a time that will depend on the nature of the problem, but any
rlm@57: problem that is solvable, and they will be able to do it in some sort
rlm@57: of tractable time --- of course, there are some problems that are
rlm@57: solvable that would require a larger universe and a longer history
rlm@57: than the history of the universe, but apart from that constraint,
rlm@57: these machines will be able to do anything []. But to be able to do
rlm@57: some of the kinds of things that humans can do, like the kinds of
rlm@57: geometrical reasoning where you look at the shape and you abstract
rlm@57: away from the precise angles and sizes and shapes and so on, and
rlm@57: realize there's something general here, as must have happened when our
rlm@57: ancestors first made the discoveries that eventually put together in
rlm@57: Euclidean geometry.
rlm@57:
rlm@57: It may be that that requires mechanisms of a kind that we don't know
rlm@57: anything about at the moment. Maybe brains are using molecules and
rlm@57: rearranging molecules in some way that supports that kind of
rlm@57: reasoning. I'm not saying they are --- I don't know, I just don't see
rlm@57: any simple...any obvious way to map that kind of reasoning capability
rlm@57: onto what we currently do on computers. There is---and I just
rlm@57: mentioned this briefly beforehand---there is a kind of thing that's
rlm@57: sometimes thought of as a major step in that direction, namely you can
rlm@57: build a machine (or a software system) that can represent some
rlm@57: geometrical structure, and then be told about some change that's going
rlm@57: to happen to it, and it can predict in great detail what'll
rlm@57: happen. And this happens for instance in game engines, where you say
rlm@57: we have all these blocks on the table and I'll drop one other block,
rlm@57: and then [the thing] uses Newton's laws and properties of rigidity of
rlm@57: the parts and the elasticity and also stuff about geometries and space
rlm@57: and so on, to give you a very accurate representation of what'll
rlm@57: happen when this brick lands on this pile of things, [it'll bounce and
rlm@57: go off, and so on]. And you just, with more memory and more CPU power,
rlm@57: you can increase the accuracy--- but that's totally different than
rlm@57: looking at /one/ example, and working out what will happen in a whole
rlm@57: /range/ of cases at a higher level of abstraction, whereas the game
rlm@57: engine does it in great detail for /just/ this case, with /just/ those
rlm@57: precise things, and it won't even know what the generalizations are
rlm@57: that it's using that would apply to others []. So, in that sense, [we]
rlm@57: may get AGI --- artificial general intelligence --- pretty soon, but
rlm@57: it'll be limited in what it can do. And the other kind of general
rlm@57: intelligence which combines all sorts of different things, including
rlm@57: human spatial geometrical reasoning, and maybe other things, like the
rlm@57: ability to find things funny, and to appreciate artistic features and
rlm@57: other things may need forms of pattern-mechanism, and I have an open
rlm@57: mind about that.
rlm@57:
rlm@57: * Abstract General Intelligence impacts
rlm@57:
rlm@57: [49:53] Well, as far as the first type's concerned, it could be useful
rlm@57: for all kinds of applications --- there are people who worry about
rlm@57: where there's a system that has that type of intelligence, might in
rlm@57: some sense take over control of the planet. Well, humans often do
rlm@57: stupid things, and they might do something stupid that would lead to
rlm@57: disaster, but I think it's more likely that there would be other
rlm@57: things [] lead to disaster--- population problems, using up all the
rlm@57: resources, destroying ecosystems, and whatever. But certainly it would
rlm@57: go on being useful to have these calculating devices. Now, as for the
rlm@57: second kind of them, I don't know---if we succeeded at putting
rlm@57: together all the parts that we find in humans, we might just make an
rlm@57: artificial human, and then we might have some of them as your friends,
rlm@57: and some of them we might not like, and some of them might become
rlm@57: teachers or whatever, composers --- but that raises a question: could
rlm@57: they, in some sense, be superior to us, in their learning
rlm@57: capabilities, their understanding of human nature, or maybe their
rlm@57: wickedness or whatever --- these are all issues in which I expect the
rlm@57: best science fiction writers would give better answers than anything I
rlm@57: could do, but I did once fantasize when I [back] in 1978, that perhaps
rlm@57: if we achieved that kind of thing, that they would be wise, and gentle
rlm@57: and kind, and realize that humans are an inferior species that, you
rlm@57: know, have some good features, so they'd keep us in some kind of
rlm@57: secluded...restrictive kind of environment, keep us away from
rlm@57: dangerous weapons, and so on. And find ways of cohabitating with
rlm@57: us. But that's just fantasy.
rlm@57:
rlm@57: Adam Ford: Awesome. Yeah, there's an interesting story /With Folded
rlm@57: Hands/ where [the computers] want to take care of us and want to
rlm@57: reduce suffering and end up lobotomizing everybody [but] keeping them
rlm@57: alive so as to reduce the suffering.
rlm@57:
rlm@57: Aaron Sloman: Not all that different from /Brave New World/, where it
rlm@57: was done with drugs and so on, but different humans are given
rlm@57: different roles in that system, yeah.
rlm@57:
rlm@57: There's also /The Time Machine/, H.G. Wells, where the ... in the
rlm@57: distant future, humans have split in two: the Eloi, I think they were
rlm@57: called, they lived underground, they were the [] ones, and then---no,
rlm@57: the Morlocks lived underground; Eloi lived on the planet; they were
rlm@57: pleasant and pretty but not very bright, and so on, and they were fed
rlm@57: on by ...
rlm@57:
rlm@57: Adam Ford: [] in the future.
rlm@57:
rlm@57: Aaron Sloman: As I was saying, if you ask science fiction writers,
rlm@57: you'll probably come up with a wide variety of interesting answers.
rlm@57:
rlm@57: Adam Ford: I certainly have; I've spoken to [] of Birmingham, and
rlm@57: Sean Williams, ... who else?
rlm@57:
rlm@57: Aaron Sloman: Did you ever read a story by E.M. Forrester called /The
rlm@57: Machine Stops/ --- very short story, it's [[http://archive.ncsa.illinois.edu/prajlich/forster.html][on the Internet somewhere]]
rlm@57: --- it's about a time when people sitting ... and this was written in
rlm@57: about [1914 ] so it's about...over a hundred years ago ... people are
rlm@57: in their rooms, they sit in front of screens, and they type things,
rlm@57: and they communicate with one another that way, and they don't meet;
rlm@57: they have debates, and they give lectures to their audiences that way,
rlm@57: and then there's a woman whose son says \ldquo{}I'd like to see
rlm@57: you\rdquo{} and she says \ldquo{}What's the point? You've got me at
rlm@57: this point \rdquo{} but he wants to come and talk to her --- I won't
rlm@57: tell you how it ends, but.
rlm@57:
rlm@57: Adam Ford: Reminds me of the Internet.
rlm@57:
rlm@57: Aaron Sloman: Well, yes; he invented ... it was just extraordinary
rlm@57: that he was able to do that, before most of the components that we
rlm@57: need for it existed.
rlm@57:
rlm@57: Adam Ford: [Another person who did that] was Vernor Vinge [] /True
rlm@57: Names/.
rlm@57:
rlm@57: Aaron Sloman: When was that written?
rlm@57:
rlm@57: Adam Ford: The seventies.
rlm@57:
rlm@57: Aaron Sloman: Okay, well a lot of the technology was already around
rlm@57: then. The original bits of internet were working, in about 1973, I was
rlm@57: sitting ... 1974, I was sitting at Sussex University trying to
rlm@57: use...learn LOGO, the programming language, to decide whether it was
rlm@57: going to be useful for teaching AI, and I was sitting [] paper
rlm@57: teletype, there was paper coming out, transmitting ten characters a
rlm@57: second from Sussex to UCL computer lab by telegraph cable, from there
rlm@57: to somewhere in Norway via another cable, from there by satellite to
rlm@57: California to a computer Xerox [] research center where they had
rlm@57: implemented a computer with a LOGO system on it, with someone I had
rlm@57: met previously in Edinburgh, Danny Bobrow, and he allowed me to have
rlm@57: access to this sytem. So there I was typing. And furthermore, it was
rlm@57: duplex typing, so every character I typed didn't show up on my
rlm@57: terminal until it had gone all the way there and echoed back, so I
rlm@57: would type, and the characters would come back four seconds later.
rlm@57:
rlm@57: [55:26] But that was the Internet, and I think Vernor Vinge was
rlm@57: writing after that kind of thing had already started, but I don't
rlm@57: know. Anyway.
rlm@57:
rlm@57: [55:41] Another...I mentioned H.G. Wells, /The Time Machine/. I
rlm@57: recently discovered, because [[http://en.wikipedia.org/wiki/David_Lodge_(author)][David Lodge]] had written a sort of
rlm@57: semi-novel about him, that he had invented Wikipedia, in advance--- he
rlm@57: had this notion of an encyclopedia that was free to everybody, and
rlm@57: everybody could contribute and [collaborate on it]. So, go to the
rlm@57: science fiction writers to find out the future --- well, a range of
rlm@57: possible futures.
rlm@57:
rlm@57: Adam Ford: Well the thing is with science fiction writers, they have
rlm@57: to maintain some sort of interest for their readers, after all the
rlm@57: science fiction which reaches us is the stuff that publishers want to
rlm@57: sell, and so there's a little bit of a ... a bias towards making a
rlm@57: plot device there, and so the dramatic sort of appeals to our
rlm@57: amygdala, our lizard brain; we'll sort of stay there obviously to some
rlm@57: extent. But I think that they do come up with sort of amazing ideas; I
rlm@57: think it's worth trying to make these predictions; I think that we
rlm@57: should more time on strategic forecasting, I mean take that seriously.
rlm@57:
rlm@57: Aaron Sloman: Well, I'm happy to leave that to others; I just want to
rlm@57: try to understand these problems that bother me about how things
rlm@57: work. And it may be that some would say that's irresponsible if I
rlm@57: don't think about what the implications will be. Well, understanding
rlm@57: how humans work /might/ enable us to make [] humans --- I suspect it
rlm@57: wont happen in this century; I think it's going to be too difficult.