Transcript for:
Keynote Insights on Digital Design Evolution

all right thank thanks everybody for joining I know it's been a long day for many who have actually stayed with us from the beginning but a lot of the work that the students have been working on and developing and researching and speculating and problematizing very much obviously has a lineage and a history and I think a big part of the way that we see research within the DRL is a kind of evolving practice particularly very design oriented and very much trying to explore systems computation the problems of materiality trying to think through time there are a lot of conditions and constraints in terms of working with things and at the same time we have a very proactive attitude towards looking at new tools ideas of information and space that are very much part of the everyday of today and to use that than to speculate ok there is a bit of a nice experiment here it will fall and crash during the lecture it is a bit of a ploy for John to see if it will distract him I put a wager that it won't as part of the events as this is day one of a two-day thing we have what we call the DRL keynote lectures and these things are really very important for us because aside about presenting the work and discussing it in sort of conversational form it's really to bring I think people that we have admiration and respect to continue the conversation to help shape I think the discourse that is happening with the work that is shared between the tutors the students and usually are working over three or four year speculative agendas that we sort of set up and this is the conclusion and the beginning I think of the second year of work behavioral complexity now John I don't think needs much of an introduction within this space John and Julia people like Cedric price and Gordon Pasque were really trying to tease out certain aspects from the 60s and 70s 80s and into the mid 90s about the role of computation I think a lot of those early experiments thought experiments discourses about cybernetics trying to explore kind of new tools and how that would refigure and reconstitute the role of the architect has been I think part of our conversation and I think continues and I'm very pleased that John is here to be part of that continuation as well as Julia and X students of John's and Julia's and so forth and with that I think I'm not going to give any more of a kind of glorified conversation on my side but I think it's really important that I think if research is supposed to really matter and practice and design is really a big part of the way that we sort of evolve projects and it's important that the history of this is very much embedded as a continuation of this kind of experiment and I think from John I've learned personally some very particular aspects particularly he made a statement once that the role of computation is basically to allow us to change how we habitually work through things it's not necessarily about computation or the literal tools but it's really about the intellectual discourse that is as much a part of the way that we think problematize and work with these things as it is in terms of how we embed and push the kind of discourse intellectual project and material realization so with that I would really like you to warmly welcome John back to the AAA though he's never really left in some capacity Thank You CEO I'll start with your remark you've just made yes we used to do a thought experiment called computing without computers this was partly to do what Theo's just said which is to try to clarify the fundamental issues underlying computation and how it would let us see the things we did maybe without computers better it was also because we hadn't got any computers or me started the a didn't have a single machine subsequently I slightly changed my mind I come to the conclusion this wasn't the question which is computing without computer it was a statement it really is best to do without them and just think what it is you could do if you had this immensely powerful machine which could and would do everything the question is what would it do for you and the continuing experiment that's what the AAA is about this is James gone from about one hundred and twenty-fifth anniversary a 1970s continuing experiment and when Theo asked me to write a little piece for his adapted the college's book and I picked on James garland the idea of a continuing experiment and how all these things fit together and Theo's kindly just suggested that some of what's going on Mayo a little bit of allegiance and and I will spend the first 15 or minutes for 20 minutes or so and doing a little bit of nostalgic self-indulgent reminiscing about what we did do and then I'll move on to what is now going on and what happens subsequently and try to end by putting together what the total theoretical picture looks like now as best I can explain it but we're here today to celebrate this event and I must say how very much I have enjoyed today and how I have indeed myself learnt things from it and it stimulated my thinking but going back to how this started and I had taught at the AAA before in the seventh and I'd left to go and do something in Ireland and and Alvin suddenly rang me out of the blue and said would I come back and run a unit again now he got fabulous units at that time they were so hard there was wrong there was everybody you can think of RAM and and I said well what he said well he said it's all great but he said it's kind of like all the same and he said can you come and do something different so I tried in the two years I tried immediate projects like an exploratory of space and time trying to push a new agenda with Melissa and with Cedric and basically it failed but let me just say a couple of things happened in that proprietary period first of all we set up a unit tutorial space in some Katharine's dock and onboard watchmaker and to do that that's the that's the logbook from watchmaker you can probably barely see the printing but here for example Cedric Paul Finch Nick Grimshaw is he met Stein and a party with Alvin and Peter Cook and hazel and everyone were there and that's how the boat came to be in st. Katharine's dock and still is that's where Julia lives during the week while she's running the computing computing here but it was a vital base and the point was that every Friday morning and Cedric and I took it in turns to invite a guest and every morning at 8 o'clock this is Pauline revenge at students wanting to come in at 10 or 11 or 12 and the interesting thing was in all that time none of our guests ever made a comment about the fact it started to date as since we had champagne and salmon and spreadable eggs at that time it was amazing that students turned up so that was great but at the end of the two years the problem was oh no before I've gone we did a few preparatory things in that time one of them was we went to visit Ron Holland this is a yacht designer in Southern Ireland and Ron works with the remarkable balance between his intuition his skill and judgment from his experience there you see him eyeballing the lines of a boat the splaining of the boat but he combined that with the most advanced computer techniques and available integral systems all tucked away in a settle college and cabin and island and from them he was able to do the difficult bits like the calculations displacement water line length and everything and produce the drawings from which the instructions went off to cut the aluminium profiles in a German shipyard and then construct the boat so in effect Ron had become a kind of digital craftsman he regained that control of the manufacturing process that we lost when we took to doing drawings Architects lost it when it joins one of the nice things about today is these robots for example have got back that control of the total construction process and I think that's very important so lessons some people like Ron Holland we were interested in everything from a computer-aided design program right through to the CNC and manufacturing and that was a beam this year but anyway I felt that we weren't changing anything and I went to Alvin and I resigned and said well the problem was that I thought I was running interesting projects but if you look at the end of year exhibitions they're exactly like everybody else's work because the students go around copying each other and whatever is fashionable gets replicated right the way through the exhibition spaces and Alvin didn't like that remark of course but he did have to concede it was true said well what would I do to change it so I explained how I would run a radically different kind of unit where we didn't do any drawings we didn't do any kind of buildings we did basically research and answer why did you do it and I said well we wouldn't get any students we wouldn't get the numbers we would never get it past the external examiner's anyway the a a hasn't got any computers so Alvin said well okay I'll um we'll run the unit under numbers I'll subsidize it and I'll give you five years because it'll take you that long to get this going I'll give you five years to try to show that you can do an experiment and I'll give you the most difficult as we got here in the first year was Peter Carlin Van Schaick and you you you'll be able to persuade them and that left the problem of the computer so Julia and I had a software company at the time so we simply gave the AAA a lot of computers and plotters so then we started this new unit it started this unit 11 it started as 14 initially and it was to do with morphogenesis basically and we still had cedric asseh an advisor and gordon task and we got Pete silver who is at the time one of the students a guy who'd also been one of the students and in his technical tutor some we've involved started this experiment so you know one Asia had a different title the O one was to do with rule-based generative systems and we've built a universal constructor which is this machine it's a reconfigurable three-dimensional array processor consisting of a 12 by 12 by 12 matrix of tubes each one of which contains a microprocessor chip eight LEDs showing the state of this space it's a 3d cellular automata except reconfigurable it's an input and output device it's input because you can rearrange the cubes to set the computer problem and it's an output device lesea and little LED lights it flashed at you if it wanted you to interact we didn't have any robots but anyway I quite liked the idea that it would solicit human interaction so that's what it did that was the output side and the the unit designed and built all the tubes the cubes are uh stamps money boxes for children they can see the little cheap that we had no money you see it's a problem but we managed to build it and it worked and it is that Helen famously said when he saw this Ronnie said well he says I suppose the first 200 million years of evolution were pretty slow he said it would be quite an exciting time to be arranged I shouldn't think he said it was boring but and then we also right from the start n intended to automate this and all the proved is a provision for putting motors throughout that whole system in the event only a small bit of this got built but these are little stepper motors dry being an automatic spline eMachine that year got its first honours from that wall none of me got a succession of honest and then amin I'm pleased to say which helped to recruit students so that was our end of first year exhibition with this object in Drey so there were no drawings there were no buildings so for the first time the AAA had finally got rid of all now to everyone in this room now it seems so normal you look at all this work we've been seen from DRL question is for a minute that the fact that may we see the plans of a house where there's no sign of the lose of the bathroom which anyway you discover is tipping and anything else and this seems perfectly reasonable implausible now but it then wasn't it was outrageous so indeed we did have some difficulty with the external examiner's but they were they were both very clever intelligent interesting people and they just argued and argued now good with us and said yes so the next year we went on to explore responsive systems in particular neural networks feedback trying to get beyond the algorithmic approach we built erm or one of the students not miles built her neural network that year an attempt to get beyond serial computing also some of the students that Stefan was still working on last year's project he'd gone off to Colombia and and he had this sequence evolves just from a particular configuration of those cubes it's a 3d Center automata but instead of having just alive or dead like a life game each cell in those days because there was eight LEDs had 256 states so those states map to things like GM I'm sorry sir map to things like geometrical transformations so it starts off by deforming stretching rotating in in response to the digital number of that space redefine redefine architecture as simply being logic in space at this point third year we went to antenna every student built an implant an annex in ten are they all interacting so we had bodysuits it's electric grass and interactive jewelry we also built um we always have a hands-on thing so we built for example the solar protractor I'm Ken yang commissioned a unit to do some work with solar angles and this is just a way of understanding for any point anywhere in the planet any particular latitude any day of the year any time of the day exactly where the Sun would be not of course linked through two computer programs but because we always felt that if you there's no point having a computer program until you could probably properly understand the Sun path from your own experience also that year akiro managed to combine for the first time a cellular automata and a genetic algorithm you'll find by the time I get to the end of our talk actually I'm very uncomfortable with any of these techniques used in isolation I don't think you start to get the real power of any of these systems until you start to have something like a learning program Genting algorithm running with a generative system like a cellular automata now return to that point but a cheero simulated the response of a structure to various environmental factors because always we had to put the project in and environmental and social context this is very important for the unity wasn't just a research in abstract it was research in response to real environmental and social problems and so these were envelopes of possible structural opportunities within that environment at the end of that year we built the universal interactor which was all the different students input and output devices all interacting in one space in the north jury room and again of course no drawings this is the first time he had an exhibition where absolutely everything worked and communicated and interacted is this about to collapse I'm watching with some interest and not because it would distract knees had just I'm just as fascinated to everybody else to know when it's gonna fail so the there were sound systems this is will McClane now and will with with an interactive sound systems body suits a suit gives you a massage Nicolas in the audience somewhere Nicola and Sweeting there's an input device you touch this and it and it and it senses you and there's another one shoe wear which provides someone else somewhere remotely it can be of course with the with the stimulation that you're providing that idea amongst many other this has been copied all over the all over the planet it was Sweeting with pizza electric grass which responds to our movement wind movement over a little kid so grass structures and all these systems are interconnected so you could plug that into the massage chute for example and be massaged by the wind there was an antenna and Hadi um this is a Fri auto structure basically but round the base here were twelve extremely powerful electromagnets so this thing thrashed about in this cage and it really was dangerous and it used to go into different moods since the exhibition knew if there was anyone in it and where they were if it came closer to do it close to the cages started a rattle about to attract your attention and show off if you're in the distance he might just tap the cage if you walked out of the exhibition without looking at it here trash around or wealth into assault it eventually trashed itself to pieces two weeks into the exhibition so we assumed it was getting I'm pretty frustrated with the audience reaction and but but but because everything was interconnected people as they went in and started playing with one thing or is very surprised sorry to discover that it was then causing something else to interact this is Nicola again with jewelry question is how you're gonna control this is basically a 3d cursor controller built into into jewelry and that was controlling this thing which we call our neck little bendy ball um I don't have a decent photograph of here but you can see it here twisting itself into different shapes like an elephant's trunk really but it was all being controlled with these varible resistors in the jewelry but the most extraordinary thing that year was this discovery which is to do with thought replication we were using close packing the geometries of close packing systems as our fundamental database as a way of structuring a data a metal I noticed something quite extraordinary there are two ways of packing I'm sure most people in the room know that shooting pack alternate layers aibee aibee aibee or you can pack ABC ABC ABC but you could combine them but so there's normal closed packing in this case some ball bearings here is a fault line that line has been induced by displacing some of those summer packing a b a b a b and summer having ABC ABC ABC this is the notion that information is dislocations in the system otherwise you have white noise there is no information until as a disturbance - there you go on time the third layer though here we go is if the pair's there's the fault line the next layer packed in repairs the fault line which is the most extraordinary discovered and we wrote to all the crystallographers that we could find who are interested knew anything about this they all said they had no idea never seen this before fascinating but this was the real shock if you build a fourth layer the fault reappears so what in fact we've got here is a physical idea a physical manifestation of a self-replicating fault system a suffer a beginning system and the question always arose okay we understand about evolution but where did you evolve from what was there before DNA and Richard Dawkins posed this question in the blind watchmaker what was the press the press the president for The Selfish Gene and what was it that was driving to replicate and he suggested perhaps it in clay and suggested this person called Ken Smith who had said this Professor Ken Smith and can Smith when he saw this was absolutely over the moon because he said here I've got a real demonstration of a self-replicating system in a mineral prior to the chemistry of of DNA and of course DNA is a crystal structure so this is not at all surprising Nicola cast that in rubber that's the fault line there and there's the fault picked out so you may be looking there at a very early ancestor some 500 million years ago in the fourth year Oh Nicolo got honours for that by the way and the fourth year we got to form generations so the geometry are we do what he had that and other folding was very fashionable at the time and I was greatly amused to watch all these units struggling to put folded structures into the computers with the dreadful interfaces which were available in 1992 so Tim as a bit of a piss-take basically built this blanket with cells like a trap and light-sensitive cells so simply all you had to do was fold and manipulate or blanket it manipulated these folding structures this was done purely as that wind up to the units who were into this kind of thing but it did make quite a serious point about um about tangible interfaces and about interfaces that went beyond scripting and I mean I said before if all we've done is replaced drawing by scripting with a typewriter what have we achieved I mean nothing is ridiculous and it's not the way to do it and that year we built an interactive structure with light-sensitive sensors on its a little little tensegrities structure here and as the light from the Sun some in the exhibition space there was a powerful light swinging backwards and forwards on a motor and as that light impinged on the structure the structure deformed and then that was projected onto a screen across the entrance to to North curity room so as you pass this you saw this structure moving and changing and responding and of course drew people in to find out why this was happening and it was just simply this structure here responding to the light changing by year five we got to the notion of a genetic language of architecture for talking about genetics how do we describe them we needed a language we also demonstrated hierarchical cellular automata and multi-state readings automata started to go into the scientific literature I'm gonna keep referring back to this point because I'm one of the things I played those are the calling and leap in the library here of the lecture I gave 20 years ago to the day more or less and I said that there was a secret agenda which was I would like we were borrowing all these ideas from science I wanted to pay something back to science so I got very excited when I started to discover that scientific journals were starting to publish serious reviews and discussions of what what we were doing this is the first of several examples and there are now something like 20,000 scientific publications that refer to the work of the unit during that time but this was quite fun just sort of diversity and Chi and had a 2d cellular automata running on a torus in a flat plane you know obviously for 2d set automata you can pretend you joined the two sides of the screen of the top and the bottom this case he just made it into a torus and the nice thing was that it it had a kind of beginning and this thing would kind of each itself up so the music kind of formed a natural climax I'm sorry then if the tape is quite nice and this 3d demonstration by MANET Rastogi who was one of the first to fully show something moving from a single cell in this case a cube to various structures so I just run a little bit of this clip I catch my breath just like the life game this 3d cell is looking at the cells around it for its neighbors only instead of having just alive or dead or being about to be born there were some I didn't have 16 million states by this point it's looking at these states including invisible environment states here and they're discussing on the basis of that how to construct the next part we're all born as single cells and those single cells start to split and split and split and after only about 50 splittings it turns into you and me and the rest of it in the room or the trees at Bedford square always amuses me that that about a third of our DNA string is identical to the trees out there so we're all all the cells are initially identical but as we develop we start to get more specialist cells a certain point the cells in our eyes for example are different and the same happens here after a particular point this thing reaches a sort of stasis and then starts to generate differentiated cells what time's dinner here anyway starting now it's starting to change its geometries change its color and turn itself into something where the cells are more specialists but but it also had a genetic algorithm so not only was that the output of that particular run but these are output some other runs so in other words we got past the idea that we're likely to have some kind of merely permutations on a fixed outcome by this point got to the stage where things really could evolve into some radically different kinds of structure by year six we built the inter activator and the exhibition which is what I'm personally celebrating with my unit colleagues tonight thirty years ago we built an exhibition in the exhibition space there and it had the first exhibition where something evolved in it and it also was on the internet was the first exhibition you could J you could virtually visit on me January 1995 just think I bet very few people in this room ever got on the internet before 1994 middle of at the earliest any hands up for before that very few some of you are obviously far too young so we have little book company huddled open the exhibition and we built the exhibition and Julia and spent most of two or three weeks on the floor wondering all this up and this is an important part of the story because we got a grant from the Arts Council to build this exhibition and you know you don't read the small print but I read the small I discovered to my horror that I'd signed that this exhibition would twelve traveled six other venues there wasn't a chance of moving all this masses of wiring and computers the whole of this space under there to stuff with them and so instead of iteration we said to the Arts Council well it will have all virtue and so what on earth do you mean as we'll send the system all over the world and people can interact with it from anywhere and up to that point we have actually thought of that so suddenly we had to build in this idea that anyone could interact on the internet from anyway the answer said yes that's all right so those you hang a notice next to each of the computers that does is saying this was funded by the Arts Council so we had to print this poster and post it to all our friends and say please photograph this next to your computer and send it back so we got away with that that was the book and that was the exhibition and it contained all the interactive objects we've been talking about including the universal constructor from five years earlier which we managed to restore and got working again but the important bit is right in the middle oh sorry must mention this in passing there's another Nicola is of one of these um data structures based on close packing spheres but in this case certainly to post fact the spheres there are nests of LEDs which change color and can be reconfigured to show that all possible crystal structures for example can be fitted into this database there's another thing that science didn't know even full who was the first proposed the idea that nature had a geometry which was ISO spatial and didn't realize that you could fit all the crystal lattices in it took Nicola to realize that by doing a phase change I might have time to explain this but this is just pure genius and but right in the middle of the exhibition was this there were three computers one evolving a space one taking information from the internet and the other examining the situation in the exhibition space and just here on the edge of the photograph are a whole lot of switches which were genetic switches so visitors could go and switch the genes on and off and see how that affected the way the structure was changing and in the meantime the environment which was always important to us was also being sensed for noise and humidity so the exhibition space was there any noisy or smoke it could be in those days a lot how long ago this was you can still smoke in the exhibition space 20 years ago it would stunt the growth of the system and the way it ran is up here are coming in messages switches in the exhibition space messages off the internet because that was the point that people could make these switches virtually on the internet here we got clothes fittings it's dead of switz I'm trying to say instead of cubes it's now spheres and up here is the evolving form the complexity of this software was considerable it relied on a whole series of cycles of evolution it was the most complex thing there are a long way that we'd achieved by that point this was done by a group of students which contingent MANET a stogie Christiana Giotto Patrick Jensen and one of my PhD students an all-star who was doing genetic algorithms with me at the time so we needed his expertise in there too and that this is the output stages it is not an architectural proposition I have to explain this because Peter Eisenman was our examiner this year he said well why don't you make a buildable proposition out of this and I got very excited about the fact you would actually section this cuz I don't know what can have done the software he had but he couldn't section it but this was a real 3d model this is a canned animation of what happened over the three weeks of the exhibition and all the interactions from literal visitors to the exhibition space and virtual visitors online managed to evolve this kind of mess and I'm looking at this piece of ice and thinking yes well in the early days there is a certain protip 'no switch you have to do to embrace in order to make your point so that was what was that it was it was evolving for increasing complexity that was all and MANET went on to do a network of computers evolving missus elaborate of communication between networks we we had a great div least stopping mallet and networking every computer on the planet he wanted to do I can basically then hack it if I me said look this is all right well but this is the AA's name and if the a is known for hacking into but he thought it'd be amazing to see what what several million computers could do overnight and he we never knew although that came up today towards the end of the questions and became the notion of what would happen it was a project as before the ice about what would happen we've got all these hyper intelligent cells and one of the critics asked a question well but we haven't got a model it's never looking at nature because Nature doesn't have that level intelligence in multi cell systems a good question we got a series of lectures following from that which really spelled out the the range of interests at the time those those tied to the evolution of form evolution of chaos and causality generation of ideas consciousness artificial life plant architectures subconscious that's all about a plant architecture plant geometry I'll come to that in a minute and in Year seven I got the end of this year seven and we went to growing gun and built an interactive model with the citizens there that allowed them to interact with the development of growing a little in in the middle of the sixteenth century it's a little isolated town and far northwest of Holland by 1600 to become severely fortified and we never did discover what it was that they were defending against whom but it was a classic sort of village with the Catholic Church the Protestant church in a marketplace in the middle and we built this interactive model which allowed the citizens there to get at mrs. Christiana Chicanos version of this a top-down economic model and a bottom-up economic that decentralized interaction and over here on the wall and looking at the same kind of thing that's with top down and bottom up it's a fundamental principle that has to be explored with all these systems and this is another level of complexity in this though because it was recursively self-similar at different levels from the individual level through the urban to the global I will be very interested to know what would happen if you took that diagram on the wall of the ice and did the similar thing with it and looked at it recursively at different scales or at Kauffman we built a chemical computer based on Gordon Parks work or it was the last student to have a to troll with Gordon before he sadly died but this chemical computer learns on the basis of feedback and she used that to build a simple cellular automata which built one this is the C bogus site we'll be using as a demonstration which led to propositions about how that city might evolve this is cristianos model of the whole system there is the evolving city model out here is the crucial thing of the user interacting both bottom-up and top-down over this side it's a gene pool of genetic descriptions of this city up here is a neural network which is looking for identify successful behaviors and down here genetic algorithms which are rewarding successful outcomes of the bottom-up this model actually has all the critical elements in it and then Chris licked them all up and here he has this is this actually his final presentation for his diploma and there are the computers at different cities around the world or communicating with each other to share information to define evolving models of a forum for a particular for a particular city fourth of july 1996 the same year johnny Botsford um his project was the way we work does he probably gathered was the way we had each that the unit had one shared project like a model of groningen or the universal constructor or whatever it was each year but then each student also did an individual project with it so in this case they're all commute collaborating to build the model of groningen but Gianni's role was to look at solar logic in this and he it's again another interesting night example of top-down and bottom-up but in this case i'm showing it as an idea of an active tool let me just define that a passive tool is one which is post hoc so you design a building and then you analyze its thermal performance that's post hoc analysis after you've done it very useful but you've got then to go right back and change the model active tools are there right at the time when your thumb a tip of the creative diamond you're having the first idea is the first sketches the first problems are being put down those are active tools which tell you right there and then so in this case these are points in spaces of the balls represent about which analysis has been done on each one before any designs taken place so that the designer knows the characteristics of all these points in space from that they can generate envelopes or possible solutions and from that you can generate an actual solution I'm gonna show Johnny applying as all all students have been talking about went on to apply the thinking they've been doing to real projects now just touch on that in a minute because I'm now moving on to the the developments post 1996 Corsa was continuing work at the a of course that was the establishment of all the groups that DRL and anti which would mean enjoying today there was take up a lot of schools of architecture which I'll touch on and there was further work by the other members of the unit either in PhDs or theory or practice and I'll touch on each of these this is Johnny in practice he the site was this old factory completely surrounded by houses and Notting Hill Gate got planning permission on the addition that there were no external windows so they're you concerned a construction there are no windows because it's being overlooked by all these adjacent houses and the problem was because of the low Sun angles in London and the cost of higher buildings all round it the question is how ever to get sunlight in daylight in so cost gianni built an active tool with the help of erips engineers these these spheres in space of a points which get particular levels of solar penetration all year do you see they don't come down all that far further down you've got another cluster which are at particular times and you can write the way down to places where you can only get daylight in two corners where you can't even get much light in at all so from that GRE they're mapped different times of the year the possible distribution of sunlight both sunlight penetration direct and general daylight different parts of the building and from that he generated quite straightforward orthogonal plans and sections it's three stories and it's goes out of three and a half there so it's all top lit and through these slots diagonally and solar direct solar penetration occurs I took this photograph just before it's handed over to the clients and just finishing I think it was the 26th of November and even with a very shallow Sun Lionel's in London there you can see as you come in at the front you can still see direct solar panel projection of the wall so you don't feel as though you're coming into this dungeon the lap pool in the basement you can see brilliantly lit with reflected so lighting of course you chose surfaces to to exploit this to the extreme electron they will see it when it was taken over by the clients there's always these shafts of sunlight coming into any of the spaces this is right down on the lowest level and that's a shaft of sunlight not all the day of course but a particular time so that every space was in Leiden that's an example of an active tool she took over from from his student project and I went to Ulster sorry the Hong Kong miss case I went to Hong Kong and some of the students came with me guy came with me Krish Carly came with me Laura slough came with me Timm Timm Jack no later came with me with the Max and we had this vast horizontal space this is worth mentioning because now Saha Hadid has made it vertical so far end of the campus is now a vertical building which links all these spaces vertically whereas before we'd link so this is the end of year exhibition where every single student project was expected to work to move crawling little kiosk but but when I arrived to see they all made these polystyrene models and I said no no more Pakistani models next any very exhibition the second year and I see a button on this I don't want a piece of paper press I expect a press of button and I expect it to work and I had to say although I really enjoyed today the big disappointment was didn't see anything running so we built them a global virtual design studio which allowed us to do this which is a krip with the AAA this is how these links kept on going there's moisten this project here is from the AAA and is being projected from this lecture room there are you'll see there in a minute with some students these students here the shadows are in Hong Kong that's moisten and I self here a front talking about his project and this is very weird because we had massive computer power and the AAA students have never seen these flood that sees a student a student standing in the air explaining his project to the students in Hong Kong and there's his fly through which he's never seems they didn't have the processing power ta to do that kind of thing so this is the most extraordinary kind of event using the technology at both ends of the planet and also there we rebuilt the universal constructor smaller better made but it was used for different purposes in this case it was being used to teach programming and again Christiana is using it to teach the students how to program what he's teaching them to program is how to operate dry little robots and it's being done by each of the chunks of instruction code being represented by one of these cubes so all they have to do is stack them up that's Tom Fisher and that's going to be Christiano no doubt yes trying to talk Cantonese I turn the sound off and what they're doing and they programmed a robot to pick up golf balls here that find the golf ball pick it up and put it in box but it's simple programming tasks but we were using the universal mapping capability of the universal constructor there's an educational so this is really just standard says how these tools of technique spread all over the place and being taken up in different ways and yeah su was all being the attention of the universal constructor would move and I just show that little kid four feet solo with the spline yeah so finally built the motorized version of this and this is how the universal constructor should have looked in 1996 it took us a while to get to the technology where we could do this but it's obvious what how its operating its essence is responsive to gesture and movement and hand proximity inside and because this is Hong Kong and I got a piece of advice if ever you take a really dull job like being head of school of design or architectural someone Dean um make sure you demand a vast dowry because you can get one Garrigus so I had some 14 million to play with so we bought every piece of rapid prototyping you could get in 1996 so we had printers we had laser cutters oh the laser comes at fashioned not just architecture students II I've had industrial design students and then graphic students and I had wonderful well they this laser cutter the first people get to it fashion students what did they do they cut that is leather cut with holes made by the laser cutter the health and safety people who'd installed the air-conditioning hadn't imagined the smell of burning laser cut leather ah it was worth it was worth the embarrassing discussion with president of the university and and and the fashions great so this is a this is a printer but it prints onto fabric so you can make clothes with this this is a weaving a knitting machine that makes clothes knitting and you can make a complete bodysuit you can have gloves hands toes the lot follows there's one the whole left and it's not made in bits that you stitch together it's woven as a knitted as one piece so having to control and understand the software for all these different kinds of machinery we're incredibly good for that the effect it had on the environmental design in the dust rezaian students so having all these disciplines together was was great now of course we had routers and 3d milling machines when the fashion students got to that first - they started machine themselves and perspex shoes and at that point they were only used by a certain kind of girl in Hong Kong and then we moved into industrial design and started to do things like evolve telephone handsets and indeed some of these pieces of software have been used in industry you may well have a telephone in your pocket that was designed by that piece of software um some of the students went on and did other things for other architects like guy this may not show it and sometimes it's very surprising but actually this these tasks have actually been used built designed using active tools to do with the lighting sunlight and so on this is north of Genesis a mallet very cleverly named his practice after the unit which I thought was brilliant and so now got projects all over the world as you can see and awards all over the world and he's still using basic unit ideas to generate form this is to do his solo angles he said the geometry wrong by the way but rather amused me and so I hope is listening to the tech man it he's coming on Monday to the party so he'll hear about this but all these things zero energy zero water balance zero waste discharge all these things are the kind of preoccupations of the unit is using those unit tools to develop these ideas and sometimes in rather surprising materials and this is Christiana of working for Zaha so her galaxy hot the whole of these exercises that we've been concerned this I have to say Christmas the presentation but this is it Chris this is they have a right at the back they have a quadcopter which flies around depositing ice to make me shake and the exactly like this is perfect you could have saved a lot of work but Chris not being able to do that how to look at the complexity of shape how to minimize the number of particularly complex difficult geometrical forms which we're going to be a very expensive to develop in order to maximize the simple plane curvature areas in order to make the overall project fit within budget and so on so it's still applying these kind of fairly sophisticated computer analysis techniques and generative techniques but for a real project variable geometry presses obviously all computer control producing the panels panels to construct the the complexity of the curse of the architecture and some of the students went back to doing theoretical work there's Nicola whose font replication I showed earlier she went back and did a PhD with me and one of the things which he developed amongst others was a proper analysis of this fault replication and again that's another idea which has passed very firmly into the scientific domain this is Patrick Janssen he went on I'll just show a little clip of his in a moment McCain will went more into publishing am i selling I'm teaching and saying with Sam who's about to publish the collected works of Cedric price having previously produced three other volumes about Cedric of course that makes another connection back to the interests of Cedric and his role in the unit and then of course that brings us to today and the exciting projects we've seen around this this room um now I just not to say a little bit about the theory out of this there are many parts of Theory out of this and I'm just gonna be able to deal with one which is the evolutionary digital design process I'll skip through this quite quickly it runs like this like every architectural project it has to have a generating concept this is not bottom-up at this point this is top-down at the start you know just top-down into a seed and I'll explain a seed in a second the seed then goes into a gene pool and at that point this becomes a more bottom-up process as you go so reproduction development prototyping analysis selection and eventually materializing or being recycled and I'll skip through this quickly my analogy has also always been the idea that deep the Acorn always makes an oak tree they're always recognizably oak leaves and oak trees but every leaf is different every tree is different depending on its environment and that analogy has just been taken up by Google with a project they're proposing where they're hoping to make 200 billion a year or whatever out of the building industry doing this plant architectures and body plans have we've heard about today fascinating there are only 24 or 26 strategies which nature has to make all the different trees and world and that economy of means is something that's very powerful to us and the notion that we need probably body plans and every plant every tree can be analyzed as one of those types and that produces an incredible richness there seems to be an obsession in architecture that if you have types or generic ideas you'll somehow get endless repetition this is nothing could be further from the truth and the same with body plants this is the same body plan being applied to different forms of animal fish and birds so the strategies of nature are fascinating and relevant so then we come to seeding and at this point I'm going to briefly go back to my own project as a student at the AAA I was struggling with this little structural system which consists of just two structural units because in those days you see we thought we had to have mass production so that dyes are very expensive to make this so you have to make minimize the number of units so they're only going to be two and then we're going to be no cut units you're gonna make straight-line edges by just reconfirmed by the way you turned around the different units in the system the problem was one of drawing it and when I was a student I had to hire a flatbed plotter by the hour and as I saying to someone it was no preview you would just have to stack up punch cards right so you had no idea how big this drawing that's gonna be so you you you fork out all your dinner money for a month to go and go and get this drawing done and that in the lab and you turn up it's the a didn't have a plot didn't have a computer it had one teletype tunnel to remote time-sharing because that's all the air you had then and when you got there you found that either it had plotted the whole thing off the piece of paper completely and this is the puddle of ink at the side or it had either dug a hole in the middle of the paper or intro wit backwards and forwards across it slashing it to pieces and there was your precious time your precious piece of Whatman paper gone and so finally the triumph of managing to get a toy onto the piece of paper and I was I just touched to the core today when I was talking about the robot that kept alive tried to put the pin back in the cut missing and I thought yes I know that the triumphs the triumphs what it actually goes into the paper cup the triumphs and the drawings actually on a piece of paper so so when I got to Cambridge I started is thinner putting all the individual cord because it was none there's no alternate you have to write every piece of software from scratch had to write everything within every vector every perspective program you had to write yourself in Fortran it was agony got a Cambridge and IRA but all that I'm gonna I'm gonna put in a minimal configuration of these things and use the power of the computer to do a bit more than just draw it and so that is the minimum possible configuration of my structural units there's just two still but they're in 32 different orientations in this in this close thing and it's closed so it will expand and not get holes in itself that in ninety 69 is the only graphics display device in the whole of Cambridge University I say that because if you look a copy of AD at the time you'd think every architect had a computer on their desk no no none of them had not a one and of course they also said there never would be oh well both those things change it's a real computer look it's got flashing lights like in the movies but the screen hadn't even got storage capacity it draws like this by flashing the beam around and and it it's slowly fades so if you try to do a drawing that's too complicated the beginning of its faded before you get to the end so that keeps using them very simple that means I can show you these very simple drawings there is the seed just simple instruction grown and magic it grew on the screen sheer at the top the structure been that been the structure and it basically was this instead of putting all the data into the data structure you just put a seeding then you had a cultivation program which enables you to build that data structure round and round doing all the work for you this is my dream of being a computer's I'm fundamentally incredibly lazy so the whole idea that this dream machine could do all my work in order to do that the problems of trying to learn to do to Atlas Titan machine Co programming at Cambridge and various other things were problematic I could only then hold one tiny dot of data and I had to do this in the middle of the nights the only time I was allowed to have the computer and the graphic display device and the important part of this story is the only other person in Cambridge who is using that at the time was John Horton Conway who invented the live game so he and I were sharing this computer every night for years not three years so that's finally got various structures you only got high resolution you see when you put them on the plotter we could put this on the screen okay so that was seeding and then you go into a gene pool then you get to reproduction your I don't need to explain all this let's cross over this mutation then rubberish generative techniques are a lot of these parametric geometry being the one which I think most people are most familiar with but there are others like combinatorial systems and these two bottom ones seller automata which we've again hit today parallel discrete computation poly automata and procedural shaped languages the one I just showed was written in something called the procedural shaped language which is my favorite but you can use any of these and that's of course what generates a 3d cellular automata like this is a set of rules so that's the generative engine then you have the genetic algorithm tributed to john holland but developed mostly for optimization in engineering we suggest that we could use this for experiments on design problems and everyone said the engineer said don't be ridiculous house really stupid everyone knows architectural problems have ill-defined and they have conflicting criteria and obviously genetic algorithms can't do with that I said hang on a minute human life um would you say that was clearly would you say that all the criteria aren't conflicting I said well genetic algorithms seem to have managed to do quite well they've even managed to evolve you and me so um I'm gonna have a go and so we did and now applying genetic algorithms the design problems is common right through engineering right through computer science so when it was done here in the AAA no one had done it before and in fact they said it was impossible we first did it Peter Graham on boats optimizing boat design and then we start to combine these techniques you start to combine the cellar automata with the genetic algorithm I showed that with Ichiro earlier where he's got both the cellar automata generating this and a genetic algorithm controlling it from one run to the next all that has to take place and in environment that is crucial tomorrow thinking and so this sequence which I showed earlier was evolved in that environment in the exhibition space that just show this from Patrick Johnson Patrick another student who was responsible for that initially he'd also did his PhD with me and here as a series of examples from his PhD where there is all those have been generated entirely automatically from one generative concept and there are proper solid models but they're able to be and first of all they're running being generated totally automatically you can see here it's gonna pop in a staircase in a second in response to some building brief it's deforming the geometry and each time this program runs it has a rigorous analysis of everything it can manage to analyze and on the basis of that will modify the genetic algorithms so that next time it will produce a more optimal solution you can see it's carving holes through the center of the building in the moment it's going to add some fenestrations things get past this oh sorry should be a maybe as in a moment virtual prototyping I thought I'd better show one sort of real-life project and this is one island east in Hong Kong which is of great interest because it was the first building to be totally procured from a virtual model Christmas working for Frank Gehry at the time before he went to Saha and Frank software digital project was being used for this and whilst Frank was not in any way involved with the design of this building the software wind digital Gehry technologies was helping to structure this program and then and then analysis these are the hydrostatic calculations for example from the auto yacht program being done to do the boats and that's in there you had to perform some kind of analysis at some point and this is post hoc analysis as opposed to the active tools I was describing earlier so there you see a typical run there's a genotype there's some plans coming out in the building form or some analysis taking place of its thermal performance from that as an identification of fitness from that and evaluation and so it goes on selection this is crucial um natural selection convergent evolution is what engineers on earth they wanted to converge an optimal solution they wanted to use natural selection which means calculations to do it we said no there are three other squares here which are as important possibly more important you can diverge you can use these techniques to explore different ideas or there's artificial selection which is you interfering as an individual and that came from Darwin himself in the original block he's he had the problem of selling the idea of natural selection so he said well you all know about variation under what he called domestic patient that is how to breed racehorses or pigeons or dogs and he then went on to say well we got variation under nature as well and then he said and nature given the massive time that can do natural selection so that's how on this diagram that is the one which nature uses but this artificial one the dog breeder is also important you have skill intuition judgment which you can use - and that goes into this box here and after that some are discarded some go around the gene pool so here you see from Patrick Johnson's program here's a perrito optimization curve all those along the bottom edge are pretty good all these ones up here are pretty hopeless and then it will rerun and reevaluate the the performance of those prototypes so that's a description later on we went on after patty - Dee's PhD patterning I managed to get a big research grant and so we started to look at all the other um quantitatively a valuable criteria leaving still the idea of your skill and judgment in that artificial selection box a finally materialization by our perhaps Building Information model and that's one island he's not very elegant building to have such an important role historically as the first building to be totally procured from a digital model and finally I come to the end it stops this notion that that elicits all these ideas have been going into computer theory into engineering and all over the place it's recently come to my great delight - motor racing and it's finally realized that pit stops are a design problem and I'm pleased to tell you that they are now using genetic algorithms with both natural and artificial selection thank you a to determine pit stop timing it's a perfect design program for a genetic algorithm software AutoCAD are now producing a genetic algorithm user software it's useless so far but it will be in the end I'm sure a tool which has access to and will want NASA have an artificial evolution design program again based on work here which is in this case dissolving devalue rather improbable antenna but the final point is this there's been a spate recently of books called things like computational design thinking digital turn in architecture about the sort of teasing notion that maybe there's a theory here now I am cautious about this there could be a theory of digital design I think we could all agree that that's a very specific thing I've just outlined a very specific theory of evolutionary digital design but you'd have to be a subset of that that's just my own particular interpretation from what we've been doing I don't think the digital in general can be a new theory of architecture in a general sense however it could catalyze its back to the point that theorem a right at the opening which is that thinking about the computer can cause you to rethink what you do otherwise so I think what's going on the digital design could cause us to do a rethink of what's going on as a general theory of architecture computer as element Muse it could inspire us to build a whole mental models differently of architecture nevermind use in the computer as to how we conceive it change the way we conceptualize architectural ideas affect our own very own internal psychology of model building in our heads the idea of conceiving of ideas not just for development of the one project but the idea that you would conceive of architectural ideas with the idea that they might evolve most architects will say I mean there was a nice little quote the other day from from Norman Foster Dulles where he he says well every time I start a new project I like to think I'm starting with a blank piece of paper or a blank screen he said at the end with hindsight I can see an evolution development between projects what if instead of seeing that with hindsight you started thinking of that up and conceiving of ideas which could evolve from the very high flock and why not conceive of ideas which could be enriched by the computer process instead of diminished by it the old joke used to be not tired but cod Co D computer obstructed design and the chairman of autodesk actually went on that project proudly the other saying you can tell which of the projects and the night is who design with autocad cuz of the primitive forms i think are you proud of that you know so there's the process summarized you can see that than i do you encode it you breathe it you develop it your generation analyze it selecting evolve materializing said in full you could start to say instead of that being justin you could start to apply that to any design process you can start just crossing out words that are to do with computing so i think we're moving to a point where we're going to be able to change the way we accommodate new and different ways of thinking and i think we've seen some very nice examples of that today so congratulations to students that we saw on but right at the tip of the creative Darland where there's a gem of a new architectural idea and i think that will change the balance of the roles between architect client user computer and environment will all change and I think that could lead to a new theory and you kind of process a new kind of creativity where we think about that creative process differently and a radically new architecture John could answer a couple of questions if anybody's interested there is a floating mic the way it works you raise your hand floating Michael come to you you ask a question you get the answer is there any questions everything was super clear we could go home and have dinner yeah is there any questions going once twice sold to nobody okay thanks everybody then let's thank Joe if I can just make a quick announcement tomorrow we pick up day two ten o'clock same room same scenario different keynote speaker thank you to our dearest