Stairs beyond our intuition

Our computational tools, from CADCAM to building information modeling, offer new ways of tackling this complexity. One of these is the method encouraged by "parametric" thinking, because it tantalizingly generates a lot of complexity, or at least a lot of variation, from a small number of initial settings. But even if a parametric city still looks like a city, that doesn't mean it behaves like a city. The danger of such an approach is precisely the ease with which such complexity is generated from the bottom up, because the more we stretch our designs beyond our intuition, the harder it is to judge them.

But there is another side to using computation that could be very useful. Artificial intelligence and machine learning technologies, such as the clever use of simpler algorithms, can be helpful when we extend our own intuition. Or they can give the computer a kind of intuition of its own. In either case, these technologies extend our abilities to deal with true complexity. This is one of the main aspects of the thinking that characterizes much of our current research in architectural design and adaptive computing, as well as the related Bartlett's currents, as I will exemplify below with a few projects that represent the two ends of the design ladder: the very small things and the very big things.

Our computational tools from CADCAM to Building Information Modeling provide new ways to deal with this complexity. The sort of method encouraged by "Parametric" thinking is one way, as it seductively generates a great deal of complexity, or at least a great deal of variation, from a few initial settings. But just because a parametric city looks like a city, it doesn't means it behaves like one. In fact it is this ease of generating such complexity from the bottom up that is a potential danger of such an approach, as the further we stretch our designs beyond our intuition, the more difficult it becomes to judge them.

There is another side to the use of computation that is potentially useful, however. The technologies of artificial intelligence and machine learning, as well as the informed use of simpler algorithms, can be aids to extend our own intuition, or give the computer a kind of intuition of its own. In either case, they extend our abilities to deal with true complexity. This is one of the core strands of thinking running through much of our current research in Adaptive Architecture and Computation and related streams at the Bartlett, as a few projects at the very smallest and the very largest ends of the design scale may help to illustrate.