Help Center
Privacy
September 2020
Digitalisation and Health
Interview with Mario Carpo
Editors Jade Bailey, Adriana Boeck, Emma Sanson & Patricia Tibu
Editors Jade Bailey, Adriana Boeck, Emma Sanson & Patricia Tibu
Mario Carpo is a professor of architectural history and theory with a focus on the early modern period as well as on contemporary digital design theory. He has published several books - his recent one being The Second Digital Turn: Design Beyond Intelligence (MIT Press, 2017). He has taught at several renowned universities including Bartlett School of Architecture, at UCL London, and since 2020 he has been the head of the Theory of Architecture departement at the University of Applied Arts Vienna.
You've been regarding the current situation as a wake up call for the necessity of transitioning towards a truly digital world. How do you think this transition will materialize in domains such as healthcare or even construction, what will the area of expertise be for the future professionals of such domains?
It is evident that the corona crisis has proven what has been said for 20-30 years, that the standardised, mass produced modes of production of modernity are wasteful. The mass customised approach of the digital world is more respectful of human resources, of material resources, and of energy. The entire mentality of modernity predicated on the assumption that resources are unlimited. That transportation cost nothing. And that energy is available at zero cost and forever. These premises are not true anymore. We now know that resources are limited.
The typical modernist approach to the problem of making a teapot is to standardise the teapot, conceive a universal teapot, and then build a colossal, humongous factory in any place where clay is cheap and labor is cheap. To build this universal factory for the global teapot, and the factory makes gazillion teapots a day and then from there teapots get distributed around the world.
But during the lockdown factories around the world shut down. And the network of global transportation, on which the entirety of our life depends, froze. In the space of one weekend in March, all the airports shut down. Which means that in Europe we thought "Well we don't need to produce facemasks because we can import them from China" - from the moment when the flights from China were stopped we ran out of face masks because import was no longer possible. So this idea of the global transportation chain during the lockdown airports shut down globally, factories shut down globally, but small robotic factories or the local fablabs kept functioning.
The typical modernist approach to the problem of making a teapot is to standardise the teapot, conceive a universal teapot, and then build a colossal, humongous factory in any place where clay is cheap and labor is cheap. To build this universal factory for the global teapot, and the factory makes gazillion teapots a day and then from there teapots get distributed around the world.
But during the lockdown factories around the world shut down. And the network of global transportation, on which the entirety of our life depends, froze. In the space of one weekend in March, all the airports shut down. Which means that in Europe we thought "Well we don't need to produce facemasks because we can import them from China" - from the moment when the flights from China were stopped we ran out of face masks because import was no longer possible. So this idea of the global transportation chain during the lockdown airports shut down globally, factories shut down globally, but small robotic factories or the local fablabs kept functioning.
Because they were local, because robots did not infect one another. If you need to put only 3 people in a robotic shop it's possible because they have enough distance. And so when we realised that we could no longer import stuff from far away, many of or our friends and colleagues, converted our fablabs into production facilities for personal protection equipment - all in the space of a weekend.
'This is exactly what robotics can do; it is versatile, it is flexible, a robot can make a spoon today and a protective mask tomorrow, and it is local.'
This is the alternative - instead of a global factory which we have learned we can no longer depend upon. The local FabLab where everything is made where you need it, when you need it, as you need it, on-time, on-spec, on-demand, and locally.
'This is exactly what robotics can do; it is versatile, it is flexible, a robot can make a spoon today and a protective mask tomorrow, and it is local.'
This is the alternative - instead of a global factory which we have learned we can no longer depend upon. The local FabLab where everything is made where you need it, when you need it, as you need it, on-time, on-spec, on-demand, and locally.
Once architects have mastered automation, or in fact maybe the digital tools that are used to create architecture have automated themselves, what do you think will come next? What is the next necessary step that architects have to face in terms of social and climate responsibility?
Automation, if you think in the terms of the history of robotics, was very trendy in the early '60s when robots were introduced in America to automate the mechanical assembly lines. Industrial robots started to be developed to replace the mechanical industrial worker. So, these were robots trained to imitate a human worker, to repeat the same gesture from sunrise to sunset because this is what the workers, which they were replacing, were meant to do. They are cheaper than actual human workers, they never get tired, they never sleep, they never strike; so from the point of view of a capitalist - that is a good thing to do. But for what we do in architecture, these kinds of robots don't have any use, because even the most advanced building site that we may conceive is still a very artisanal environment.
So, for our job, robotics only becomes useful if we can have intelligent robots who can perform intelligent movements by adapting themselves to different working conditions. This is called 'adaptive robotics', 'versatile robotics' or 'flexible robotics' which now exists. This is not a robot meant to replace an industrial worker, this is a robot meant to replace an artisan, an artisan dealing with the unpredictability of natural materials. This is the next frontier of robotics, which is probably important for some industries, but it is crucial for us. Because in architecture, in buildings, we mostly deal with unpredictable conditions. And we have to deal with a lot of natural material, and not with industrial material. Steel is a case in point. Steel is standard because it is an industrial material. Steel is always the same, which is why engineers can calculate steel so easily.
'This is the future of automation. Automation which allows us to avoid the wastefulness of the modern way of building; going back to the economics, to the daily rules of a society which was used to coping with limited resources, limited workforce, and almost no energy.'
So, for our job, robotics only becomes useful if we can have intelligent robots who can perform intelligent movements by adapting themselves to different working conditions. This is called 'adaptive robotics', 'versatile robotics' or 'flexible robotics' which now exists. This is not a robot meant to replace an industrial worker, this is a robot meant to replace an artisan, an artisan dealing with the unpredictability of natural materials. This is the next frontier of robotics, which is probably important for some industries, but it is crucial for us. Because in architecture, in buildings, we mostly deal with unpredictable conditions. And we have to deal with a lot of natural material, and not with industrial material. Steel is a case in point. Steel is standard because it is an industrial material. Steel is always the same, which is why engineers can calculate steel so easily.
'This is the future of automation. Automation which allows us to avoid the wastefulness of the modern way of building; going back to the economics, to the daily rules of a society which was used to coping with limited resources, limited workforce, and almost no energy.'
Now, this is the modern mentality: Standardise materials so you can calculate them using modern mathematics, and when you put all the pieces together you know in advance what each piece will do structurally because it has been tested before it leaves the factory. This is the modern, industrial frame of mind. Today we can envisage a new way of using machines. Imagine a robot that can scan a tree when it is still in the field, and can model the different branches and can understand from the building you're building which branch you'll need. And then it will cut that branch, to measure, knowing in advance that it's the branch you will need, so there is minimal waste from what it has taken from the tree, from the forest to the building site.
So it is possible to use advanced robotics and computation to work with natural timber, in the way artisans always had before the Industrial Revolution. So, paradoxically, the next step of computational automation is to go back to artisanship. We have to go back to a way of all being artisans because it is more efficient. The carbon footprint of traditional artisanship is much smaller than the one created by all the waste generated by the modern, industrial way of working. From the point of view of the use of materials, the use of energy and the carbon footprint, going back to artisanship is the future, because traditional artisanship was parsimonious by virtue of its own mentality.
This is the future of automation: automation which allows us to avoid the wastefulness of the modern way of building; going back to the economics, to the daily rules of a society that was used to coping with limited resources, limited workforce, and almost no energy. All the waste, the carbon waste, the energy waste, the material waste that we have today is based on this colossal misunderstanding: bringing the standardised logic of modern industrialisation to the building site. With today's machines we are not obliged to standardise.
In your opinion, robots can replace physical and sensory work. Now with recent developments of artificial intelligence, do you think AI could replace the 'thinking' work of designers in the future?
You know the idea that artificial intelligence can replace design - can produce design automatically - was a very strong idea at the beginning of the age of cybernetics. It was strong in the '60s, it imploded in the '70s, it disappeared in the '80s, and then computers came back in the '90s as personal computing. Nobody used computers to replace the designer, the computer was a tool that the designer was using and exploiting to create new forms. It was a tool to support the designer - not to replace the designer.
But now it is coming back and today it works! Artificial intelligence today can solve problems, basically problems of optimization. Say you want to find the best span for a bridge to save steel or the best size of a city if you want to minimize the wastefulness of air conditioning. These optimization problems, which are mathematical problems, is what artificial intelligence can do today.
However, even the most banal architectural problem has such a huge number of variables that there is no way to optimise them all at the same time. You cannot ask the machine to optimise everything because that is a paradox. You can only optimise some parameters to the detriment of some others. So the idea that the computer will replace our design solutions is first of all technically meaningless, because you cannot optimise all parameters at the same time, and it is not within reach.
But now it is coming back and today it works! Artificial intelligence today can solve problems, basically problems of optimization. Say you want to find the best span for a bridge to save steel or the best size of a city if you want to minimize the wastefulness of air conditioning. These optimization problems, which are mathematical problems, is what artificial intelligence can do today.
However, even the most banal architectural problem has such a huge number of variables that there is no way to optimise them all at the same time. You cannot ask the machine to optimise everything because that is a paradox. You can only optimise some parameters to the detriment of some others. So the idea that the computer will replace our design solutions is first of all technically meaningless, because you cannot optimise all parameters at the same time, and it is not within reach.
'It is possible to use advanced robotics and computation to work with natural timber, in the way artisans did before the industrial revolution. So, paradoxically the next step of computational automation is to go back to artisanship.'
Even if it were at some point, if a computer could emulate the maximisation of the optimisation of certain numbers of parameters - this is not going to happen next year, not even in 10 years. Because if it did, it would always be a very expensive solution, and honestly I don't see who would want to have a very expensive machine to replace humans to do what humans can do so well, which is to create, after all, forms of art. To do that we need machines, we need computers, as tools that we use to optimise problems that we formulate. But the idea that we devolve the formulation of the problem to a machine is an abomination. First of all, it's technically unlikely. Secondly, it's against the spirit of what we do. We solve problems in architecture in an intuitive, problematic, ideologic, biased, personal way - because that's what a work of art is about.
Even if it were at some point, if a computer could emulate the maximisation of the optimisation of certain numbers of parameters - this is not going to happen next year, not even in 10 years. Because if it did, it would always be a very expensive solution, and honestly I don't see who would want to have a very expensive machine to replace humans to do what humans can do so well, which is to create, after all, forms of art. To do that we need machines, we need computers, as tools that we use to optimise problems that we formulate. But the idea that we devolve the formulation of the problem to a machine is an abomination. First of all, it's technically unlikely. Secondly, it's against the spirit of what we do. We solve problems in architecture in an intuitive, problematic, ideologic, biased, personal way - because that's what a work of art is about.
'Artificial intelligence today can solve problems. [...] But the idea that the computer will replace our design solutions is first of all technically meaningless, because you cannot optimise all parameters at the same time, and it is not within reach. [...] And honestly I don't see who would want to have a very expensive machine to replace humans to do what humans can do so well, which is to create, after all, forms of art.'
Once architects have mastered automation, or in fact maybe the digital tools that are used to create architecture have automated themselves, what do you think will come next? What is the next necessary step that architects have to face in terms of social and climate responsibility?
When we design, we always have the future in mind, because a design is something that does not exist yet; so you should design it because you have an expectation, you can anticipate the forecast something that does not yet exist, but you want to make it happen. If you are a designer, you have a certain confidence that the future can be better than the past - it's as simple as that, it's optimism - and if you think about it, many people would disagree. I remember at the beginning at the first Coronavirus lockdown, the French writer Michel Houellebecq, like many other people, left Paris and went to the countryside, waiting. He published an article in a French daily newspaper and his last line was "and you think that the Coronavirus is going to change everything? No.
'If you are a designer, you have a certain confidence that the future can be better than the past - it's as simple as that, it's optimism.'
'If you are a designer, you have a certain confidence that the future can be better than the past - it's as simple as that, it's optimism.'
It is going to change nothing; when it is over, everything will be the same as before, only a little bit worse."
This is the real mentality of someone who does not want to design. We design because we think that things can be the same as before, but a little bit better; which is why we try to design. Without optimism we wouldn't be designers. We have this messianic optimism. We always think we can make a drawing and that will help to make things better; to make better things for more people, using fewer resources, diminishing our carbon footprint - that's what we try to do, better, cheaper things, especially for housing, infrastructure and cities; this is after all our vocation. If we do not have some optimism, whereby we think that by working six months, day and night on a project, we can bring our little contribution to making better cheaper housing for more people, then we should not be in this trade.