14 May 2015
London / New York
As the final event in the Embedded Intelligence Lecture Series, Skylar Tibbits presented on the topic of Material Intelligence, describing the role of materials, their inherent properties and their possible potentials in architectural design through a series of experimental and applied research projects undertaken in the MIT’s Self-Assembly Lab which he directs. In his talk, Skylar interrogated the design opportunities provided by techniques that act upon our standard material set and questioned the role of new materials in the development of novel design strategies.
After Skylar’s talk, Evan Greenberg and Gil Akos, who direct Embedded Intelligence, the AA’s Visiting School in New York along with Ronnie Parsons, led a Q&A session which is transcribed below.
Gil Akos (GA): I’d like to start by circling back to the origins of your presentation. You were talking about how code is a kind of new or “newish” language for design. That’s something that now has a heritage of 50 years or so. We know that material is a language for design as well, and that heritage goes back centuries. Is programmable matter a new design language or is it a hybrid of what has come before?
Skylar Tibbits (ST): I think it’s a convergence of a lot of technologies and developments. I actually got started on this when I came to MIT. I was on a Programmable Matter Grant from DARPA under Director Neil Gershenfeld at the Center for Bits and Atoms and there were a lot of different people on that grant. If you look at every one of them today, they’ve transitioned from the notion at the time, which is nearly 10 years ago now, that robots were programmable matter, that if we just got robots to be autonomous, modular, and small enough…then that was programmable matter. And almost every single one of those researchers, we just had a summit here at MIT called Active Matter Summit, a lot of those people were there (it almost became a joke, we were calling it the DARPA Summit because it was a bit of a reunion). Every one of those researchers had transitioned away from that [idea] and is now doing active, soft, highly agile, flexible, reconfigurable things.
Which brings me back to this idea of materials. You can look at the software revolution in any space, but in the design space specifically, this led to many of us thinking, “Okay, after I produce this crazy thing in Maya or Rhino, or Grasshopper, or whatever it is, what do I do with it?” So that led to this whole boom of digital fabrication, every school had CNC routers, laser cutters, more recently there’s an arms race with robot arms, for example, and that led to more people focusing on materials. It’s like, okay, we have all these crazy machines, well now what do I do?
I want to actually investigate the materials we build with. How can I make better materials? How can I print with new materials? How can I make hybrids, etc.? You know EmTech has been doing that for a long time, and many of us have been investigating that, but there’s a lineage centuries old of looking at materials. In some ways I think we lost that ability, maybe because of the industrial revolution. We just standardized stuff, so we look at 2×4 construction with wood as a standard, industrially produced material instead of looking at the anisotropic properties of it, for example. But now there’s this resurgence, both because we have computing possibilities, we have machine possibilities to fabricate things. We’re now interested in the properties of materials as well as the revolutions that are happening in synthetic biology and material science right now. Those two fields are booming. So there’s this huge technological development happening in materials, which fundamentally allow us to program them like we did computing and machines before.
Evan Greenberg (EG): That’s a great segue into my next question. You talk about the arms race with robotic arms and all of the different kind of fabrication tools that we have and the techniques that we’re using now. But how do you see the future of materials? Is it about innovating around the materials themselves or the processes that we design to act upon them?
ST: I think all of it. There are material scientists that are making amazing developments and fundamental new materials. There’s the Center for Directed Self-assembly, which is just looking at how you arrange particles to create new formations. Like Graphene for example, how you invent new materials just by rearranging the fundamental components. You can go to synthetic biology and you can look at programming bacteria as another world (for example, the iGem competitions). People are looking at materials in every single length and scale, every single domain up to the macroscale perspective, which is what we take. It’s about how you combine those to get new properties. We don’t necessarily invent fundamental new materials but rather we try to combine in new ways. And so then the question becomes, what do you want the material to do? What is the application you’re going for? What kind of radical properties can you get? And also, how do you design around it? So then that comes back to the tools. What kind of design tools, machine tools, hardware tools do we need in order to program these materials? So I think it’s happening because of the convergence but then they’re all feeding back to one another. Just like the hardware boom is feeding back to software, and wouldn’t have happened without software, and is now leading to new software developments, etc. They’re all kind of intertwined.
GA: So if there’s, let’s call it a new approach to materiality that’s enabled by other trends that are happening by digital, physical, robotic properties, etc. Taking it back to process and approach: are there key things in regards to this approach to materials that design professionals today need to keep in mind in order to be successful?
ST: It’s a good question. I don’t know that I have a direct answer for it other than to say a lot of times it feels like we’re in the early days of chemistry or in the early days of cooking, or the early days of any of these art forms, gardening or whatever. I look at these domains because, let’s say gardening or cooking: you have no control over the local interactions. The only things you have control over are the ingredients and the energy. By understanding the ingredients and the interaction with their environment, you can produce things you wouldn’t have been able to produce otherwise. So, you can’t sculpt the flower, it just won’t work, there’s no possible way to do that. The only thing you can do is let it grow by giving it the right environment–kind of dancing with it. You’re really collaborating with the materials and the energy. So I look at those domains, but think about the early days. We’re just screwing around and we’re just trying to figure it out. We’re throwing shit together and seeing what happens and trying to build up a language, trying to build up the tools. We don’t even have the microwave; we don’t even have the fundamental tools of the oven or whatever it is. So everyone is kind of out in different domains working on different scales in different applications. There is no fundamental ingredients cookbook that we all share. There are not even software tools that go between our lab and a biologist’s lab or a material science lab. We’re trying to build all that stuff up, but it’s going to take a while. I think on the actual materials transformation side we’re a lot further along. On the assembly side it’s like the way early days.
GA: What about lessons? What are one or two lessons someone who is newly approaching materials in this way might use to get started?
ST: One relates to issues of entropy–talking about forcing something vs. dancing with something. We always go back to that. The more we try to force something to happen, the worse it gets. So maybe two quick examples. If you look at the flasks that you shake and they assemble. If you try to make it assemble, you’re super bad. It’s like the worst you can do. If you close your eyes, you’re actually really good. So you have to find a way to let go. Release control and good things can happen and be open to surprises. Let it move, take an opportunity when you see a surprise and then kind of dance with it, both as a design thing as well as an assembly thing.
Another example is that I look back at precision or tolerance control. If you want to build a structure, the more you try to force accuracy in every single component, the worse you’re going to be. So what you really need to do is find a way for redundancy to creep in, for error correction to creep in, and use your degrees of freedom in the right way. There are lots of principles in mechanical engineering about that. As a simple example, if you want to build a circle out of discreet elements you probably shouldn’t try to precisely control that angle at every single step because by the time you get around you’re going to be way off because there is tolerance in every one. But if you allow a little bit of freedom, as it comes back it’s going to then close itself as it error corrects as a whole. That’s a principle that we try to achieve. Not forcing everything. Letting go of control a little bit so you can potentially achieve better results.
EG: In discussing the issue of tolerance, this brings back scale and how you deal with material at different scales. You showed a series of projects that crossover from the handheld or desktop experiment to the much larger balloon experiment. A lot of your references also maybe materially don’t scale. How do you scale through logic and what do you see as your limitations?
ST: A core belief that I have and that I’m trying to instill in everyone in the lab is that the principles scale, nothing else scales. The materials don’t scale, the forces don’t scale, but the principle scales. Every time you switch scales, you need to change materials. You need to rethink force. We have gravity at human scales and nanoscales don’t have that. We don’t have van der waal forces and they do. You probably can’t use the same material at the nanoscale that you can at the macroscale. But if you don’t look at it as a negative thing and you’re not trying to literally translate everything, but you’re trying to translate the principle, then self-assembly works at every possible scale. You need to change things so you see it as work at the scale that you’re given. You look at the right environments, the right forces, the right interactions and materials and then it works. So in that sense logic and information are scale independent. There are a lot of references about looking at information as the thing that translates across all of it. How do you embed information into the construction process? How do those components know how to error correct or know when they’re the right one? Where’s the blueprint in the parts so that it adds up to the right thing? All of that scales, so the question is: how do you embed information in DNA or how do you embed information at the cell level? How do you embed information at the organism level? How do you embed information at a macroscale or synthetic material level? How do you embed it at an astrophysics level? Information scales fundamentally, so do the phenomena that we’re working with, but you’ve got to change all of the other dynamics.
GA: My next question is going up to about 10,000 feet, not at the conceptual level, but at the organization level. So you’ve got the Self-assembly Lab at MIT with is its own organization that’s in the midst of a lot of really interesting things going on, not only at Cambridge but in the related fields that you talked about. You mentioned mechanical engineering, biology, and the molecular sciences. One of the first things you said you did when you were looking to explore some of the applied experiments was to gather a set of collaborators. Is the cross-disciplinary nature of the Self-assembly Lab something that is absolutely essential to the work that you’re doing and if so how does that mindset impact how you go about developing the experiments in time?
ST: Yeah, it’s fundamental to it. We’re in a center called the International Design Center at MIT. It’s quite unique because it’s not related to any other departments and it’s not a department itself. It’s an independent center and all the labs that come there and are housed there are from all different departments: mechanical engineering, biology, material science, architecture, etc. All of the students that come there are from every field, every department you can think of. In the lab we have students from every domain like computer science, architecture and design, material science, mechanical engineering, astrophysics, all different domains. The undergrads at MIT are just super brilliant. They have knowledge that is crazy and super inspiring and they converge here. We can teach them some things and they can teach us a lot of things. I think about it in the same way as when we collaborate. It’s trying to find a way to be comfortable in uncomfortable domains. It’s trying to use non-intuitive processes and find ways to make it intuitive. If you collaborate with molecular biologists or material scientists or physicists or whatever it is, you’re going to be uncomfortable, they’re going to be uncomfortable and awesome things usually happen from that. You learn a lot. You’re not going to be an expert in those domains but maybe you can make contributions collectively in both domains. So we try to be a bit humble when we’re in other domains. We don’t even really know what ours is. Like, what field are we? We don’t really now. But if you can collaborate with people you can achieve better things and then make contributions in many different areas. That’s how we approach it.
EG: You say you’re not really sure where you sit in the world, but I’m sure a question everyone wants to know is, what’s next? What’s on the horizon for the Self-assembly Lab?
ST: We have lots of crazy projects going on, both industrially sponsored projects of very real-world stuff like apparel and sportswear and furniture, to some building materials and infrastructure, orthodontics work, and medical stuff. Those are the ones that are super applied. Then there is the super weird, wild, utopian fundamental research, just trying to understand these things, scale it up, look at new properties, new environments and all sorts of collaborations in between. There are some larger proto-architectural works and all sorts of different things. We never really know where we’re going but it’s an endless search. We’re always trying to find different opportunities and ways to push the boundaries of different domains. We’re hopefully never making incremental progress, but we want every step to be large, radical progress.
For more information:
The Embedded Intelligence Lecture Series aims to interrogate three main lines of inquiry — material systems, natural systems, and machinic systems — which comprise the notion of Embedded Intelligence in architecture and design through an online platform of live-streaming lectures. Each presenter will deliver a lecture built around a series of curated questions pertaining to the subject matter, streamed live via Google Hangouts on Air and the Architectural Association’s web platform.
In addition to Skylar’s lecture, David Benjamin discussed Biological Intelligence, while Michael Weinstock spoke on Fabrication Intelligence.
For more information on Embedded Intelligence or to take part in this summer’s Visiting School visit newyork.aaschool.ac.uk
Embedded Intelligence, AA Visiting School New York
Embedded Intelligence Lecture Series
Material Intelligence Lecture Video
Biological Intelligence Lecture Video
Fabrication Intelligence Lecture Video
David Benjamin, The Living
Michael Weinstock, Emergent Technologies and Design
Skylar Tibbits, MIT Self-Assembly Lab
Embedded Intelligence on AA Conversations