0:00:00.346,0:00:03.116 This is me building a prototype 0:00:03.116,0:00:06.036 for six hours straight. 0:00:06.036,0:00:09.957 This is slave labor to my own project. 0:00:09.957,0:00:14.947 This is what the DIY and maker movements really look like. 0:00:14.947,0:00:19.713 And this is an analogy for today's construction and manufacturing world 0:00:19.713,0:00:22.501 with brute-force assembly techniques. 0:00:22.501,0:00:25.335 And this is exactly why I started studying 0:00:25.335,0:00:29.604 how to program physical materials to build themselves. 0:00:29.604,0:00:31.364 But there is another world. 0:00:31.364,0:00:33.359 Today at the micro- and nanoscales, 0:00:33.359,0:00:36.110 there's an unprecedented revolution happening. 0:00:36.110,0:00:40.242 And this is the ability to program physical and biological materials 0:00:40.242,0:00:42.966 to change shape, change properties 0:00:42.966,0:00:45.952 and even compute outside of silicon-based matter. 0:00:45.952,0:00:48.459 There's even a software called cadnano 0:00:48.459,0:00:51.242 that allows us to design three-dimensional shapes 0:00:51.242,0:00:54.326 like nano robots or drug delivery systems 0:00:54.326,0:00:58.629 and use DNA to self-assemble those functional structures. 0:00:58.629,0:01:00.702 But if we look at the human scale, 0:01:00.702,0:01:03.758 there's massive problems that aren't being addressed 0:01:03.758,0:01:06.126 by those nanoscale technologies. 0:01:06.126,0:01:08.225 If we look at construction and manufacturing, 0:01:08.225,0:01:12.418 there's major inefficiencies, energy consumption 0:01:12.418,0:01:14.745 and excessive labor techniques. 0:01:14.745,0:01:17.250 In infrastructure, let's just take one example. 0:01:17.250,0:01:18.733 Take piping. 0:01:18.733,0:01:22.458 In water pipes, we have fixed-capacity water pipes 0:01:22.458,0:01:26.917 that have fixed flow rates, except for expensive pumps and valves. 0:01:26.917,0:01:28.208 We bury them in the ground. 0:01:28.208,0:01:30.829 If anything changes -- if the environment changes, 0:01:30.829,0:01:33.466 the ground moves, or demand changes -- 0:01:33.466,0:01:37.716 we have to start from scratch and take them out and replace them. 0:01:37.716,0:01:41.084 So I'd like to propose that we can combine those two worlds, 0:01:41.084,0:01:46.255 that we can combine the world of the nanoscale programmable adaptive materials 0:01:46.255,0:01:47.932 and the built environment. 0:01:47.932,0:01:50.118 And I don't mean automated machines. 0:01:50.118,0:01:52.577 I don't just mean smart machines that replace humans. 0:01:52.577,0:01:56.461 But I mean programmable materials that build themselves. 0:01:56.461,0:01:58.517 And that's called self-assembly, 0:01:58.517,0:02:03.034 which is a process by which disordered parts build an ordered structure 0:02:03.034,0:02:05.527 through only local interaction. 0:02:05.527,0:02:08.702 So what do we need if we want to do this at the human scale? 0:02:08.702,0:02:10.677 We need a few simple ingredients. 0:02:10.677,0:02:13.516 The first ingredient is materials and geometry, 0:02:13.516,0:02:16.985 and that needs to be tightly coupled with the energy source. 0:02:16.985,0:02:18.643 And you can use passive energy -- 0:02:18.643,0:02:23.327 so heat, shaking, pneumatics, gravity, magnetics. 0:02:23.327,0:02:26.402 And then you need smartly designed interactions. 0:02:26.402,0:02:28.535 And those interactions allow for error correction, 0:02:28.535,0:02:32.535 and they allow the shapes to go from one state to another state. 0:02:32.535,0:02:35.767 So now I'm going to show you a number of projects that we've built, 0:02:35.767,0:02:39.051 from one-dimensional, two-dimensional, three-dimensional 0:02:39.051,0:02:42.118 and even four-dimensional systems. 0:02:42.118,0:02:44.004 So in one-dimensional systems -- 0:02:44.004,0:02:46.911 this is a project called the self-folding proteins. 0:02:46.911,0:02:51.618 And the idea is that you take the three-dimensional structure of a protein -- 0:02:51.618,0:02:54.299 in this case it's the crambin protein -- 0:02:54.299,0:02:57.945 you take the backbone -- so no cross-linking, no environmental interactions -- 0:02:57.945,0:03:01.096 and you break that down into a series of components. 0:03:01.096,0:03:03.370 And then we embed elastic. 0:03:03.370,0:03:05.711 And when I throw this up into the air and catch it, 0:03:05.711,0:03:10.562 it has the full three-dimensional structure of the protein, all of the intricacies. 0:03:10.562,0:03:12.662 And this gives us a tangible model 0:03:12.662,0:03:16.396 of the three-dimensional protein and how it folds 0:03:16.396,0:03:18.680 and all of the intricacies of the geometry. 0:03:18.680,0:03:22.128 So we can study this as a physical, intuitive model. 0:03:22.128,0:03:24.930 And we're also translating that into two-dimensional systems -- 0:03:24.930,0:03:29.229 so flat sheets that can self-fold into three-dimensional structures. 0:03:29.229,0:03:33.682 In three dimensions, we did a project last year at TEDGlobal 0:03:33.682,0:03:35.646 with Autodesk and Arthur Olson 0:03:35.646,0:03:37.344 where we looked at autonomous parts -- 0:03:37.344,0:03:41.663 so individual parts not pre-connected that can come together on their own. 0:03:41.663,0:03:44.429 And we built 500 of these glass beakers. 0:03:44.429,0:03:46.896 They had different molecular structures inside 0:03:46.896,0:03:49.120 and different colors that could be mixed and matched. 0:03:49.120,0:03:51.369 And we gave them away to all the TEDsters. 0:03:51.369,0:03:53.870 And so these became intuitive models 0:03:53.870,0:03:57.362 to understand how molecular self-assembly works at the human scale. 0:03:57.362,0:03:59.179 This is the polio virus. 0:03:59.179,0:04:01.111 You shake it hard and it breaks apart. 0:04:01.111,0:04:02.557 And then you shake it randomly 0:04:02.557,0:04:06.061 and it starts to error correct and built the structure on its own. 0:04:06.061,0:04:09.028 And this is demonstrating that through random energy, 0:04:09.028,0:04:13.656 we can build non-random shapes. 0:04:13.656,0:04:17.180 We even demonstrated that we can do this at a much larger scale. 0:04:17.180,0:04:19.334 Last year at TED Long Beach, 0:04:19.334,0:04:22.545 we built an installation that builds installations. 0:04:22.545,0:04:26.082 The idea was, could we self-assemble furniture-scale objects? 0:04:26.082,0:04:28.583 So we built a large rotating chamber, 0:04:28.583,0:04:31.796 and people would come up and spin the chamber faster or slower, 0:04:31.796,0:04:33.381 adding energy to the system 0:04:33.381,0:04:36.827 and getting an intuitive understanding of how self-assembly works 0:04:36.827,0:04:38.206 and how we could use this 0:04:38.206,0:04:42.952 as a macroscale construction or manufacturing technique for products. 0:04:42.952,0:04:44.619 So remember, I said 4D. 0:04:44.619,0:04:48.314 So today for the first time, we're unveiling a new project, 0:04:48.314,0:04:50.200 which is a collaboration with Stratasys, 0:04:50.200,0:04:52.035 and it's called 4D printing. 0:04:52.035,0:04:53.995 The idea behind 4D printing 0:04:53.995,0:04:56.995 is that you take multi-material 3D printing -- 0:04:56.995,0:04:59.184 so you can deposit multiple materials -- 0:04:59.184,0:05:01.068 and you add a new capability, 0:05:01.068,0:05:02.934 which is transformation, 0:05:02.934,0:05:04.251 that right off the bed, 0:05:04.251,0:05:08.578 the parts can transform from one shape to another shape directly on their own. 0:05:08.578,0:05:12.078 And this is like robotics without wires or motors. 0:05:12.078,0:05:13.729 So you completely print this part, 0:05:13.729,0:05:16.533 and it can transform into something else. 0:05:16.533,0:05:21.494 We also worked with Autodesk on a software they're developing called Project Cyborg. 0:05:21.494,0:05:24.617 And this allows us to simulate this self-assembly behavior 0:05:24.617,0:05:27.819 and try to optimize which parts are folding when. 0:05:27.819,0:05:30.549 But most importantly, we can use this same software 0:05:30.549,0:05:33.457 for the design of nanoscale self-assembly systems 0:05:33.457,0:05:36.300 and human scale self-assembly systems. 0:05:36.300,0:05:39.813 These are parts being printed with multi-material properties. 0:05:39.813,0:05:41.530 Here's the first demonstration. 0:05:41.530,0:05:43.434 A single strand dipped in water 0:05:43.434,0:05:45.783 that completely self-folds on its own 0:05:45.783,0:05:49.701 into the letters M I T. 0:05:49.701,0:05:51.523 I'm biased. 0:05:51.523,0:05:54.972 This is another part, single strand, dipped in a bigger tank 0:05:54.972,0:05:59.595 that self-folds into a cube, a three-dimensional structure, on its own. 0:05:59.595,0:06:01.446 So no human interaction. 0:06:01.446,0:06:03.335 And we think this is the first time 0:06:03.335,0:06:05.582 that a program and transformation 0:06:05.582,0:06:08.832 has been embedded directly into the materials themselves. 0:06:08.832,0:06:11.562 And it also might just be the manufacturing technique 0:06:11.562,0:06:15.696 that allows us to produce more adaptive infrastructure in the future. 0:06:15.696,0:06:16.882 So I know you're probably thinking, 0:06:16.882,0:06:21.114 okay, that's cool, but how do we use any of this stuff for the built environment? 0:06:21.114,0:06:23.369 So I've started a lab at MIT, 0:06:23.369,0:06:25.251 and it's called the Self-Assembly Lab. 0:06:25.251,0:06:28.395 And we're dedicated to trying to develop programmable materials 0:06:28.395,0:06:29.586 for the built environment. 0:06:29.586,0:06:31.549 And we think there's a few key sectors 0:06:31.549,0:06:33.863 that have fairly near-term applications. 0:06:33.863,0:06:35.918 One of those is in extreme environments. 0:06:35.918,0:06:38.464 These are scenarios where it's difficult to build, 0:06:38.464,0:06:40.952 our current construction techniques don't work, 0:06:40.952,0:06:44.512 it's too large, it's too dangerous, it's expensive, too many parts. 0:06:44.512,0:06:46.879 And space is a great example of that. 0:06:46.879,0:06:49.245 We're trying to design new scenarios for space 0:06:49.245,0:06:52.545 that have fully reconfigurable and self-assembly structures 0:06:52.545,0:06:56.235 that can go from highly functional systems from one to another. 0:06:56.235,0:06:58.346 Let's go back to infrastructure. 0:06:58.346,0:07:02.245 In infrastructure, we're working with a company out of Boston called Geosyntec. 0:07:02.245,0:07:05.045 And we're developing a new paradigm for piping. 0:07:05.045,0:07:08.668 Imagine if water pipes could expand or contract 0:07:08.668,0:07:11.418 to change capacity or change flow rate, 0:07:11.418,0:07:15.952 or maybe even undulate like peristaltics to move the water themselves. 0:07:15.952,0:07:18.562 So this isn't expensive pumps or valves. 0:07:18.562,0:07:22.845 This is a completely programmable and adaptive pipe on its own. 0:07:22.845,0:07:24.660 So I want to remind you today 0:07:24.660,0:07:28.044 of the harsh realities of assembly in our world. 0:07:28.044,0:07:31.509 These are complex things built with complex parts 0:07:31.509,0:07:34.294 that come together in complex ways. 0:07:34.294,0:07:37.493 So I would like to invite you from whatever industry you're from 0:07:37.493,0:07:41.546 to join us in reinventing and reimagining the world, 0:07:41.546,0:07:45.245 how things come together from the nanoscale to the human scale, 0:07:45.245,0:07:48.320 so that we can go from a world like this 0:07:48.320,0:07:51.270 to a world that's more like this. 0:08:00.632,0:08:02.542 Thank you. 0:08:02.542,0:08:04.843 (Applause)