9:59:59.000,9:59:59.000 How many of your are creatives? 9:59:59.000,9:59:59.000 Designers, engineers,[br]entrepreneurs, artists, 9:59:59.000,9:59:59.000 or maybe you just have[br]a really big imagination. 9:59:59.000,9:59:59.000 Show of hands? 9:59:59.000,9:59:59.000 That's most of you. 9:59:59.000,9:59:59.000 I have some news for us creatives. 9:59:59.000,9:59:59.000 Over the course of the next 20 years, 9:59:59.000,9:59:59.000 more will change around the way[br]we do our work 9:59:59.000,9:59:59.000 than has happened in the last 2,000. 9:59:59.000,9:59:59.000 In fact, I think we're at the dawn[br]of a new age in human history. 9:59:59.000,9:59:59.000 Now, there have been four major historical[br]eras defined by the way we work. 9:59:59.000,9:59:59.000 The hunter-gatherer age lasted[br]several million years. 9:59:59.000,9:59:59.000 And then the Agricultural Age[br]lasted several thousand years. 9:59:59.000,9:59:59.000 The Industrial Age lasted[br]a couple of centuries, 9:59:59.000,9:59:59.000 and now the Information Age[br]has lasted just a few decades. 9:59:59.000,9:59:59.000 And now today, 9:59:59.000,9:59:59.000 we're on the cusp of our next[br]great era as a species. 9:59:59.000,9:59:59.000 Welcome to the Augemented Age. 9:59:59.000,9:59:59.000 In this new era, 9:59:59.000,9:59:59.000 your natural human capabilities are going[br]to be augmented by computational systems 9:59:59.000,9:59:59.000 that help you think, 9:59:59.000,9:59:59.000 robotic systems that help you make, 9:59:59.000,9:59:59.000 and a digital nervous system 9:59:59.000,9:59:59.000 that connects you to the world[br]fay beyond your natural senses. 9:59:59.000,9:59:59.000 Let's stated with Cognitive Augmentation. 9:59:59.000,9:59:59.000 How many of you are augmented cyborgs? 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 I would actually argue that we're[br]already augmented. 9:59:59.000,9:59:59.000 Imagine you're at a party, 9:59:59.000,9:59:59.000 and somebody asks you a question[br]that you don't know the answer to. 9:59:59.000,9:59:59.000 If you have one of these, 9:59:59.000,9:59:59.000 in a few seconds, 9:59:59.000,9:59:59.000 you can know the answer. 9:59:59.000,9:59:59.000 But this is just a primitive beginning. 9:59:59.000,9:59:59.000 Even Siri is just a passive tool. 9:59:59.000,9:59:59.000 In fact, for the past[br]three-and-a-half million years, 9:59:59.000,9:59:59.000 the tools that we've had[br]have been completely passive. 9:59:59.000,9:59:59.000 They do exactly what we tell them[br]and nothing more. 9:59:59.000,9:59:59.000 Our very first tool only cut[br]where we struck it. 9:59:59.000,9:59:59.000 The chisel only carves where[br]the artist points it. 9:59:59.000,9:59:59.000 And even our most advanced tools[br]do nothing without out explicit direction. 9:59:59.000,9:59:59.000 In fact, to date -- 9:59:59.000,9:59:59.000 this is something that frustrates me -- 9:59:59.000,9:59:59.000 we've always been limited 9:59:59.000,9:59:59.000 by this need to manually[br]push our wills into our tools, 9:59:59.000,9:59:59.000 like manual -- 9:59:59.000,9:59:59.000 like literally using our hands, 9:59:59.000,9:59:59.000 even with computers. 9:59:59.000,9:59:59.000 But I'm more like Scotty in "Star Trek." 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 I want to have a conversation[br]with a computer. 9:59:59.000,9:59:59.000 I want to say, "Computer,[br]let's design a car," 9:59:59.000,9:59:59.000 and the computer shows me a car. 9:59:59.000,9:59:59.000 And I say, "No, more fast-looking,[br]and less German," 9:59:59.000,9:59:59.000 and bang, the computer shows me an option. 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 That conversation might be[br]a little ways off, 9:59:59.000,9:59:59.000 it's actually probably less[br]than any of us think, 9:59:59.000,9:59:59.000 but right now, 9:59:59.000,9:59:59.000 we're working on it. 9:59:59.000,9:59:59.000 Tools are making this leap from being[br]passive to being generative. 9:59:59.000,9:59:59.000 Generative design tools[br]use a computer and algorithms 9:59:59.000,9:59:59.000 to synthesize geometry -- 9:59:59.000,9:59:59.000 to come up with new designs[br]all by themselves. 9:59:59.000,9:59:59.000 All it needs are your goals[br]and your constraints. 9:59:59.000,9:59:59.000 I'll give you an example. 9:59:59.000,9:59:59.000 In the case of this Aerial[br]Drone [Chassy?], 9:59:59.000,9:59:59.000 all you would need to do is tell it[br]something like, 9:59:59.000,9:59:59.000 it has four propellers, 9:59:59.000,9:59:59.000 you want it to be[br]as lightweight as possible, 9:59:59.000,9:59:59.000 and you need it to be[br]aerodynamically efficient. 9:59:59.000,9:59:59.000 And then what the computer does 9:59:59.000,9:59:59.000 is it explores the entire solution space: 9:59:59.000,9:59:59.000 every single possibility that solves[br]and meets your criteria -- 9:59:59.000,9:59:59.000 millions of them. 9:59:59.000,9:59:59.000 It takes big computers to do this, 9:59:59.000,9:59:59.000 but it comes back to us with designs 9:59:59.000,9:59:59.000 that we by ourselves[br]never could've imagined. 9:59:59.000,9:59:59.000 And the computer's coming up[br]with this stuff all by itself, 9:59:59.000,9:59:59.000 no one ever drew anything, 9:59:59.000,9:59:59.000 and it started completely from scratch. 9:59:59.000,9:59:59.000 And by the way, 9:59:59.000,9:59:59.000 it's no accident 9:59:59.000,9:59:59.000 that the drone body looks just like[br]the pelvis of a flying squirrel. 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 It's because the algorithms are designed[br]to word the same way that evolution does. 9:59:59.000,9:59:59.000 What's exciting is we're starting to see[br]this technology out in the real world. 9:59:59.000,9:59:59.000 We've been working with [Airbus][br]for a couple of years 9:59:59.000,9:59:59.000 on this concept plane for the future. 9:59:59.000,9:59:59.000 It's aways out still, 9:59:59.000,9:59:59.000 but just recently we used[br]a generative design AI 9:59:59.000,9:59:59.000 to come up with this. 9:59:59.000,9:59:59.000 This is a 3D printed cabin partition[br]that's been designed by a computer. 9:59:59.000,9:59:59.000 It's stronger than the original[br]yet half the weight, 9:59:59.000,9:59:59.000 and it will be flying[br]in the Airbus [A-320] later this year. 9:59:59.000,9:59:59.000 So computers can now generate. 9:59:59.000,9:59:59.000 They can come up with their own solutions[br]to our well-defined problems. 9:59:59.000,9:59:59.000 But they're not inutitive. 9:59:59.000,9:59:59.000 They still have to start from scratch[br]every single time, 9:59:59.000,9:59:59.000 and that's because they never learn, 9:59:59.000,9:59:59.000 unlike Maggie. 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 Maggie's actually smarter than our[br]most advanced design tools. 9:59:59.000,9:59:59.000 What do I mean by that? 9:59:59.000,9:59:59.000 If her owner picks up that leash, 9:59:59.000,9:59:59.000 Maggie knows with a fair[br]degree of certainty 9:59:59.000,9:59:59.000 that it's time to for a walk. 9:59:59.000,9:59:59.000 And how did she learn? 9:59:59.000,9:59:59.000 Well, every time the owner[br]picked up the leash, 9:59:59.000,9:59:59.000 they went for a walk. 9:59:59.000,9:59:59.000 And Maggie did three things: 9:59:59.000,9:59:59.000 she had to pay attention, 9:59:59.000,9:59:59.000 she had to remember what happened 9:59:59.000,9:59:59.000 and she had to retain and create[br]a pattern in her mind. 9:59:59.000,9:59:59.000 Interestingly, 9:59:59.000,9:59:59.000 that's exactly computer scientists[br]have been trying to get AIs to do 9:59:59.000,9:59:59.000 for the last 60 or so years. 9:59:59.000,9:59:59.000 Back in 1952, 9:59:59.000,9:59:59.000 they build this computer[br]that could play Tic-Tac-Toe. 9:59:59.000,9:59:59.000 Big deal. 9:59:59.000,9:59:59.000 Then 45 years later, 9:59:59.000,9:59:59.000 1997, 9:59:59.000,9:59:59.000 Deep Blue beats [Casperoff] at Chess. 9:59:59.000,9:59:59.000 2011, Watson beats these two[br]humans at Jeopardy, 9:59:59.000,9:59:59.000 which is much harder for a computer[br]to play than Chess is. 9:59:59.000,9:59:59.000 In fact, rather than working[br]from predefined recipes, 9:59:59.000,9:59:59.000 Watson had to use reasoning[br]to overcome his human opponents. 9:59:59.000,9:59:59.000 And then a couple of weeks ago, 9:59:59.000,9:59:59.000 [Deep Minds' Alpha-Go] beats[br]the world's best human at Go, 9:59:59.000,9:59:59.000 which is the most difficult[br]game that we have. 9:59:59.000,9:59:59.000 In fact in Go, 9:59:59.000,9:59:59.000 there are more possible moves[br]than there are atoms in the universe. 9:59:59.000,9:59:59.000 So in order to win, 9:59:59.000,9:59:59.000 what Alpha-Go had to do[br]was develop intuition, 9:59:59.000,9:59:59.000 and in fact, 9:59:59.000,9:59:59.000 at some points, 9:59:59.000,9:59:59.000 Alpha-Go's programmers didn't understand[br]why Alpha-Go was doing what it was doing. 9:59:59.000,9:59:59.000 And things are moving really fast. 9:59:59.000,9:59:59.000 I mean, consider -- 9:59:59.000,9:59:59.000 in the space of a human lifetime, 9:59:59.000,9:59:59.000 computers have gone from a child's game 9:59:59.000,9:59:59.000 to what's recognized as the pinnacle[br]of strategic thought. 9:59:59.000,9:59:59.000 What's basically happening 9:59:59.000,9:59:59.000 is computers are going[br]from being like Spock 9:59:59.000,9:59:59.000 to being a lot more like Kirk. 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 Right? 9:59:59.000,9:59:59.000 From pure logic to intuition. 9:59:59.000,9:59:59.000 Would you cross this bridge? 9:59:59.000,9:59:59.000 Most of you are saying, 9:59:59.000,9:59:59.000 "Oh, hell no." 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 And you arrived at that decision[br]in a split second. 9:59:59.000,9:59:59.000 You just sort of knew[br]that that bridge was unsafe. 9:59:59.000,9:59:59.000 And that's exactly the kind of intuition 9:59:59.000,9:59:59.000 that our deep learning systems[br]are starting to develop right now. 9:59:59.000,9:59:59.000 Very soon you'll literally be able[br]to show something that you've made, 9:59:59.000,9:59:59.000 you've designed, 9:59:59.000,9:59:59.000 to computer, 9:59:59.000,9:59:59.000 and it will look at it and say, 9:59:59.000,9:59:59.000 "Mm, sorry homey, that will never work, 9:59:59.000,9:59:59.000 you'll have to try again." 9:59:59.000,9:59:59.000 Or you could ask it if people[br]are going to like your next song, 9:59:59.000,9:59:59.000 or your next flavor of ice cream. 9:59:59.000,9:59:59.000 Or, much more importantly, 9:59:59.000,9:59:59.000 you could work with a computer to solve[br]a problem that we've never faced before. 9:59:59.000,9:59:59.000 For instance climate change. 9:59:59.000,9:59:59.000 We're not doing a very[br]good job on our own, 9:59:59.000,9:59:59.000 we could certainly use[br]all the help we can get. 9:59:59.000,9:59:59.000 That's what I'm talking about: 9:59:59.000,9:59:59.000 technology amplifying[br]our cognitive abilities 9:59:59.000,9:59:59.000 so we can imagine and design things 9:59:59.000,9:59:59.000 that were simply out of our reach[br]as plain old unaugmented humans. 9:59:59.000,9:59:59.000 So, what about making[br]all of this crazy new stuff 9:59:59.000,9:59:59.000 that we're going to invent and design? 9:59:59.000,9:59:59.000 I think the era of human augmentation[br]is as much about the physical world 9:59:59.000,9:59:59.000 as it is about the virtual,[br]intellectual realm. 9:59:59.000,9:59:59.000 So how will technology augment us? 9:59:59.000,9:59:59.000 In the physcial world, 9:59:59.000,9:59:59.000 robotic systems. 9:59:59.000,9:59:59.000 OK, there's certainly a fear that robots[br]are going to take jobs from humans, 9:59:59.000,9:59:59.000 and that is true in certain sectors. 9:59:59.000,9:59:59.000 But I'm much more interested in this idea 9:59:59.000,9:59:59.000 that humans and robots working together[br]are going to augment each other, 9:59:59.000,9:59:59.000 and start to inhabit a new space. 9:59:59.000,9:59:59.000 This is our applied[br]research lab in San Francisco, 9:59:59.000,9:59:59.000 where one of our areas of focus[br]is advanced robotics, 9:59:59.000,9:59:59.000 specifically human-robot collaboration. 9:59:59.000,9:59:59.000 And this is Bishop, 9:59:59.000,9:59:59.000 one of our robots. 9:59:59.000,9:59:59.000 As an experiment, 9:59:59.000,9:59:59.000 we set it up to help a person working[br]in construction doing repetitive tasks. 9:59:59.000,9:59:59.000 Tasks like cutting out holes for outlets[br]or light switches in dry wall. 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 So Bishop's human partner can tell[br]what to do in plain English 9:59:59.000,9:59:59.000 and simple gestures, 9:59:59.000,9:59:59.000 kind of like talking to a dog. 9:59:59.000,9:59:59.000 And then Bishop executes[br]on those instructions 9:59:59.000,9:59:59.000 with perfect precision. 9:59:59.000,9:59:59.000 We're using the human for what[br]the human is good at, 9:59:59.000,9:59:59.000 right? 9:59:59.000,9:59:59.000 Awareness, perception and desicion making. 9:59:59.000,9:59:59.000 And we're using the robot[br]for what it's good at: 9:59:59.000,9:59:59.000 precision and repetitiveness. 9:59:59.000,9:59:59.000 Here's another cool project[br]that Bishop worked on. 9:59:59.000,9:59:59.000 The goal of this project, 9:59:59.000,9:59:59.000 which we called, "The Hive," 9:59:59.000,9:59:59.000 was the prototype the experience[br]of humans, computers and robots 9:59:59.000,9:59:59.000 all working together to solve[br]a highly complex design problem. 9:59:59.000,9:59:59.000 The humans acted as labor. 9:59:59.000,9:59:59.000 They cruised around the construction site, 9:59:59.000,9:59:59.000 they manipulated the bamboo -- 9:59:59.000,9:59:59.000 which by the way, 9:59:59.000,9:59:59.000 because it's a non-icomorphic material, 9:59:59.000,9:59:59.000 is super hard for robots to deal with. 9:59:59.000,9:59:59.000 But then the robots[br]did this fiber winding, 9:59:59.000,9:59:59.000 which was almost impossible[br]for a human to do. 9:59:59.000,9:59:59.000 And then we had an AI[br]that was controlling everything. 9:59:59.000,9:59:59.000 It was telling the humans what to do, 9:59:59.000,9:59:59.000 it was tell the robots what to do, 9:59:59.000,9:59:59.000 and keeping track of thousands[br]of individual components. 9:59:59.000,9:59:59.000 What's interesting is building[br]this pavillion was simply not possible 9:59:59.000,9:59:59.000 without human, robot and AI[br]augmenting each other. 9:59:59.000,9:59:59.000 OK, I'll share one more project. 9:59:59.000,9:59:59.000 This one's a little bit crazy. 9:59:59.000,9:59:59.000 We're working with Amsterdam-based artist,[br][Youris Marvin and his team at MX3D] 9:59:59.000,9:59:59.000 to generatively design[br]and robotically print 9:59:59.000,9:59:59.000 the world's first autonomously[br]manufactured bridge. 9:59:59.000,9:59:59.000 So, Youris and an AI are designing[br]this thing right now, as we speak, 9:59:59.000,9:59:59.000 in Amsterdam. 9:59:59.000,9:59:59.000 And when we're done, 9:59:59.000,9:59:59.000 we're going to hit Go, 9:59:59.000,9:59:59.000 and robots will start 3D printing[br]in stainless steel, 9:59:59.000,9:59:59.000 and then they're going to keep printing[br]without human intervention 9:59:59.000,9:59:59.000 until the bridge is finished. 9:59:59.000,9:59:59.000 So as computers are going[br]to augment our ability 9:59:59.000,9:59:59.000 to imagine and design new stuff, 9:59:59.000,9:59:59.000 robotic systems are going to help us 9:59:59.000,9:59:59.000 build and make things that we've[br]never been able to make before. 9:59:59.000,9:59:59.000 But what about our ability[br]to sense and control these things? 9:59:59.000,9:59:59.000 What about a nervous system[br]for the things that we make? 9:59:59.000,9:59:59.000 Our nervous system -- 9:59:59.000,9:59:59.000 the human nervous system -- 9:59:59.000,9:59:59.000 tells everything that's[br]going on around us. 9:59:59.000,9:59:59.000 But the nervous system of the things[br]we make is rudimentary at best. 9:59:59.000,9:59:59.000 For instance, 9:59:59.000,9:59:59.000 a car doesn't tell the city's[br]Public Works department 9:59:59.000,9:59:59.000 that it just hit a pothole at the corner[br]of Broadway and Morrison; 9:59:59.000,9:59:59.000 A building doesn't tell its designers 9:59:59.000,9:59:59.000 whether or not the people[br]inside like being there, 9:59:59.000,9:59:59.000 and the toy manufacturer doesn't know[br]if a toy is actually being played with -- 9:59:59.000,9:59:59.000 how and where and whether[br]or not it's any fun. 9:59:59.000,9:59:59.000 Look, I'm sure that the designers[br]imagined this lifestyle for Barbie 9:59:59.000,9:59:59.000 when they designed her -- 9:59:59.000,9:59:59.000 right? 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 But what if it turns out that Barbie's[br]actually really lonely? 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 If the designers had known what was[br]really happening in the real world 9:59:59.000,9:59:59.000 with their designs -- 9:59:59.000,9:59:59.000 the road, the building, the Barbie -- 9:59:59.000,9:59:59.000 they could've used that knowledge 9:59:59.000,9:59:59.000 to create an experience[br]that was better for the user. 9:59:59.000,9:59:59.000 What's missing is a nervous system 9:59:59.000,9:59:59.000 connecting us to all of the things[br]that we design, make and use. 9:59:59.000,9:59:59.000 What if all of you had that kind[br]of information flowing to you 9:59:59.000,9:59:59.000 from the things you create[br]in the real world? 9:59:59.000,9:59:59.000 With all of the stuff we make, 9:59:59.000,9:59:59.000 we spend a tremendous amount[br]of money and energy -- 9:59:59.000,9:59:59.000 in fact last year about[br]two trillion dollars -- 9:59:59.000,9:59:59.000 convincing people to buy[br]the things that we've made. 9:59:59.000,9:59:59.000 But if you had this connection[br]to the things that you design and create, 9:59:59.000,9:59:59.000 after they're out in the real world, 9:59:59.000,9:59:59.000 after they've been sold,[br]or launched or whatever, 9:59:59.000,9:59:59.000 we could actually change that, 9:59:59.000,9:59:59.000 and go from making people want our stuff, 9:59:59.000,9:59:59.000 to just making stuff that people want[br]in the first place. 9:59:59.000,9:59:59.000 The good news is we're working[br]on digital nervous systems 9:59:59.000,9:59:59.000 that connect us to the things we design. 9:59:59.000,9:59:59.000 We're working on one project[br]with a couple of guys down in Los Angeles 9:59:59.000,9:59:59.000 called the Bandito Brothers, 9:59:59.000,9:59:59.000 and their team, 9:59:59.000,9:59:59.000 and one of the things these guys do[br]is build insane cars 9:59:59.000,9:59:59.000 that do absolutely insane things. 9:59:59.000,9:59:59.000 These guys are crazy. 9:59:59.000,9:59:59.000 (Laughter) 9:59:59.000,9:59:59.000 In the best way. 9:59:59.000,9:59:59.000 And what we're doing with them 9:59:59.000,9:59:59.000 is taking a traditions racecar [Chassy], 9:59:59.000,9:59:59.000 and giving it a nervous system. 9:59:59.000,9:59:59.000 So we instrumented it[br]with dozens of censors, 9:59:59.000,9:59:59.000 and then we put a world-class driver[br]behind the wheel, 9:59:59.000,9:59:59.000 took it out to the desert 9:59:59.000,9:59:59.000 and drove the hell out of it for a week. 9:59:59.000,9:59:59.000 And the car's nervous system captured[br]everything that was happening to car. 9:59:59.000,9:59:59.000 We captured four billion data points. 9:59:59.000,9:59:59.000 All of the forces[br]that it was subjected to. 9:59:59.000,9:59:59.000 And then we did something crazy. 9:59:59.000,9:59:59.000 We took all of that data, 9:59:59.000,9:59:59.000 and plugged it into a generative[br]design AI that we call, "Dreamcatcher." 9:59:59.000,9:59:59.000 So what do get when you give[br]a design tool a nervous system, 9:59:59.000,9:59:59.000 and you ask it to build you[br]the ultimate car [chassy]? 9:59:59.000,9:59:59.000 You get this. 9:59:59.000,9:59:59.000 This is something that human[br]could never have designed. 9:59:59.000,9:59:59.000 Except a human did design this, 9:59:59.000,9:59:59.000 but it was a human that was augmented[br]by a generative design AI, 9:59:59.000,9:59:59.000 a digital nervous system, 9:59:59.000,9:59:59.000 and robots that can actually[br]fabricate something like this.