 Picasso once said, computers are useless. All they can do is answer questions. Oh my goodness. And he was right. We need humans to ask the timely and important questions that really matter. Like will machines make better decisions than humans? Recently everyone from Elon Musk to Stephen Hawking have advised about the concerns about the singularity. So today I want to propose a new idea, the multiplicity, where diverse groups of humans work together with diverse groups of machines to make ideal decisions. And this idea has a long history. If we went back 300 years to a village near here in the Alps, we'd find a Swiss watchmaker in his studio crafting an automaton to study the changing relationships between men and machines. And there's been an enormous amount of research since then. And today we routinely trust machines like autopilots and pacemakers to make crucially important decisions. Machines have an enormous advantage in terms of speed and precision. And today the cutting edge of research in places like UC Berkeley are on machines that can learn. For example, it was recently shown that a sufficiently diverse group of machines will always learn to make better decisions than a single machine working alone. I study robots, the automata of our generation. And there are now over a million robots working in factories around the world, but we still don't have them in our homes. This is because of something called Moravex Paradox. What's very hard for humans, like giving a talk under intense time conditions, is actually very easy for robots. But what's easy for humans, like clearing the dinner table, is very hard for robots. Now if you put yourself in the position of being a robot, your sensors and motors are noisy and imprecise, everything around you is uncertain and inconsistent, including your own body. Outside the factory, the central problem for robots is uncertainty. Now consider Google's robot car. This car is making enormous progress. In fact, today it's actually capable of making better decisions than humans who might be sleepy or intoxicated or catching up on Instagram. But we'll know that these machines, by the way, are truly intelligent and we instruct them to go to work and they decide instead to go to the beach. So what Google figured out was that driving is similar to clearing the dishes. In both cases, robots can cope with uncertainty using spatial probability distributions, convolution and statistical learning to maximize expected utility. But the problem is that this requires enormous computation. Google's second insight is that the computation can be performed by remote processors in the cloud. This also allows robots to share information so that the collective learns and becomes smarter over time. This new paradigm of cloud robotics is revolutionizing my field. I'll give you three quick examples from my own lab. My students and I as an experiment set up the first cloud robot. We installed an industrial arm in the center of a garden and invited anyone in the world to help us come in and plant and water scenes. Well, over a hundred thousand humans showed up and they learned how to work together to keep this garden alive for over nine years. We're now studying how cloud robots can assist surgeons in the operating room. And this fall we were able to show for the first time that a surgical robot could learn to perform procedures like this by observing diverse examples provided by human surgeons. The key to learning is diversity. Now, my third point goes back to when I was born in Nigeria. My parents were teaching there and on a recent visit back to West Africa, I met students whose interest in science, technology, engineering and math was being inspired by robots. So with a professor from Ghana, we formed the African Robotics Network with the goal to design an ultra affordable robot for education. Many ideas came in, but the best one was someone took a surplus game controller, modified it into a robot car and attached two lollipops to detect contact. Our hope is that this lollipop will now increase the diversity for the next generation of roboticists. So rather than AI, what we need is IA, intelligence amplification. So in the spirit of Picasso, let's smash this singularity and instead consider the multiplicity where diverse groups of humans ask the questions that matter and then work with diverse groups of machines to make the best decisions. Thank you.