Anima Anandkumar is a world-renowned AI researcher with brilliant answers to some of the most complicated AI questions. In this episode of More Intelligent Tomorrow, we talk about AI’s role in science and how future AGI systems might behave.
When we think of AGI we think of massive brains and massive compute but more and more of the computation will move to the edge.
Our brains use much less power than current GPUs do, how can we create more effective hardware and training and inference on the edge?
How do you define intelligence? For me it is the ability to learn, to learn and adapt. For instance, the Boston Dynamics robot can do impressive dancing, backflips, open the door, but it is pre-programmed to do exactly that.