AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Can We Collect a Large Quantum of Data That Can Be Labeled Well?
The question Tesla asks is, can we create a data engine that keeps sending an imperfect machine learning system out there? There are 17 different acronyms of different ways you can try to apply this. And in reinforcement learning, I feel like the consensus is like, it totally depends. So even at these like extremely large scales, if we are not yet touching what reinforcement learning can truly do, imagine what the next 10 years looks like. It's amazing.