AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Challenges of Large Language Models
Microsoft has invested $10 billion in GPT. The company is training models from scratch, bigger and bigger. It's very strange as a human to interact via natural language with a system that produces correct sounding answers - but sometimes just correct sounding answers. Are we on a trajectory that converges towards models like this being able to evaluate truthfulness?