In Machines We Trust: Exploring AI's evolution (Live)
MIT Technology Review Narrated
00:00
What Do You Think of the Progress in the Last Decade?
The revolution in natural language understanding systems that we see in large language models and stuff came from two major advances several years before. One is the very wide use of self-supervised learning to train those systems, so they're not trained to do a particular task. And then there was a new set of architectures called Transformers that really sort of brought a new type of flexibility in the models that are used. So, you don't see this progress as being continuous unless you are down in the trenches and you realize what's going on.
Transcript
Play full episode
