- Large Language Models are trained on the largest amount of language possible, so everything on the internet. It uses that training to generate completely new information, such as writing essays to poems to even writing code. Chatbot magic is based on simple math concepts that uses probability. On top of that, it needs context so it needs a neural network. People train these neural networks to predict what is going to come next. It has LOTS of human training.
- The difference between generative and predictive AI is the data. Predictive AI does not make new data. It creates new content using historical data and patterns. This helps it make predictions from past data. Generative AI creates (generates) new content that is based on learned patterns from the training data.
