LLMs are trained using vast amounts of information sourced from across the internet. They rely on mathematical probabilities to produce new text, such as sentences, paragraphs, or stories. This process is inspired by how neurons in the brain work, which is why these models are called Neural Networks. Generative AI uses the patterns and data gathered by Predictive AI to create new content. Predictive AI identifies patterns from past information, and Generative AI then builds on this to generate new outputs, like images or text. However magical it seems, AI still requires "human tuning" to produce reasonable results.
