1. Algorithms
The logic and models that enable learning and decision-making.

2. Data
The raw material from which patterns and insights are derived.

3. Computing power
The engine that makes large-scale processing possible.

1. Supervised Learning
The model is trained on labelled data – datasets that include both inputs and correct outputs. Think of it like a teacher providing math problems with solutions so students can learn patterns and apply them to new problems.

2. Unsupervised Learning
Here, the data has no labels. The model identifies patterns or structures on its own, such as grouping similar items together. It’s like giving students a box of puzzle pieces without the final picture and asking them to figure out how they fit.

3. Reinforcement Learning
This method involves an agent interacting with an environment and learning through feedback – rewards or penalties. It’s similar to how children learn to ride a bike: falling, adjusting, and eventually mastering the skill through experience.

1. Tokenization
Text is broken down into discrete units called tokens. A token can be a word, subword, or even a character, depending on the model's design. 

2.  Token Embeddings
Each token is mapped to a high-dimensional vector, known as an embedding. These embeddings capture semantic and syntactic relationships between tokens, enabling the model to understand context and meaning.

3. Sequence Modeling
Using deep neural networks – typically transformer architectures – the model processes the sequence of token embeddings to learn dependencies and patterns across the text.

4. Autoregressive Generation
Text is generated one token at a time. For each step, the model predicts the most likely next token based on the previous ones. This continues until a predefined stop condition is met, such as a special end-of-sequence token or a maximum token limit.