Archive
Natural Language Processing, or NLP, is the field focused on enabling computers to analyse, understand, process and generate human language. Its applications include translation, summarisation, sentiment analysis, speech recognition and conversation.
A simple explanation of embeddings, how they represent text and images as numerical vectors, and their uses in semantic search, RAG, recommendations, and similarity.
Ask someone to write a short message and you get something generic. Add who it is for and why, and the result changes completely. Same with an AI model.
A simple explanation of tokens, how AI models divide text into smaller units, and how token counts affect context limits and API costs.
Someone telling you a long story starts forgetting the early details. Models have the same limit, and knowing where it sits changes how you work with them.
A friend texts you: can you fix it? Fix what? The sentence alone is not enough. AI has the same problem, and it explains most bad answers you get.
You hear half a sentence and your mind fills in the rest. A language model does something close to that, and it explains both its power and its limits.
One sentence can mean the opposite of itself depending on your tone. We catch it without thinking. Teaching machines to do it took decades, and Darija is still losing.
I put the book on the table because it was heavy. You knew instantly what it meant. The architecture behind ChatGPT does the same thing, and that is what changed everything.
What is an LLM and how does it generate text? A simple guide to large language models, how they work, their relationship with ChatGPT and generative AI, and why they sometimes produce incorrect answers.
