top of page

How Generative AI Works: A One Minute Guide

Ever thought how generative AI is able to mimic human capabilities? 


If not, then allow me to explain. In an oversimplified way.


So, suppose you want to create an AI that creates internet content about philosophy. You want your AI to be so good, that it can spit out information on any philosophical topic - be it the ancient teachings of Plato or the modern ideas of Sartre.


For this, you have to create three components:


1. Dataset

2. Interpreter

3. Predictor


The dataset is the information you feed to the AI. In this case, it would be all the philosophy books, internet articles, lecture notes, etc., you can get your hands on.


The interpreter identifies patterns in the information. Such as identifying that Marcus Aurelius, Seneca, and Epictetus all belonged to a branch of philosophy called Stoicism.


The predictor matches your users' prompts with the identified patterns. For example, if your user requests a poem about stoicism, it takes the teachings of Aurelius, Seneca, and Epictetus to craft the said poem.

Commentaires


bottom of page