Artificial General Intelligence (AGI) is the concept of a machine that can understand, learn, and apply knowledge in a way that matches or surpasses human intelligence.

Unlike tools designed for specific tasks, AGI can theoretically handle any task a human can—and much better.

An Application Programming Interface (API) allows different software applications to communicate with each other.

In AI, an API can let your application use AI capabilities provided by another service, like generating text or images, without needing to develop the AI technology yourself.

An agent is a piece of software that makes decisions or performs actions on behalf of a user or another program.

Think of it as a digital assistant that can learn and act based on the information it's given.

An algorithm is a set of rules or instructions designed to solve problems or perform tasks.

In generative AI, algorithms process data and patterns to create new content, like writing a story or generating an image.

Automation involves using technology to perform tasks without human intervention.

In AI, this can mean automatically creating content, making decisions, or analyzing data.

Bias in generative AI refers to unfair prejudices or favoritisms in the AI's output, often reflecting the biases in the data it was trained on. This can result in unfair or discriminatory results.

Working to identify and reduce bias is a critical part of developing ethical AI systems.

A "corpus" is a large collection of text data used to train generative AI models.

It can include books, articles, websites, or any other text. The quality and diversity of the corpus can significantly affect the AI's performance and ability to generate relevant and varied content.

Generative Pre-trained Transformer (GPT) is a type of AI model known for its ability to generate human-like text.

It learns from a vast amount of text data and can then create new text that sounds like it was written by a person, based on a given prompt.

Generative AI is a type of AI that can create new content, such as text, images, or music, that resembles human-created content.

It learns from existing data and applies what it has learned to produce original outputs.

Large Language Models (LLM'er) er avancerede AI-systemer, der er trænet på store datasæt af tekst.

De kan forstå og generere sprog på en måde, der efterligner menneskelig skrift.

ChatGPT, Gemini og Claude er alle eksempler på en LLM.

In AI, a model is a system that has been trained on data to recognize patterns and make predictions or generate content.

It's like teaching a computer to understand certain information and then use that understanding to do tasks, such as answering questions or creating new text.

OpenAI is an organization that conducts research in AI and develops advanced AI technologies like ChatGPT.

Their main objective is to create safe AGI that benefits all of humanity.

Parameters in AI are the settings or variables that the model uses to make decisions and generate outputs.

You can think of them as the model's internal knowledge, shaped by training, which it applies when performing tasks.

Prompt engineering is the art of crafting questions or prompts that guide generative AI to produce the desired output.

It involves understanding how to effectively communicate with the AI to get the best results, like asking the right questions to generate specific or creative responses.

"Shot Prompting" term refers to how many examples you give an AI model when asking it to do something.

For instance, "zero-shot" means giving it no examples, "one-shot" means giving it one example, and "few-shot" means giving it a few examples.

It's a way to guide the AI on what you're asking it to do.

A token can be a word, part of a word, or just characters like punctuation.

AI models like GPT break down text into tokens to understand and generate language.

It's the basic unit of text that the AI works with.