# Introduction to Generative AI
## What is Generative AI?
Generative AI, or GenAI, is a type of AI that creates new content. It can write text, generate images, produce audio, and even make videos.
It learns patterns from large amounts of data. Then it uses those patterns to create something new that feels natural and useful.
Tools like chat assistants, image generators, and AI copilots are all examples of GenAI.
***
## What are the use cases of GenAI?
GenAI is useful in many everyday tasks and business workflows. Some common use cases are:
* **Writing and communication**: Draft emails, summarize documents, translate text, and create reports or marketing content.
* **Chat and support**: Power chatbots, virtual assistants, and copilots that answer questions in natural language.
* **Creative work**: Generate or edit images, audio, video, and design ideas from a simple prompt.
It can also support research, learning, product design, and data analysis.
***
## What is an LLM?
Large Language Models, or LLMs, are AI models built to understand and generate human language. They can answer questions, write content, summarize information, and help with many text-based tasks.
They learn from very large amounts of text. When you give them a prompt, they predict what should come next. That is how they generate useful and natural responses.
They are called "large" because they are trained with a huge number of internal settings, often called parameters. These settings help the model understand patterns, meaning, and context.
Some examples of LLMs are:
* **GPT**: A family of models from OpenAI that can write, explain, summarize, and answer questions.
* **Gemini**: A family of models from Google that can work with text and, in some versions, images and other media.
* **Claude**: A family of models from Anthropic designed for helpful and thoughtful text generation.
* **Dola Seed**: A family of models from BytePlus that can help with text generation, reasoning, and other language-based tasks.
* And more...
***
## What is an Embedding?
An embedding turns content into numbers that AI systems can compare. It helps a machine understand meaning, not just exact words.
For example, two sentences with similar meaning can have similar embeddings, even if they use different wording. This is useful for search, recommendations, and question answering.
***
## What is a Vector Database?
A vector database stores embeddings so they can be searched quickly. Instead of searching by exact keyword, it searches by meaning.
This makes it useful for semantic search, recommendation systems, and AI assistants that need to find the most relevant content.
***
## What is a Multimodal Model?
A multimodal model can understand or generate more than one type of content. That can include text, images, audio, or video.
This helps the model understand context more fully. For example, it can answer questions about an image, generate captions for a video, or create an image from a text prompt.
Common examples include:
* A chatbot that reads text and understands images.
* A model that turns text into images or video.
* A tool that analyzes both sound and visuals in a video.
***
## What is the memory of an LLM?
A model's memory usually means how much earlier context it can still use in the current conversation. This is often called the context window.
For example, if you ask a follow-up question, the model can use the previous messages to understand what you mean. The more context it can keep, the better it can handle long chats, detailed documents, or multi-step tasks.
This is different from training data. Training teaches the model general knowledge. Memory helps it stay aware of the current conversation.
**GPT**: A family of models from OpenAI that can write, explain, summarize, and answer questions.
* 
**Gemini**: A family of models from Google that can work with text and, in some versions, images and other media.
* 
**Claude**: A family of models from Anthropic designed for helpful and thoughtful text generation.
* 
**Dola Seed**: A family of models from BytePlus that can help with text generation, reasoning, and other language-based tasks.
* And more...
***
## What is an Embedding?
An embedding turns content into numbers that AI systems can compare. It helps a machine understand meaning, not just exact words.
For example, two sentences with similar meaning can have similar embeddings, even if they use different wording. This is useful for search, recommendations, and question answering.
