What is AI Context

Preface

With the emergence of ChatGPT in December 2022, AI has rapidly become the focal point of global scientific and technological development and social tool innovation (primarily referring to LLM large language models). Expanding the boundaries of AI capabilities, exploring AI applications, and other AI-related fields have become one of the hottest and most valuable investments worldwide.

Overview

AI Context refers to a series of resources provided to AI before it actually begins processing tasks, including but not limited to prompts, text knowledge bases, specific tool calls, etc. The purpose is to precisely define AI's task processing behavior, making the results more aligned with expectations.

Resources

Currently, from which perspectives can we approach building an AI Context? The following are resources I understand as belonging to AI Context:

• Prompts: Prompts are the most fundamental text-based language forms that directly communicate with LLM large models, serving as "scenario constraints" or "instructions" that can very directly influence AI behavior
• Knowledge Bases: Knowledge bases refer to collections of data containing professional information in specific domains, commonly used to provide reference information for AI task processing to standardize and precisely define AI's processing behavior
• MCP Servers: MCP (Model Context Protocol) is an open protocol that has gradually become standardized for how LLM large models connect to external tools, under the premise that LLM large models widely support Function Calling to invoke external tools. An MCP Server can provide AI with a bundle of specific AI Context, including tool calls, text knowledge queries, and prompt templates, thereby expanding AI capabilities, enhancing AI's domain expertise, and maximizing consistency in homogeneous workflow results.

Summary

LLM large model training is a game for world-class consortiums, while AI Context construction is an individual's ode to AI.