context tax
A concept in AI product interaction design referring to the hidden efficiency loss caused when users must repeatedly supply information that the system should already remember or prefill in order for the AI to understand a specific task. It is often used to distinguish demo-friendly interfaces from production-ready interfaces.
Timeline (last 90 days)
Every piece of text in a context file is injected into the model's context window on every call, causing increased token cost, i.e., the "context tax".
The context tax refers to users repeatedly providing information that the system should have remembered or pre-filled in order for the AI to understand a specific task.
In AI agent engineering, each tool mounted occupies the model's context window through the tool's name, description, and parameter schema, and this overhead is often called the 'context tax'.
Traditional AI chat windows use a stateless conversation mode where every new conversation starts from scratch, requiring users to manually paste background materials — what the industry calls the "context tax".
All Facts (4)
Every piece of text in a context file is injected into the model's context window on every call, causing increased token cost, i.e., the "context tax".
50%UnverifiedThe context tax refers to users repeatedly providing information that the system should have remembered or pre-filled in order for the AI to understand a specific task.
50%UnverifiedIn AI agent engineering, each tool mounted occupies the model's context window through the tool's name, description, and parameter schema, and this overhead is often called the 'context tax'.
50%UnverifiedTraditional AI chat windows use a stateless conversation mode where every new conversation starts from scratch, requiring users to manually paste background materials — what the industry calls the "context tax".
50%