Unverified50% confidenceFactExact time
A large model (LLM) itself only outputs text and cannot directly operate a computer; an agent (Agent) is a system framework built on top of a large model, containing modules such as tool calling, memory management, task planning, and execution environment.
1
Sources
50%
Confidence
Long-term
Relevance
10/29/2026
First Seen
Sources
WorkBuddy in Action: How a Desktop Agent Automates Your Work
bilibiliWorkbuddy系统课10/29/2026
Related Entities
Related Claims
Unverified智能体的许多能力实际上是由Agent所运行的框架在LLM外部实现的，而非LLM本身77% similarUnverifiedThe technology of having large models control a computer to perform operations is commonly referred to as Computer Use or Agent mode, requiring the model to generate and execute system calls, which places higher demands on instruction-following capability than pure text generation.76% similarUnverifiedIn LLM-driven agents, the model acts as the "brain," responsible for understanding user intent, planning steps, and calling tools.74% similarUnverifiedThe hierarchical design (LLM for planning, lightweight models for execution) is becoming the mainstream paradigm in scenarios such as game AI, robot control and automated workflows.74% similarUnverifiedThe Agents in this framework are not special programs, but rather LLM invocation instances carrying dedicated system prompts and independent context windows.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1113004API
curl https://kongchang.com/api/v1/knowledge/claims/1113004MCP
get_claim(id=1113004)