alt Hacker News"Does the AI system perform actions under its own identity?"
I don't agree with this definition.
I view an agent has having the ability to affect the world, and then sense how it affected the world and then choose to make additional actions. Thus there is an act, sense, re-act feedback loop going on that does not require a human to mediate this. This to me is an agent.
"But why isn't, say, ChatGPT an agent?"
ChatGPT (the web app where you send it chats and it responses) by default doesn't act on the world and sense the changes it is making. Although once you take the GPT o4 model and hook it up with tool calling that affects the world in a feedback loop is definitely an agent.
I believe this definition generally aligns with most people's definitions as well.
I wrote an essay about building an agentic coder and it really is when you establish the tool-calling feedback loop that things move from an assistant to an agent: https://benhouston3d.com/blog/building-an-agentic-code-from-...
Replies
Your right, the “own identity” part is the problem. You can act on your own agency or you act as an agent for someone else.
AI today is only the second. We tell it what we want, it acts by our impetus, but what it does or how it does it, is up to it.
Is ChatGPT with its Code Interpreter tool an agent?
[dead]
I agree with you. People really overcomplicate this.
From wiktionary:
""" Agent (plural agents)
- One who exerts power, or has the power to act.
- One who acts for, or in the place of, another (the principal), by that person's authority; someone entrusted to act on behalf of or in behalf of another, such as to transact business for them.
- [various more specific definitions for real estate, biology, etc]
From Latin agēns, present active participle of agere (“to drive, lead, conduct, manage, perform, do”). """
An agent is simply someone or something that does something, usually for someone else. An AI agent is thus an AI that does something, usually for someone else. An AI assistant could be an AI agent, or it could be a glorified chatbot who merely offers you spoken or written word, possibly after reacting to real-world information (but not itself modifying it)