Source URL: https://simonwillison.net/2025/Oct/6/work-independently/
Source: Simon Willison’s Weblog
Title: a system that can do work independently on behalf of the user
Feedly Summary: I’ve settled on agents as meaning “LLMs calling tools in a loop to achieve a goal" but OpenAI continue to muddy the waters with much more vague definitions. Swyx spotted this one in the press pack OpenAI sent out for their DevDay announcements today:
How does OpenAl define an "agent"? An Al agent is a system that can do work independently on behalf of the user.
Adding this one to my collection.
Tags: ai-agents, openai, agent-definitions, swyx
AI Summary and Description: Yes
Summary: The text discusses definitions of “agents” within the context of Artificial Intelligence (AI), particularly focusing on how OpenAI defines an AI agent as a system that performs tasks independently for users. This redefinition reflects ongoing discussions and potential misunderstandings in the AI community, making it relevant for professionals in AI and related fields.
Detailed Description:
The text emphasizes the evolving understanding of AI agents, particularly in relation to large language models (LLMs) and their capabilities. The definition provided by OpenAI contrasts with other interpretations, indicating the potential confusion in the industry regarding what constitutes an AI agent.
– **Definition Context**: It mentions that agents are defined as “LLMs calling tools in a loop to achieve a goal”, suggesting a practical perspective on AI functionality.
– **Vagueness in Definitions**: The concern about OpenAI’s vague definitions highlights the need for clarity in the rapidly developing field of AI, particularly for professionals who must ensure precise communication and understanding of AI’s capabilities and limitations.
– **Industry Implications**: The ambiguity around the term “agent” could lead to misinterpretations in the development, deployment, and governance of AI systems.
– **Relevance for Professionals**: This definition issues a call to AI practitioners to engage critically with evolving terminology and to demand clearer frameworks as they navigate compliance, security, and operational challenges posed by AI solutions.
Overall, this text is important for those in AI security, cloud, and infrastructure fields, urging them to consider the implications of terminology and definitions when designing or implementing AI systems.