Your AI Agent Is Not a Chatbot. It Should Be a Class.

A new paper from NVIDIA Labs called NOOA (Native Python Object-Oriented Agents) argues that AI agents should be built as Python classes rather than as complex prompt-based systems. The framework treats agents as objects with methods and fields, aiming to simplify development and avoid the fragmentation common in current agent-building approaches. The paper backs its claim with benchmark results, suggesting a shift toward treating agent engineering as standard software engineering.

Member-only story Your AI Agent Is Not a Chatbot. It Should Be a Class. Why NVIDIA’s new agent framework treats “prompt engineering” as just… software engineering For Non members , please check the Responses of this story Every team building AI agents right now runs into the same wall. You start with a simple prompt. Then you add a tool. Then another tool. Then a memory file. Then a workflow graph to control the order of steps. Before long, your “agent” is spread across five different files, written in three different formats, and nobody on the team can explain the whole system in one sitting. A new paper from NVIDIA Labs called NOOA Native Python Object-Oriented Agents, pronounced “double-O Agents” argues that this mess is unnecessary. Their pitch is simple: an AI agent should just be a Python object. Not a special new language. Not a graph of boxes and arrows. A class, with methods and fields, the same way software has been built for decades. This idea sounds almost too plain to be a research contribution. But the paper backs it up with real benchmark numbers, and the reasoning behind it says something important about where AI agents are heading.