There's significant disagreement on what constitutes an AI agent, with varying definitions emerging from both technical and marketing perspectives.

The concept of agents includes both those that replace human tasks and those that operate as low-level system processes, complicating the definition further.

The definition of an agent may evolve as we explore the capabilities of AI, particularly in how they interact with users and perform tasks.

As AI models improve, agents are becoming more prevalent and capable, prompting the need for formal definitions to clarify their use.

An agent is something a human can delegate tasks to, and we're aiming for agents to handle more complex tasks over time.

The confusion around agents stems from the expectation that AI will develop human-like intent and creativity, which may not be theoretically possible.

The distinction between agents and functions is blurred, as both can involve multiple functions with language models in the middle, making them appear similar from an external perspective.

AI agents are currently overhyped, especially in consumer applications.

Agents for consumers are currently overhyped. Specifying preferences for tasks can be as complex as doing the task yourself, and verifying the agent's actions can be very expensive.