In the agent loop, what counts as an 'action' and how is it different from a thought?

**An action** in an agent loop is a step that does something in the world: calling a tool, querying a database, running code, sending an email, clicking a button. **A thought** is the model's internal reasoning, the text it writes to plan or explain its next move. Both come out of the same LLM call, but only actions cause real effects. The distinction comes from the ReAct pattern, which made the difference explicit. In ReAct, the model alternates between two kinds of output. First it writes a thought, like "I need to find out the user's recent orders before answering." Then it emits an action, like a tool call to the order database. The runtime intercepts the action, executes it, and feeds the result back as an observation. The cycle repeats: thought, action, observation. The model's reasoning text is useful for debugging and for guiding its own next decision, but it does not affect the outside world. The action does. With the structure explicit, the runtime knows which parts to execute and which parts to log. In modern function-calling APIs, the action is the structured tool-use block in the response. Everything else (reasoning, partial answers, planning) counts as a thought.

**An action** is a step that affects the world or fetches new information. Examples include calling a tool, querying a database, running code, sending a message, or executing a shell command. **A thought** is internal reasoning the model writes for itself before deciding what action to take. The distinction comes from the ReAct (Reason + Act) pattern, introduced in 2022. Earlier agent designs interleaved reasoning and tool use implicitly. ReAct made the structure explicit. The model alternates: thought, action, observation, and repeat. The thought is plain text. The action is a structured tool call the runtime can parse and execute. The observation is whatever the tool returned. This explicit interleaving made trajectories debuggable and gave the runtime a clean place to intercept and execute actions. Modern function-calling APIs encode the structural distinction directly. Anthropic returns content blocks of type `text` (thoughts) and `tool_use` (actions). OpenAI returns assistant messages with optional `tool_calls`. Both providers treat the tool-call block as the action. The surrounding prose is reasoning the runtime forwards to the conversation but does not execute. The model decides what to emit on each turn: an action means "runtime, please do this"; no action means "I am done" or "I am still thinking." The split matters in practice. Actions have side effects, latency, and cost. Thoughts are free in comparison. Actions need permission checks, idempotency keys, and audit logging. Thoughts only need to be logged for trace inspection. Confusing the two is the cause of most beginner mistakes in agent design.