What "reasoning" means

Placeholder — full lesson coming.

For an LLM, "reasoning" means spending test-time compute on intermediate steps before committing to an answer — generating a chain of thought, exploring, self-correcting — rather than emitting the answer in one shot. To cover: chain-of-thought, the "reasoning model" family (o1, DeepSeek-R1) and how they're trained (RL on verifiable rewards), test-time-compute scaling, and how it differs from plain sampling. Relates to thinking modes.