A large language model (LLM) is a machine-learning model trained on extensive text data to generate and interpret natural language. Claude uses LLMs to support tasks such as coding, documentation, and planning. section 1.0 section 1.3
An LLM generates responses by guessing what should come next, using patterns learned from large amounts of training data. Its answers can be useful and confident, but confidence is not proof of correctness. Results also depend on the training data and the decisions of the organization building the model. section 1.3
Work with an LLM through critical dialogue: give it a task, review the response, and use corrections or clarifications to steer the next step. The model cannot know your systems or start work without direction, so your experience and judgment remain essential. section 1.0 section 1.3
Parallel project work
Use separate Claude sessions to generate code, write tests, draft documentation, and plan another project. Review their outputs rather than accepting them automatically; managing several streams takes practice. section 1.3
Claude offers different models for different tasks: Opus for complex reasoning, Sonnet for daily work, and Haiku for lightweight tasks. Its tools include web chat, a desktop app, a command-line coding agent, and a developer SDK. section 1.7





