When an AI confidently states something that isn't true. It might invent a fake citation, misquote a real person, or describe an event that never happened, all with the same confident tone it uses for accurate information. Hallucinations happen because LLMs generate text based on patterns, not verified facts. They don't know when they don't know. Always verify important AI outputs before relying on them.

A technique that improves AI accuracy by letting a model look up relevant information before generating a response, rather than relying solely on what it learned during training. Think of it as giving the AI access to a filing cabinet it can consult in real time. RAG is why some AI tools can answer questions about your specific company documents, recent news, or proprietary data without needing to be retrained.

A term coined in 2025 describing the practice of building software primarily through natural language prompts to AI, with the developer guiding the direction rather than writing every line of code. Instead of engineering a solution, you describe what you want and iterate. It lowers the barrier to building software dramatically, and raises genuine questions about code quality, maintainability, and what it means to be a developer.