Vision-language-action models connect cameras, human instructions, and robot movement. See how they work, where they fail, and why safety still matters. ⚙️🗣️🤖