This talk sketches some connections between approximability of NP-complete problems, analysis and geometry, and the role played by the Unique Games Conjecture in facilitating these connections.