Mechanism Design: Constructing Algorithms for Strategic Agents — LessWrong