Two years ago for Christmas, I got an ‘impossible to solve’ Magic Hexagon puzzle. Using the power of evolutionary genetic algorithms, numpy and millions of iterations, ‘impossible’ turned to less than a second of computation. The key ideas are simple: 1. Represent each state of the board as a gene. 2. Create a population of boards. 3. Randomly mutate each gene in the population. 4. Based on a fitness measure, select the best candiates from the mutated population and 5. clone the best candidates to create a new population and repeat until convergence. Through this cycle of evolution an apex gene emerges, and the unsolvable puzzle is solved.