Implement K-means and run two iterations

Given points P={(0,0),(0,2),(2,0),(2,2),(8,8),(8,10),(10,8),(10,10)} and k=2, (1) initialize centroids with k-means++ using seed=42 and Euclidean distance; show the sampled centers. (2) Perform exactly two Lloyd iterations (assign → update) by hand, breaking ties by choosing the lower-index centroid; list cluster memberships and centroids after each iteration. (3) Compute SSE after the second iteration. (4) Describe how you would implement this efficiently in vectorized NumPy (no loops over points), including the time complexity per iteration, and explain how you would handle empty clusters robustly.