The advantages of ML are said to be: - Highly parallelizable (depends on μ) But in reality in depends on λ because I just have to evaluate the offspring, right?