Under the Lab-in-the-Loop model, the difference between completing one iteration in a month versus one iteration in a week is an order of magnitude in model performance.