I'm curious what the author's take on scale.ai is since he was considering building an ml-ops / data labeling company and they seem to be the most successful startup in that space.