Just like with data science, don't we expect data engineering to get to a point where they codify best practices into a tool, so 1x data engineers can be closer to be as effective as 10x data engineers? The amount of data engineering work won't change, but the number of people needed to do it will reduce, reducing the open headcounts.