Given design, setup/prep, printing/molding, and trim work, that's still quite impressive. Mass producing one thing over and over is easy. Changing your tooling to deal with a new part is what's hard. When I worked in factories, we'd get laid off for a week when it was time to switch products. The engineers needed time to swap everything out. It was equivalent to rearranging a huge house where all the furniture weight over 30tons. I'd imagine these places are setup for lots of rapid changes so it wouldn't be so bad, but it's still probably requires a lot of work. Also, I doubt the workers are your regular linemen. They'd almost have to all be engineers.
When I wrote a bunch of software for InvisAlign over 10 years ago, we were ramping up to a capacity of 20,000 unique plastic parts per day while printing over half of that every day. I can only imagine what they're doing today. The actual printing was mostly stereolithography making molds, pressure forming, then CNC cutting them off, but there was also scanning, modeling, approvals, labeling, mesh cleanup, supports, etc., which all had to happen in 3d. The automation required to get all that humming along was substantial (lots of patents, and not just "on the internet" ones...)
