Well, I don't think it's a completely binary money or UX type decision, but I do know that working at a larger scale of users shifts your priorities significant

Well, I don't think it's a completely binary money or UX type decision, but I do know that working at a larger scale of users shifts your priorities significantly. For example, I worked on a website that would sometimes process as much as a million dollars in donations in a single day for a nonprofit. If the site went down for even a half hour, that could be tens of thousands of dollars in missed donations. That is like someone not getting paid for almost a year. We had a cool autocomplete feature that went out and due to a bug that made it past QA caused a on of DB queries and took down the server. It was maybe a better user experience, but the way it was implemented would cost too much in reliability for it to be worth it. The team reengineered the concept until it worked at scale. A lot of hackers I know who aren't working on that scale don't worry about QA or bugs so much because hardly anybody is using their software and it's not mission critical. When you have millions of people touching your software and it is processing millions of dollars in transactions, the metrics that matter are different. I am sure Facebook cares about UX, but they are a publicly traded company and if a slightly better UX delivers a significantly worse ad revenue number, that isn't just sort of bad for investors, it's bad for the employees job security (and stock value), it's bad for the advertisers who can't get the same ROI on their ad spend, and ultimately it's a little bit bad for the end users because if "good UX" doesn't make money, then over the long haul, Facebook will learn to not invest in "good UX". I think the scale that Facebook's engineers are working at is a lot bigger than most of us can easily appreciate.