I appreciate what you're saying, but I don't think it quite applied. What I meant was that it's easy to create an architecture for an application that doesn't scale well at all. Eg - poorly sharded data, lots of cross dependencies etc. However, if you properly think through your data model and data flows and use cases, it's generally possible to create a system that is extremely scalable in all directions. This is certainly not easy, but it's a hell of a lot easier than creating some huge ai driven data slurping ad empire.