My experience: I was interested in machine learning and was working in industry with no CS background. I was accepted to an MS in CS based on my undergraduate work and also programming experience. The program was in-person, not online. Through the courses and contact with professors both in CS and in other departments (Mathematics, Statistics) I developed a far better understanding of the field than I would have without being at a university.

All of the courses were the same regardless of whether or not you were an MS student or a PhD student. Professors actively encouraged MS students to continue on to the PhD.

The article says "If your undergrad degree was in some other field, you can get through an MS in CS without ever taking an algorithms or data structures class." In my department you had to take a minimum number of courses from a few categories. For the category that included the Algorithms course, I would say at least 90% of all students took that course rather than the others on offer. It followed CLRS and moved really quickly for someone with no undergrad background like me. The course had no programming in it so I have no idea what the author is talking about when per mentions programming experience.

I think if the program were online it would have been harder to have the multidisciplinary experience I got in-person. But for people who didn't want that aspect of it, I think an online program might work just fine.