I was a TA in the first OMSCS cohort. I think the biggest difference between the OMSCS experience and what other people are giving here as anecdotes is Georgia Tech was prepared for this. They created an online program on purpose that was well thought-out and planned in advance. Many universities that did this in the past year, in contrast, were forced into it by the pandemic and scrambling to do whatever they could with no real plan, bad or no policies in place, professors and TAs that were not properly prepared.
Though I actually agree here with the people saying this type of software is usually a rootkit by design and unacceptable for that reason alone. I understand why the school did it, but in my experience, we never caught a person cheating on an exam. Where we caught cheating was in project submissions, and that was often quite blatant with people turning in identical code or code identical to code on the Internet, sometimes even identical comments. No need for malware on student's personal home networks to catch that.
There is somewhat of a fundamental conflict in the structure of the OMSCS courses, though. Personally, I don't think there is much value in even having the exams at all on top of the usually extremely comprehensive practical project assignments, for the terminal masters students. But that's everyone in OMSCS. The exams tend to test your knowledge of prior research, which is mostly valuable for people going on to a PhD afterward.
Of course, some students do that, even if you need to separately apply after completing the terminal masters and can't just move straight into it like the on-campus grad students. So I don't really know the answer, but the structure right now is imposing a heavily research-focused curriculum on students who are overwhelmingly not going into research. And the OMSCS program largely exists because of sponsorship from industry that wanted stronger industry-focused CS education for their own workers from a school more reputable than a plain MOOC with no university behind it.
They do an amazingly good job of best of both worlds right now, but it is a conflict as I think many of the exams are unnecessarily difficult for people who are never going to become researchers. I guess it's the classic Google approach to hiring, though. You end up filtering a bunch of students who may have been perfectly good as programmers, but you're probably never going to let anyone through who wasn't, and that is ultimately more important for maintaining the school's strong reputation in industry.