Well, Claude Shannon's 1937 master's thesis in EE pretty much invented digital circuit design with boolean algebra ... but I wouldn't say everyone who came after him was lowering the bar :)
There were less students at his time. If there was a way to plot "average added value" of a master's theses trough out the last century, I'd expect it to be non zero historically and then to exponentially converge to zero.
Let's attempt to keep two conflicting thoughts in our heads at the same time:
1) Linus is fantastic. We owe him so much! He has changed computing in a very forceful and persistent way - people will rightfully remember him long after we're all gone - along with dmr and ken, of course.
2) Stockholm University didn't even have a computer science education at the time (I think) (in Stockholm, this was only offered by KTH, The Royal Institute of Technology). Either way, KTH is far more prestigious in this area.
In my opinion: this particular honorific event was just a result of trying to get a famous name, sort of how like the frankly speaking pathetic Nobel Peace Prize committee in Oslo keeps trying to meet famous people. (Like Obama, before he had even entered office.) It's embarassing to everyone involved.
Depends on a country and educational system, right? Yet usually there is someone or some committee which decides whether a piece of work is sufficient for a student to get a master's degree. In other words, they set a bar.
The bar can vastly differ among universities and generations. That's why I asked in the first place.
A masters theses is basically just a project and a report. Usually little to no original research. It's more about practicing writing reports than anything else.
Depends. A PhD thesis with negative results (one that does not disprove the null hypothesis) is still original research and still counts for the poor chap that made it for the degree.
You cannot know a priori that you will have positive results that will advance the field.
https://en.wikipedia.org/wiki/Claude_Shannon#Logic_circuits