As a student that uses Google scholar quite often "almost everyone" is a lot lower than you think it is. At least in computer science and related fields, I have to use my schools access to journals almost every time.
That's odd. I'm a computer scientist who uses Google scholar quite often, and I can't remember when I last needed to read a paper which Google couldn't find for me.
Maybe the situation is different in my fields (algorithms and cryptography) than in yours?
Maybe the situation is different in my fields (algorithms and cryptography) than in yours?