IMHO, Microsoft could be seeing a similar problem as IBM 10-15 years ago. They couldn't hire and keep the Great Ones any more. First it meant getting out of the OS business. Then it meant buying software firms rather than developing software. They realized that they were an Enterprise Sales and Services organization.
The big difference here is that Microsoft owns a large segment of consumer OS, something that IBM did not (was stolen by Microsoft) in the past.
While I've heard or read prediction/assumption/guesses/perspective of MS smells like IBM 10-15 years ago (and they are legit, I'm not saying it has no merit), that difference alone is a huge (and fundamental) one.
Speaking of "great ones", there are plenty of great ones that work at MS, just like there are plenty that work for Google. It's just that the "great ones" solve different problems in different domain: most people only heard about smart people working at Google solving problems at scale, while the .NET team (Visual Studio, Compiler, IL, Libraries, etc) are not necessary "trendy" any more.
Judging MS engineers as "average" engineers are simply naive or too cocky without proper education at best.
