I have many times reported bugs to Apple, been told that what I'm describing is not a known issue and must be something specific to my setup, corrupted on my machine/device, etc, only to have the advisor with whom I'm speaking finally follow the steps I've outlined, and discover that they can reliably reproduce the bug on their own machines/devices. The notion that millions of users would be reporting something if it were actually a widespread bug is just wrong. Most users do not use their computers, tablets, or phones deeply enough to knowingly experience intermittent or subtle bugs, and neither do they have the knowledge, experience, or willingness required to properly communicate the issues they are facing.
"I've never encountered this problem; if it was actually a problem with Sierra itself, wouldn't everyone be encountering it?"
Not necessarily. The issue might be specific to those with certain hardware, specific settings, or a multitude of other variables. I'm not going to argue that Apple should always catch every bug before releasing an OS update, but I do think that they should work harder and faster on solving the problems that are noticed by their users, and I find it highly frustrating that they don't answer questions and deal with issues which are reported and discussed in their own forums here.
Safe Mode can indeed be a great diagnostic tool, but results of using it can also be quite misleading. Let's say I'm suffering from a bug which is consistently cured temporarily by logging out of my account or rebooting. If I start in safe mode, and I don't immediately encounter the issue, how long do you think I'm going to keep using my computer in safe mode before I think that the problem doesn't exist there? What if it takes three days of use for a problem to present itself? How many forum threads have I read here where some "helpful" person suggests rebooting, several readers try that, report back that their issue was solved, credit the helpful person with solving their issues, then report back a day later that the problem has recurred? At some point, these helpful people who suggest the same things in thread after thread after thread need to stop presenting their advice as likely solutions, and attach to them the requisite disclaimers: "This may only help temporarily, so wait for hours/days before concluding that your issue is resolved" etc.