Microsoft famously canned their entire group of people who do testing (SDETs). If the problem you hit isn't caught by an automated test at build, Microsoft employees running it in some form of "dogfood," or the people running beta builds, it is getting shipped.
Maybe that's good enough. Maybe people are getting used to software coming out with multiple updates per quarter to fix some "random glitch."
I look at it as the computing industry getting away from the roles and pace that made computers reliable instead of just functional. We used to have Operations staff to keep a service reliable; we used to have Testing staff to make sure a code base was reliable. They had their own failings, but to cut them out entirely did wonders for stock prices and not much for the end user experience.
But maybe being able to launch a pre-VC startup with one person and a stack of cloud services is a better trade-off.
Maybe that's good enough. Maybe people are getting used to software coming out with multiple updates per quarter to fix some "random glitch."
I look at it as the computing industry getting away from the roles and pace that made computers reliable instead of just functional. We used to have Operations staff to keep a service reliable; we used to have Testing staff to make sure a code base was reliable. They had their own failings, but to cut them out entirely did wonders for stock prices and not much for the end user experience.
But maybe being able to launch a pre-VC startup with one person and a stack of cloud services is a better trade-off.