Sure, you can build highly reliable supercomputers. But the majority of us are running on google style networks of mostly-reliable boxes. I worked for a compa

Sure, you can build highly reliable supercomputers. But the majority of us are running on google style networks of mostly-reliable boxes. I worked for a company that built distributed ML software, and even at eg hundreds of boxes you will regularly see failures. Customers will be very unhappy unless your code tolerates that.