I like that errors can be quantified as it removes some doom ambiguity from mistakes. The first company I worked for sold access to data on a per minute basis

I like that errors can be quantified as it removes some doom ambiguity from mistakes. The first company I worked for sold access to data on a per minute basis and it was like $1.50/minute or something (dialup access in the 90s to an interactive ClarionDB session). So any outage would be quantified by some curious programmer. I remember even small outages being quantified in hot washes as 10 minutes x 2000 users x $1.5 = $30k outage. I remember one of my bosses was nicknamed “Commander Crash” as he once crashed the whole system for hours based on a single live config file edit. No one was fired but it was actually liberating knowing how much mistakes cost.