Post #4235102

@teledyn@mstdn.ca I think it's best summed up as a change from thinking about software in terms of mean time between failure, to thinking about it in terms of mean time to recovery. And because compute is so cheap now that nobody much cares about crufty performance, either.