I am certainly not saying everything you say is not all true. My comment is dark humour. I really like your last point. Years ago I replaced a huge hadoop clust

I am certainly not saying everything you say is not all true. My comment is dark humour. I really like your last point. Years ago I replaced a huge hadoop cluster data processing job with a single app on one machine with a few CPUs, that reduced a job that took over 8 hours to 20 minutes. What is even dumber is, it was just a python script and gnu parallel, which used to be perl.