And to remind you from a job perspective, the people using GPU clusters are doing the “best work” there is.Thought Provoking?The use cases of Data Science which were envisioned in the starting phases have not panned out pretty well..Our own god, Nate Silver, has faltered while making election predictions..The use cases for image classification and detection which we have solved in the past few years using Deep Learning, though great research problems, leave a lot to be desired.Businesses not really ready to accept change in the ways of working..Finance domain comes as the foremost example to mind..With all the regulations and corporate clutter, I noticed that a lot of work is needed to even deploy anything other than a simple logistic regression to production..When I look at the needs of the Business partners or the CEO’s, they really don’t want models but “Advanced Visualizations”..A glorified term for basic analytics..A dashboard is pretty much every project ever..While from a data scientist’s perspective I should be happy if I could just provide good business results using just basic visualizations and patterns, it takes away the sexiness that we as practitioners were promised.All in all, we are stuck in the name of the new rebranding doing the exact same work we were doing before..We need to break this chain and strive for excellence.So What Next?Data Scientists need to be at the decision-making table..This will help in envisioning better use cases as the problems become more clear..Also, any new project that we envision comes with its own problems around data.. More details