Why @Microsoft Azure ML offering is failing

We have all seen the marketing around Microsoft's Azure ML offering. For those of us that have actually tried it we know it's a train wreck: On the blog post there is a quote from a supposed data scientist: "Look at the amazing choice of datasets from many sources and so many data processing/feature engineering options. Oh, oh… and don't forget the vast array of powerful ML algorithms at my fingertips, and then web service modules so that I can work in harmony with the app developers" I'm sorry, but reading this my first reaction was: "Said no data scientist ever....". If you were what I would consider a real data scientist you would be able to pump out kickass solutions using sklearn/pandas/numpy or R faster than anything Microsoft has to offer on this WYSIWYG platform. The idea sounds cool: a Labview/Simulink type environment for building models (see below).