I think the author makes a point but also misses (or voluntarily glosses over) important counter points.<p>1) Machine learning (which includes deep learning) still hasn't been applied to a lot of field that would draw benefits even from a basic CNN.<p>2) While AI/Smart branded crap is indeed all the rage right now (and part of the bubble imo), most of those are not learning algorithms and rely on cheap marketing tricks (looking at you Oracle).<p>Bottomline: lot of hype comes with empty promises and ultimately the people's dissatisfaction, but there is a real demand and need for applied machine learning in many fields.