Situations like that are not why statistics is becoming the new in-demand skill. Expensive trials have existed for a long time. Petabytes of barely-structured data haven't.
I didn't comment on the article but on the statement above about applied statistics being dominated by ml and about Bayesian statistics being the future.