Well, previously you were a 'researcher' working on 'open problems'. Now the problem has been solved, time to become an 'engineer' working on implementing/optim

Well, previously you were a 'researcher' working on 'open problems'. Now the problem has been solved, time to become an 'engineer' working on implementing/optimising the solutions. The usage of AI will spike 100x going forward, from a mere curiosity inside the company which only affects management decisions (Say sentiment analysis), to extremely intensive direct interactions with customers. This will drastically increase the need for reliability and tractability of the results. NLP researchers simply have to accept 50% of what they did is useless, and reorient around the other 50%, its the risk of researching in an open field.