My reaction to the article title was recalling how old assembly teachers would drill into our heads to make sure we have proof of 200% gains before implementing an optimization. However the article shows how sometimes such planning doesn't work.