The problem with the Bitter Lesson is that it doesn't clearly define what is a computational "hack" and what is a genuine architecturally breakthrough. We would be no where without transformers for example.