Greg discusses the contrast between general systems and specific task optimization, highlighting how humans, much like machines, can achieve greater efficiency through practice and mastery. He draws parallels between computational learning and human reasoning, emphasizing the importance of muscle memory in enhancing performance.