Language translation continues to be one of the most common AI-moderated tasks performed in this multicultural world, especially with the rising popularity of large language models such as GPT-4. In this paper, we revisit a human-centered approach to language translation: post-editing. We compare language translations between English and Mandarin – two of the most-spoken languages globally – as performed by the Neural Machine Translator (NMT) employed by Google Translate to one of the state-of-the-art post-editing services in Translate.com. Through a mixed-methods approach combining qualitative analysis of results by 13 bilingual interviewees and quantitative analysis performed by 5 large-language models, we make a case for post-editing and the human element in language translation continuing to stay relevant in this ML-driven age. Though we recognize the cost differential is a significant pain point for users, we demonstrate post-editing producing translations of higher quality over NMT results in almost all cases, especially in conversations containing colloquialisms.