[thrdbndndn]

As I mentioned in another reply, the scenario here is translating "artistic" or "real-world" (for lack of a better term) literature accurately—whether it's a novel, a YouTuber's video, casual conversation, or blog posts/tweets with internet slang and abbreviations. In these cases, getting things 95% right isn’t enough to capture the nuances, especially when the author didn’t create the content with translation in mind (which I believe matches your experience).

Machine translation for instructional or work-related texts has been "usable" for years, way before LLM emerged.

LLM-based translation has certainly made significant progress in these scenarios—GPT-4, for example, is fully capable IMHO. However, it's still not quite fast enough for real-time use, and the smaller models that can run offline still don't deliver the needed quality.