Anonymous asked:

Do you think deepl is worth using for a decent enough translation or its not worth using at all?

That depends on how accurate you want the translation to be. It's definitely not ATLAS-tier, but you're still only going to get a rough translation out of it. In my experience, the chief problem with DeepL is that it simply makes stuff up. You'll get anything from basic paraphrasing, to invented details, to sentences that are completely wrong in every way. But no matter which of these it gives you, it will always output a sentence that sounds convincingly accurate. It's good by machine translation standards, but be aware that it still has the same exact pitfalls—it's just better at hiding them.