My understanding is that it isn't but recent-ish statistical approaches have had lots of success. I know that almost overnight Arabic went from an unreadable mess to about as good as a basic non-native speaker.

I wish it'd find its way into usage for East Asian languages more, but I suppose finding dual-equivalent corpuses to build the models off of is hard.

https://en.wikipedia.org/wiki/Statistical_machine_translatio...

Interesting, thanks!

I had to struggle to read it. I mean:

After logout, one of the key interlocutors for chat will regenerate, and to check that I have the same key as the source, I can only look in his eyes phone.

...did it translate "iPhone" to "eyes phone"? I'm not sure. If that isn't what happened, then something far more horrible must have.

The original:

После логаута одного из собеседников ключ для чата будет перегенерирован, а проверить то, что я имею тот же ключ что и собеседник я могу только посмотрев в его телефон глазами.

My (human) translation:

After one of the participants will log out, the key from the chat will be re-generated, but in order to check that I have the same key as them, I would need to see their phone with my own eyes.

This sentence has a particularly non-English word order, plus some missed punctuation. I can see how it would be a hard case for machine translation.

The thing is, I though it was poorly written English, written by a non native speaker, rather than a computer.

I must also clarify that I was a little distracted.

I noticed it instantly (the title). I also noticed many irregularities throughout the article. I don't know if you're a native speaker, but to me it was obviously computer translated.

English to Russian is a bit easier than Russian to English. Word order doesn't mean as much in Russian as in English, so the translation is lossy.

Does this seem good to you? Damn, I feel I should stop worrying about my English.