If You Don’t Have a Job Repairing a Robot, You Don’t Have a Job

The United States and a number of other countries have lost many millions of manufacturing jobs in the last two decades, in particular in the last decade.

The main reason why Donald Trump was elected last year may very well be the fact that he was the only candidate talking about this problem and promising to do something about it, still left standing after the Democratic Party unceremoniously got rid of Bernie Sanders by sabotaging his every move.

The truth is, the Democratic establishment prefers Donald Trump to Bernie Sanders. Bernie’s presence in the White House would remind Democrats day after day what they once used to be and no longer are: a party that fought for the middle class and working people.

I heard the sentence that I picked for the title of my post today on a call-in show about disappearing manufacturing jobs in the United Sates. It was said in a bitter sounding voice by a former automotive worker as an expression of the hopelessness he was feeling about his job prospects.

I’ve had many bitter comments on my blog from translators who are also discouraged by the current situation of translators and the conditions created for them by the “translation industry” that they are looking for a way out, preferably a different occupation.

Although they love, or used to love translating, they don’t think it is possible to make a decent living being a translator at this point. That is what the “translation industry” has done to our profession only in the last decade or so.

I think this is also one reason why there are so many instant translation gurus everywhere who claim to have a magical solution for all translators, though mostly for those who sign up for their paid seminars. To me, the funny thing, or one of the funny things about these instant gurus, is that most of them are so young they could be my grandchildren. And I am not really that old yet … I have adult children, but no grandchildren yet.

I think that many of these “experienced translators” are cashing in by giving seminars and webinars on how to make more money translating by using extra cool technology, social media, or by finding direct clients … because they can’t make enough money just by translating.

You can tell really good writers by how they write really good or at least popular books, which is to say books that many readers want to buy. Writers who are not as good teach creative writing at a community college because they too have bills to pay.

And really good translators are so busy translating, even in the brave new world created for them by the “translation industry”, that they don’t need to give webinars to make ends meet because their customers are keeping them busy translating at very good rates.

The equivalent of an autoworker who is fixing robots, robots doing what the worker used to be doing because he can’t find any other work, is a translator who fixes machine translations because he can’t find any real translation work.

The difference here, or at least one of the differences, is that while it is possible to design a robot that will replace a worker in a car plant who performs manual labor, it is not possible to design software that will replace the thinking that goes on inside the brain of a translator.

Human languages are too complicated. They are as complicated as human thought processes. If the “translation industry” were honest with its clients, it would have to admit this simple fact and make it clear that machine translation is only a tool. Although it might look like real translation, it is not really translation.

But they can’t say that because if they did, who would buy it?

Even beautiful sounding sentences can be a trap, if these sentences are translated by machine translation software rather than a human being. The problem with machine translations that are based on the statistical approach pioneered by Google is that beautiful sentences that sound just like sentences created by human translators often say the exact opposite of what was said in the original language.

I have seen many examples of texts in descriptions of patents that were translated by machine translation and that looked very impressive to me, a patent translator who has been translating patents for a living since May 1987 – nearly 30 years now. The sentences looked perfect, but they said the opposite of what was said in the original. After all, the software is only based on statistical probability.

The “translation industry” says that this problem will be solved by what the industry calls post-editing of machine translations by humans. Sometimes, the industry even calls this kind of unappetizing editing of raw machine output by humans “copy editing” to make it sound more appealing to translators and customers.

But several fatal flaws are hidden in this concept of humans whose job it is to repair a robotic system for translation of human thoughts.

When machine-translated sentences are obvious mistranslations because they sound hilarious or because they clearly make no sense, it is easy to see where the likely problem is in the machine-translated output. But when they sound like real sentences created by a real human being (because that is what they originally were, except for the fact that the meaning of the original text may now be the complete opposite of what the machine translation says), it is very difficult and extremely time-consuming to find out where the problem might be, even more time-consuming than translating everything from scratch.

It is highly unlikely that humans hired at slave wages will bother trying to find mistakes in machine-translation output for one simple reason: the “translation industry” wants to pay them so little that even if the “post-processors” want to do a good job, they won’t have the time to do it.

I have been contacted by translation agencies several times already to work as a “copy editor of machine translations”. Out of curiosity, I called twice to inquire how much would I be paid for this work.

Both times I was offered the princely sum of 1 cent a word, as I wrote in a post titled So Now I Know How Much They Want to Pay Us a little over a year ago.

Some former car plant workers did become robot repairmen who were fixing robots that had replaced them … until they were fired even from this job once the robots were fixed.

But I can’t imagine that real translators would be willing and able to repair the product of machine translation robots for a living, no matter how desperate these translators might be, at the pitiful rates that the “translation industry” is willing to pay for such a horrible job.