Before I got into freelance translation, I thought of it as completely detached from computer translations. Indeed, when we talk about Computer-Assisted Translation (CAT), we are not referring to a computer translating for us, rather we're referring to the fact specialised software is used to assist in the translation.
Every now and then, I come across some complaints in the translation industry about technology, usually this centres on CAT tools, but here is an extreme example: In it, a translator is described who essentially was using a typewriter and fax machine into the age of email and word processing. Granted, to continue doing that, they must have been both an excellent translator and precise typist; however, resisting technology in today seems like a very, very bad idea.
Let me explain. A good CAT tool today will analyse the sentence you are currently translating, search through your previous translations, and if a decent enough match is found, it will automatically perform a translation. This spares the translator anything between 10-100% of the time translating that sentence (depending on how well the current context matches the context of the previously translated sentence). Clearly, anyone already using such tools has a competitive advantage in terms of speed. They can also offer lower rates.
The problem with CAT tools is that often one is forced to offer a lower rate, because the translation agencies know the translator is doing less less hours of work for the same work. Thus, you either keep up with technology or spend longer on a project for which you'll be paid less. At the moment, the situation isn't too bad: only one of the agencies I work for requests a lower rate for translations due to the CAT tools' ability to assist with some of the document.
However, are CAT tools going to become more or less efficient at this? I think you can guess the answer. At the moment, the CAT tool spiel touts that they make translation 20% faster (or more, depending on which spiel you're looking at). As technology progresses translators will be spending less time translating, and more time checking the CAT tool's work.
Then of course, at some point one of the big technology companies is going to come out with a computer that can do the whole thing for you, then we'll only need to review. Seriously. Did you imagine 15 years ago that you could ask a phone the weather, or mark appointments in the calendar for you? I'm possibly not being realistic*, but it is undeniable that the situation can only become more technology dependent.
*I imagine that when it comes to increasing translation speed, CAT tools will find it progressively difficult as they push past +50%. But machine translation is a bit of a black box to me: words go in one side and half-understandable gibberish comes out the other. The problem is that machine translation only needs to get to the point where words go in one side and really bad English comes out the other before one can basically say "we don't need translators any more! But we are hiring reviewers!". There are already companies that offer human-edited machine translations. The other problem with this being a "black box" is that is difficult for people on the outside to work out when the algorithms are going to reach that level of proficiency.