Career Change?

[John Stroman]

Fellow 'Yakkers,

Just for your information, here is an article from the weekly Slator newsletter that summarizes a survey of 260 linguists and how their careers are trending.

https://slator.com/50-of-freelance-linguists-have-thought-about-changing-career-slator-survey-finds/?mc_cid=b4f5fd632f&mc_eid=fb7469eae2

I'm not sure how relevant this is to Japanese, but it does provide some general information that may be useful

John Stroman

----------------

[Tom Gally]

Thanks for posting that article, John. Interesting results.

One comment:

... with freelance interpreters more likely to report an upward trend compared to freelance translators.

Last Saturday, I gave a talk about AI trends to a group of interpreters in Tokyo. Those I spoke with confirmed what I had guessed: that AI interpreting is not yet at a level where human interpreters’ jobs are threatened. One interpreter (who is on this list) told me that he has heard no reports yet of interpreters losing work to AI.

I think that will probably change in the coming months and years as better and faster AI voice interpretation becomes available, especially AI with large context windows and with visual input for recognizing gestures, facial expressions, etc. My guess, though, is that there will continue to be demand for human interpreters in situations where the interpreter’s physical presence and identity and feelings as a real person are valued.

I have been waiting eagerly for months to try out OpenAI’s Advanced Voice feature to see how well it can interpret spoken conversations, but they are rolling it out very slowly.

Another comment:

One complaint I have seen about AI-generated texts and images, and which I agree with, is that, at least with simple prompting, the output tends towards a homogenous mediocrity. Texts lack an individual voice, and faces—though superficially diverse—have a similar level of symmetricity and attractiveness. One explanation is that the models are just producing high-level averages of the data they were trained on, so they tend to converge on similar midpoints. Thus the blandness of so much AI-generated slop.

As I continue to use LLMs to help me with my translation work, though, I am coming to realize that such blandness can actually be helpful in translation. I have mainly done writer-driven work that, ideally, should not seem like it was translated from Japanese, but it has always been hard for me to scrub the Japanese-ness completely from my translations. Now, if I am unhappy about how a translation I have produced sounds, I will give a couple of LLMs the original Japanese and my draft translation and ask them to rewrite my English so that it reads more smoothly and naturally without departing significantly from the meaning of the Japanese. I don’t accept all of their suggestions, of course, but often their output is better than what I could have produced on my own. They have been trained on vastly more English than I have been exposed to in my entire life, and their “sense” of what is the most natural, commonplace English is better than mine.

And one more:

A few days ago I did a translation job for which, as I usually do now, I started with the translations produced by the latest versions of several LLMs. This time I used ChatGPT 4o, Claude 3.5 Sonnet, Gemini 1.5 Pro Experimental 0827, Mistral Large 2, and Llama 3.1 405B. It might have been due to how I prompted them and the nature of the text, but, to my surprise, for the first time Gemini produced the best draft, followed by Claude and ChatGPT. Mistral and Llama were obviously inferior, and I ended up not using anything from their drafts.

Tom Gally

--
You received this message because you are subscribed to the Google Groups "Honyaku E<>J translation list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to honyaku+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/honyaku/CAJVTx94Vxgz2mJDULd%3DVALUvSukviwtmfHZGq0nYQF81G_cd5w%40mail.gmail.com.

[Matthew Schlecht]

Thanks for posting this, John!

I imagine that the data in the survey pertain mainly to European languages, and that the changes to the language services workspace for Japanese, Korean, and Chinese(s) are less pronounced at present, but that would be mainly due to the fact that developers have focused on Western languages during development.

That will change.

The majority attributed this drop to AI and believe that the trend will continue over the next five years.
Most definitely, especially if you define an effect by AI broadly, and I believe the trend will continue well past five years.

The data also revealed that one in five freelance translators and interpreters have actively applied for new jobs during the same period.

Based on discussions with translator colleagues, again mainly those who work in Western languages, I would say that the "one in five" figure is conservative, and that many more are considering career changes, either actively or aspirationally.

Matthew Schlecht, PhD
Word Alchemy Translation, Inc.
Newark, DE, USA
wordalchemytranslation.com

[Warren Smith]

I have to wonder how much of the drop in work is because of the overly strong dollar as well.

 

 

 

---

--

You received this message because you are subscribed to the Google Groups "Honyaku E<>J translation list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to honyaku+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/honyaku/CALnYxczpD3-u%2BFDmcyXq3GNCJeidjTVx-hS58_WXU1T6t46t1g%40mail.gmail.com.