translation is dead as a profession

Charles R. wrote:



We'll soon see rates that even the AI will refuse to work for.

Vladimir Pochinov wrote:

A study by researchers at Oxford University and Deloitte carried out in 2015 concluded that about 35% of current jobs in the UK are at high risk of computerisation over the following 20 years.

Below you can see 10 job titles facing the highest risk of losing out to robots.

Ice Scream wrote:

I agree with everything else you say, Exy, but I'm not sure about this. Most translators seem happy to hasten their own demise by collaborating with MT. Extrapolate that to humanity in general and...

For the record, I would dispute pretty much everything you said about English, Welsh, Swedish, Danish and Norwegian, all of which are languages I know a thing or two about. I might be able to do a really good impression of an American but I would never actually be one, so I would never want to try to write like one.

Vladimir Pochinov wrote:

A study by researchers at Oxford University and Deloitte carried out in 2015 concluded that about 35% of current jobs in the UK are at high risk of computerisation over the following 20 years.

David GAY wrote:
In 2015, there was no PEMT.

expressisverbis wrote:



According to the European Terminology Coordination:
Post-editing is possibly the oldest form of human-machine cooperation for translation, having been a common practice for just about as long as operational machine translation systems have existed.
https://termcoord.eu/wp-content/uploads/2013/08/Post-editing-of-Machine-Translation.pdf

In 2011, as far as I can remember, SDL BeGlobal was implemented in SDL Trados Studio 2011 providing free access to users.

David GAY wrote:


I meant no PEMT jobs. I ve never seen any PEMT job offered before Deepl was released.

Adieu wrote:

Also, I've seen plenty of MTPE in language pairs not supported in DeepL

David GAY wrote:


Sorry, in my pairs, there were no PEMT jobs in 2015 and no agency, even the cheapest, has ever offered me to do PEMT before 2019.But your mileage may vary. Perhaps you ve seen EN>RU PEMT jobs pretranslated using GT. Who knows. Deepl is just one of the so called NMT engines.
I can t tell for other pairs.But I know some translators who were using extensively MT before 2019 even though they were paid to do translations, not PEMT.Perhaps it s what you mean. And you are right. There are PEMT jobs which are not supported by Deepl, because NMT engines nowadays mushroom and thrive, but it was not the case in 2015. Only a few translators used MT and MT quality used to be quite poor. LSPs, even the cheapest ones, did not use MT



[Edited at 2022-07-03 23:21 GMT]

[Edited at 2022-07-03 23:22 GMT]

Adieu wrote:

No DeepL there (officially, at least, although it can somehow approximate some texts nonetheless)

Actually, I'd guess that your feelings about DeepL are entirely dictated by your place of residence. It does in fact do translations out of Bulgarian pretty well (except when it decides to disregard words or lose whole clauses).

As to Russian-English MT, that has existed for ages. I first played around with it

Adieu wrote:

But in the languages I'm familiar with, DeepL is nothing special.

Slightly better flow BUT FAR LESS RELIABLE (disregards words or whole clauses) vs. Google w/ Bulgarian, doesn't support Ukrainian, worse than Yandex with Russian.

David GAY wrote:

WIWO What I m working on or what translators are working on
Otherwise I don t work with any of your pairs.

Adieu wrote:



You mean social network posts? I think established translators have become increasingly wary of accidentally disclosing who they work for or where well-paid work can be found and thus encouraging low-cost competition

Or do you mean on ProZ? This site is dead as a doornail. There are ~20 active posters.

[Edited at 2022-07-05 23:38 GMT]