Pages in topic:   < [1 2 3 4 5 6 7] >
Who is not using MT today?
Thread poster: Gary Evans

Tom Hoar
United States
Local time: 07:16
English
Valid points, poorly presented Aug 19

Gary Evans wrote:

My question relates to the use (or not) of MT by translators. I'm curious to know why it's not being used. My suspicion that many agencies refuse to allow translators to make use of it has largely been confirmed on this thread. Data protection is not justified if the data is safe. And if the TM works well, why not use it? Agencies are increasingly using MT and farming out postediting, and therefore threatening our income with this model. Neither you, not I are interested in working for less. MT is seems is a double-edged sword. Just shunning it would be a mistake IMO.


I agree with this paragraph. I suspect that if you had presented this as your opening statement, much of the confusion would have been avoided.

I've been selling Slate Desktop to translators, teams and agencies for 3 years. I'll share what I've learned about why translators use or don't MT. I'll keep it brief and I'm happy to dig deeper in another thread if you start one.


  1. Katalin Szilárd was right. The hype surrounding MT, ML and AI has turned off many potential customers. MT vendors flat-out lie about the improvements they're trying to sell to an unsuspecting translator. Their hype sets set unrealistic expectations with harsh disappointments.

  2. Superstitions. That is, "a notion maintained despite evidence to the contrary." Many translators learned what they know about MT about 10 years ago. That's a LONG time in computer years. Computers have changed. Software has changed. We've learned to understand how the technology. Yet, many translators live under the superstition that technology still performs like it did 10++ years ago.

  3. Exploitative business intentions. Agencies/LSPs have used the technology to exploit their position in the market. So, while a translator might be open to the technology, the business surrounding it is more trouble. The alternatives to cloud-based technologies are difficult to find.

  4. Old dogs, new tricks. A translator recently commented that she just doesn't want to learn another new technology. Fair enough.



Gary Evans
Jorge Payan
 

Katalin Horváth McClure  Identity Verified
United States
Local time: 07:16
Member (2002)
English to Hungarian
+ ...
CAT vs. MT Aug 19

Gary,
I use various CAT-tools, including MT engines.

I think there is a fundamental difference in your thinking than that of others (including myself) in terms of CAT vs. MT.
You seem to look at MT as a CAT-tool, while I take it as a completely different thing.
If I were to draw a Venn-diagram, I would have two circles (one for CAT, one for MT), with a small overlap, while you seem to include MT in the CAT circle - so you would have a larger circle for CAT, and a smaller circle inside the larger one, for MT.
I think the use of CAT-tools is widely accepted among translators as a productivity-booster and an aid in consistency, especially in areas where this is important. MT is a different story. (But I have talked about that earlier, so I stop here.)


Katalin Szilárd
Jo Macdonald
Christine Andersen
 

Katalin Szilárd  Identity Verified
Hungary
Local time: 13:16
Member (2006)
English to Hungarian
+ ...
Human translation vs. MT Aug 19

Tom Hoar wrote:


I agree with this paragraph. I suspect that if you had presented this as your opening statement, much of the confusion would have been avoided.

I've been selling Slate Desktop to translators, teams and agencies for 3 years. I'll share what I've learned about why translators use or don't MT. I'll keep it brief and I'm happy to dig deeper in another thread if you start one.


  1. Katalin Szilárd was right. The hype surrounding MT, ML and AI has turned off many potential customers. MT vendors flat-out lie about the improvements they're trying to sell to an unsuspecting translator. Their hype sets set unrealistic expectations with harsh disappointments.

  2. Superstitions. That is, "a notion maintained despite evidence to the contrary." Many translators learned what they know about MT about 10 years ago. That's a LONG time in computer years. Computers have changed. Software has changed. We've learned to understand how the technology. Yet, many translators live under the superstition that technology still performs like it did 10++ years ago.

  3. Exploitative business intentions. Agencies/LSPs have used the technology to exploit their position in the market. So, while a translator might be open to the technology, the business surrounding it is more trouble. The alternatives to cloud-based technologies are difficult to find.

  4. Old dogs, new tricks. A translator recently commented that she just doesn't want to learn another new technology. Fair enough.



Tom it is not only about hype turning off potential customers, it is not only about rejection of exploitation, it is not only about old dogs/new tricks.... So MT is rejected not because there is an aggressive campaign behind it, but

It is like convincing a true professional artist (a painter) not to use any tools to create a painting just AI.
A painting is so much more, only a professional artist will be able to understand this.

The same is true for translation. And it doesn't have to be an art/poetry translation, it is also true for highly medical, legal and technical texts. This is a thing only true linguists will understand and will stand for.

This is a painting done by AI:

https://www.bbc.com/news/av/entertainment-arts-45951803/ai-painting-to-go-under-the-hammer

I'm not a painter, but I'd never call it an art. In a painting created by a human, you notice emotions, why she/he used certain colors, you can feel the emotion of the creator. In a translation it is also true. Yes, even in highly technical texts, we are able to give 'a little bit of us' within the texts, creating new words, when there is a new invention, we can use our imagination, logic and emotions to tell what it looks like, creating metaphors for newly invented methods or when the text is about explaining something for example.

Translation is so much more than machine translation.
And if somebody works after a machine, she/he will lose this ability, this skill.
Texts will be dull, monotone or humans will not understand them, because the machine will create such terms that nobody will understand them like it happened before: "Facebook abandoned an experiment after two artificially intelligent programs appeared to be chatting to each other in a strange language only they understood."


[Edited at 2019-08-19 17:31 GMT]

for typos

[Edited at 2019-08-19 17:59 GMT]


Tom Hoar
 

Tom Hoar
United States
Local time: 07:16
English
All true and nothing mutually exclusive Aug 19

Katalin,

Everything you said is true. None of it is mutually exclusive of my points. Much of your emphases actually supports my second point about superstitions.

Today’s mainstream MT and translators’ notions about it originated from decisions made over 15 years ago. Many of those decisions were business related, not linguistic.

There’s new evidence that contradicts those notions, but people cling to those notions despite the evidence.

... See more
Katalin,

Everything you said is true. None of it is mutually exclusive of my points. Much of your emphases actually supports my second point about superstitions.

Today’s mainstream MT and translators’ notions about it originated from decisions made over 15 years ago. Many of those decisions were business related, not linguistic.

There’s new evidence that contradicts those notions, but people cling to those notions despite the evidence.

Extending your metaphor, it’s not only about the kind of paintbrush, the brush’s actual composition is important. The artist’s skills, abilities learn and willingness to how to use the brush are also important parts of the equation.

Give the best paintbrush and quality paints to an uninspired artist without skills or ability or willingness to learn, and the painting is, well, crap.

Give a toothpick and inkwell to an inspired painter with skills, an ability to learn and the willingness to learn, and the painting will be a masterpiece.

Put the best MT system in the hands of a sceptic and the sceptic will always reject the raw MT segments. IMPORTANT: I did not point that statement at anyone in this topic. I merely point out that this personality exists. I believe they’re a small minor, but they’re toxic when they get vocal.

Tom

[Edited at 2019-08-20 13:41 GMT]
Collapse


 

Tom Hoar
United States
Local time: 07:16
English
Lightening the mood,,, Aug 20

Gary Evans wrote:

...translating them into another language and then back again into the source,..


This YouTuber did just that with the lyrics of Frozen’s Let It Go soundtrack. —ENJOY!

https://youtu.be/2bVAoVlFYf0


 

Katalin Szilárd  Identity Verified
Hungary
Local time: 13:16
Member (2006)
English to Hungarian
+ ...
Unfortunately these are NOT fake news Aug 20

Tom Hoar wrote:

Wow, Katalin!

What you call "programmed hype." I'd call "fake news." I agree with all of them except #5.

Adapting to changing technology is not necessary to live a happy, healthy, fufilling life. On the whole, however, it sure makes life easier. I don't see many of us using quill and inkwell now-a-days.

Tom

Katalin Szilárd wrote:

Programmed hypes:

1. https://www.seattletimes.com/business/amazon/amazon-program-has-warehouse-workers-tweet-positive-comments-about-working-conditions/

2. Programmed hype based on the above scheme:
Freelancers please join and translate for free to help others (khmm "or help us to make more money") and post your psychedelic happiness

3. Programmed hype:
MT/AI/NLP

4. Programmed hype:
Go get retired (we don't need professional translators, who don't want to work with MT) and mentor/teach fresh translators who may be willing to use MT

5. Programmed hype:
Adapt to the recent technology


Unfortunately it is not fake news, because I have seen with my own eyes (and others with their own eyes as well), how this type of mass psychosis can be reached by bots at certain translators to work for free to help others or to use certain tools by telling positive comments about them.

I'm very far from the world of "influencers" but the same method is used there, too: bot influencers try to lead the "mass". https://sproutsocial.com/insights/fake-influencers/
https://www.nytimes.com/2018/03/12/business/media/instagram-bots.html

That's why it is extremely dangerous.


Tom Hoar
 

Rachel Fell  Identity Verified
United Kingdom
Local time: 12:16
French to English
+ ...
This reminds me... Aug 20



...of the "improvements" made to that fresco in a church Spain done by a parishioner to restore it:

https://www.bbc.co.uk/news/world-europe-19349921



Katalin Szilárd
 

Katalin Szilárd  Identity Verified
Hungary
Local time: 13:16
Member (2006)
English to Hungarian
+ ...
The human story behind Aug 20

Rachel Fell wrote:



...of the "improvements" made to that fresco in a church Spain done by a parishioner to restore it:

https://www.bbc.co.uk/news/world-europe-19349921



Hi Rachel,

Yes.
And actually the Spanish fresco has more "art" in it, because it has a human story behind it, and it was done by human hands, yes, amateur hands but still I would call it more an art than the painting done by AI.

Bests,
Katalin


Rachel Fell
 

Katalin Szilárd  Identity Verified
Hungary
Local time: 13:16
Member (2006)
English to Hungarian
+ ...
The motive behind Aug 20

Tom Hoar wrote:

Adapting to changing technology is not necessary to live a happy, healthy, fufilling life. On the whole, however, it sure makes life easier. I don't see many of us using quill and inkwell now-a-days.

Tom


I'm not against technology at all but what is going on here with the agressive campaign of MT/AI and bots and other tools, goes way beyond improving productivity and "helping" translators or end clients, it is rather about how to make certain companies mega companies, so making certain people disgustingly rich on the expense of destroying freelancers' own businesses. So the motive is greed. And although being greedy is not illegal, I think many of these tactics what we freelancers have experienced in the last 2-3 years are illegal in many countries.

First a universal law should had to be accepted before using these tools so to avoid misusing them.


 

Christine Andersen  Identity Verified
Denmark
Local time: 13:16
Member (2003)
Danish to English
+ ...
I am not superstitious about MT. I just do not find it useful Aug 20

BTW Gary, have you discovered how to edit a post when you quote it, and highlight the relevant sections or eliminate the bits we have already ploughed through several times?
Some of your postings remind me of a Yahoo group I belong to - the technology just so 20th century…

Tom Hoar wrote:
….
  • Superstitions. That is, "a notion maintained despite evidence to the contrary." Many translators learned what they know about MT about 10 years ago. That's a LONG time in computer years. Computers have changed. Software has changed. We've learned to understand how the technology. Yet, many translators live under the superstition that technology still performs like it did 10++ years ago.

  • ...
  • Old dogs, new tricks. A translator recently commented that she just doesn't want to learn another new technology. Fair enough.



  • That was me. I have always been quite positive about new technology, but I do not accept superstitiously that every new development is something I just have to have.

    I go with an open mind to talks and panel discussions about MT. I have in fact tried it within the last couple of years. So I know it is a lot better now than it used to be.
    But I am past pension age, and I have, over the years, installed and tried out several CATs. I have been through WordPerfect, Trados Workbench, and quite a lot of technology since my schooldays, when anything more technical than chalk and talk and often pre-war textbooks was considered quite advanced.
    We had a language lab with reel-to-reel tape recorders when I was at college, but there was no screen on the computer (singular!!), and one of our lecturers seriously believed there was no real point in introducing computers into libraries. We programmed in COBOL to prove his point - it really was far more laborious than just putting the books physically in alphabetical order, which all second-year student librarians could do in their sleep.

    My first job after college included helping to convert a manual card index (pieces of cardboard arranged in a box) for online searching. Still no screen on the company computer. I never saw it. I wrote abstracts in longhand and proofread blue-striped leporello printouts. (It was the BHRA Fluidex database.)
    I have learnt to use tape cassettes and CDs instead of LP gramophone records, and I can even handle a smartphone if I have to.

    I am not superstitiously opposed to MT, but for the translation jobs I do and in my language pairs, it leaves a lot to be desired, and I can find clients who are happy for me to translate without it.

    The return on investment in time and energy is just too small. I do not consider it worthwhile to get my head around PEMT - it DOES take a different approach from translating from scratch. Using MT means looking critically at its offerings, instead of simply translating in a flow, and it is not what I signed up for.

    Newly qualified translators who start now will have to learn to use the new technology as it is today, and whatever comes in their lifetimes. I can drop out if and when the way I work is really outdated, but I suspect I can carry on as long as I am able to, and others will go on for some time after that.


    John Fossey
     

    Tom Hoar
    United States
    Local time: 07:16
    English
    Another case of changed meaning... Aug 21

    Christine Andersen wrote:

    I am not superstitious about MT...


    Christine, I chose the word "superstition" carefully, not "superstitious," not "superstitiously," but "superstition." I used the noun as defined in the Merriam-Webster online dictionary's second definition, meaning "notion" (idea or belief). I'm not sure why you changed the meaning of my comment. I don't know you and had no intention to call you "superstitious" about anything (or anyone else in this topic for that matter).

    The fact is, many comments expressed in this topic, including yours, reflect notions that people hold to despite evidence to the contrary (please revert to the definition). I won't repeat myself. If you reread my previous comment within the context of that definition and you'll find what I believe is the genesis of those notions.

    [Edited at 2019-08-21 02:10 GMT]


     

    Maxi Schwarz  Identity Verified
    Local time: 06:16
    German to English
    + ...
    answering the question directly Aug 21

    I've been following the discussion for a while. I'd like to answer the question as posted in the beginning.
    Gary Evans wrote:

    I've been testing machine translation tools for a while now. Considering how many CAT tools now have MT APIs, isn't the use of MT simply logical? So my question is: who is not using MT and why?

    I'd be interested in some honest answers here!

    Well, actually, I'd like to answer the question without the preamble. So the honest answer is to the question
    Who is not using MT and why?

    I am not using it, because I have not found a need for it. From time to time I pop part of a text in an MT out of curiosity, and the results are not that useful. We probably translated different types of material.

    Going to the question as posted:
    Considering how many CAT tools now have MT APIs, isn't the use of MT simply logical? So my question is: who is not using MT and why?

    In my choices, I don't consider what CAT tools have. I consider what my translation needs are, which is thing I assess constantly. I also don't use CAT tools - what these contain has no logical bearing on what I choose to use. What does have a logical bearing is what actually works for my needs, and that is found out largely by experimentation and observation.


     

    Gary Evans  Identity Verified
    Germany
    Local time: 13:16
    Member (2007)
    German to English
    TOPIC STARTER
    So if you don't use CAT tools... Aug 21

    Hi Maxi,

    One of my first translation jobs was to translate three reports. I translated the first one using a CAT tool. I think it was Trados. The remaining two reports were practically the same. Trados did 95% of the remaining translating using the translation memory (TM) it had built from the first report. Obviously this saved me a lot of time!

    I quickly recognised the value of a good term base (TB) as it ensured consistency, especially when translating technical repo
    ... See more
    Hi Maxi,

    One of my first translation jobs was to translate three reports. I translated the first one using a CAT tool. I think it was Trados. The remaining two reports were practically the same. Trados did 95% of the remaining translating using the translation memory (TM) it had built from the first report. Obviously this saved me a lot of time!

    I quickly recognised the value of a good term base (TB) as it ensured consistency, especially when translating technical reports etc. What's really cool is that you can export/import the TM and the TB when working with others on big projects. Quality assurance (QA) tools are also highly valuable. Memsource has an excellent QA. Both consistency and quality are improved when you use them. The alternative of downloading approved glossies in pdfs from Marketing is simply nuts.

    I now work with FoundCAT, my own CAT tool, which has multiple machine translation APIs. I use this for my teaching work, so it's not so relevant to this debate. Students find it educational. Some have found that by using CAT tools, their interest in the translation industry has been awakened.

    Of course it all comes down to what you have to translate. I wouldn't recommend CAT tools for poetry (or AI for painting for that matter), but they are really useful tools, which is demonstrated by the size of the industry supplying CAT tools and the increasing interest in MT.

    Regards,
    Gary


    Maxi Schwarz wrote:

    I've been following the discussion for a while. I'd like to answer the question as posted in the beginning.
    Gary Evans wrote:

    I've been testing machine translation tools for a while now. Considering how many CAT tools now have MT APIs, isn't the use of MT simply logical? So my question is: who is not using MT and why?

    I'd be interested in some honest answers here!

    Well, actually, I'd like to answer the question without the preamble. So the honest answer is to the question
    Who is not using MT and why?

    I am not using it, because I have not found a need for it. From time to time I pop part of a text in an MT out of curiosity, and the results are not that useful. We probably translated different types of material.

    Going to the question as posted:
    Considering how many CAT tools now have MT APIs, isn't the use of MT simply logical? So my question is: who is not using MT and why?

    In my choices, I don't consider what CAT tools have. I consider what my translation needs are, which is thing I assess constantly. I also don't use CAT tools - what these contain has no logical bearing on what I choose to use. What does have a logical bearing is what actually works for my needs, and that is found out largely by experimentation and observation.
    Collapse


     

    Gary Evans  Identity Verified
    Germany
    Local time: 13:16
    Member (2007)
    German to English
    TOPIC STARTER
    If you don't like the message, just claim it's fake news Aug 21

    Hi Katalin,

    Katalin Szilárd wrote:

    Unfortunately it is not fake news, because I have seen with my own eyes (and others with their own eyes as well), how this type of mass psychosis can be reached by bots at certain translators to work for free to help others or to use certain tools by telling positive comments about them.

    I'm very far from the world of "influencers" but the same method is used there, too: bot influencers try to lead the "mass". https://sproutsocial.com/insights/fake-influencers/
    https://www.nytimes.com/2018/03/12/business/media/instagram-bots.html

    That's why it is extremely dangerous.


    You fail to spot the difference here. You claim I'm a fake influencer, possibly assisted by an army of bots...

    Well, I've been a translator for nearly twenty years. I joined Proz in 2007, got certified here and have worked to certify other applicants here on Proz. I take no pay from anyone to promote CAT tools. Clearly Tom has a commercial interest in selling his MT tool. I don't blame him for promoting it frankly. I however, am genuinely interested in how MT can help us in our work. I realise some agencies are forcing PEMT on the industry. It's not a good business model for us so we should be debating how MT can work for (some/most) of us instead.

    If I am seeking to influence people here, then hopefully for our mutual benefit. Believe me, no one is being forced to work for free here. That's just nonsense. Now please stop making false claims about me.

    Regards,
    Gary


     

    Katalin Szilárd  Identity Verified
    Hungary
    Local time: 13:16
    Member (2006)
    English to Hungarian
    + ...
    Roles Aug 21

    Gary Evans wrote:

    Hi Katalin,

    Katalin Szilárd wrote:

    Unfortunately it is not fake news, because I have seen with my own eyes (and others with their own eyes as well), how this type of mass psychosis can be reached by bots at certain translators to work for free to help others or to use certain tools by telling positive comments about them.

    I'm very far from the world of "influencers" but the same method is used there, too: bot influencers try to lead the "mass". https://sproutsocial.com/insights/fake-influencers/
    https://www.nytimes.com/2018/03/12/business/media/instagram-bots.html

    That's why it is extremely dangerous.


    You fail to spot the difference here. You claim I'm a fake influencer, possibly assisted by an army of bots...

    Well, I've been a translator for nearly twenty years. I joined Proz in 2007, got certified here and have worked to certify other applicants here on Proz. I take no pay from anyone to promote CAT tools. Clearly Tom has a commercial interest in selling his MT tool. I don't blame him for promoting it frankly. I however, am genuinely interested in how MT can help us in our work. I realise some agencies are forcing PEMT on the industry. It's not a good business model for us so we should be debating how MT can work for (some/most) of us instead.

    If I am seeking to influence people here, then hopefully for our mutual benefit. Believe me, no one is being forced to work for free here. That's just nonsense. Now please stop making false claims about me.

    Regards,
    Gary


    Hi Gary,

    I rather see a "good cop, bad cop" scenario here with the same goal.
    The problem is that roles are reversed and we freelancers are definately not the "BAD guys".

    I'm out of this conversation now.

    Bests,
    Katalin


     
    Pages in topic:   < [1 2 3 4 5 6 7] >


    There is no moderator assigned specifically to this forum.
    To report site rules violations or get help, please contact site staff »


    Who is not using MT today?

    Advanced search







    Déjà Vu X3
    Try it, Love it

    Find out why Déjà Vu is today the most flexible, customizable and user-friendly tool on the market. See the brand new features in action: *Completely redesigned user interface *Live Preview *Inline spell checking *Inline

    More info »
    SDL Trados Business Manager Lite
    Create customer quotes and invoices from within SDL Trados Studio

    SDL Trados Business Manager Lite helps to simplify and speed up some of the daily tasks, such as invoicing and reporting, associated with running your freelance translation business.

    More info »



    Forums
    • All of ProZ.com
    • Term search
    • Jobs
    • Forums
    • Multiple search