Pages in topic:   [1 2 3] >
What do we consider a sustainable MTPE rate?
Thread poster: Adieu
Adieu
Adieu  Identity Verified
Ukrainian to English
+ ...
Mar 28, 2021

I'm starting to suspect that I have quoted far too low, on the mistaken assumption that I would probably just end up rejecting most such proposals anyway.

What would you guys deem sustainable, in USD per word and % of your full service translation rate?


 
Sadek_A
Sadek_A  Identity Verified
Local time: 21:48
English to Arabic
+ ...
..... Mar 28, 2021

I'm afraid "sustainable" & "MTPE" can NOT be in the same sentence together, as the latter is meant to remove you from the equation altogether. That eventual ZERO can in no way represent a sustainable value!
But, if you meant a 'hedge-rate' for what is to come as a result of helping MTPE, then charge north of $0.50 (Fifty Cents)/w. Mind you, it's not word-equivalents they're looking for anymore, that phase is almost complete, now they are perfecting, transferring the human translator's way
... See more
I'm afraid "sustainable" & "MTPE" can NOT be in the same sentence together, as the latter is meant to remove you from the equation altogether. That eventual ZERO can in no way represent a sustainable value!
But, if you meant a 'hedge-rate' for what is to come as a result of helping MTPE, then charge north of $0.50 (Fifty Cents)/w. Mind you, it's not word-equivalents they're looking for anymore, that phase is almost complete, now they are perfecting, transferring the human translator's way of thinking into the machine.
Collapse


Tom in London
Pal Lederer
 
Anton Konashenok
Anton Konashenok  Identity Verified
Czech Republic
Local time: 18:48
French to English
+ ...
Measure it yourself Mar 28, 2021

In general, a fair price means you earn the same amount per hour for all linguistic services you provide. Suppose you can translate X words per hour or post-edit Y words per hour for the same kind of text, and your translation price is Z per word.
Then your fair MTPE price = Z × X/Y.

The ratio X/Y varies widely depending on the subject field, source text complexity, MT output quality, your professional experience and your criteria of acceptable quality. Don't be surprised if
... See more
In general, a fair price means you earn the same amount per hour for all linguistic services you provide. Suppose you can translate X words per hour or post-edit Y words per hour for the same kind of text, and your translation price is Z per word.
Then your fair MTPE price = Z × X/Y.

The ratio X/Y varies widely depending on the subject field, source text complexity, MT output quality, your professional experience and your criteria of acceptable quality. Don't be surprised if it takes you longer to post-edit than to translate the same text from scratch - to me, this is the case all the time, and since nobody wants to pay more for MTPE than for human translation, I don't do MTPE.
Collapse


Peter Shortall
Kuochoe Nikoi-Kotei
Tony Keily
Laura Kingdon
Tretyak
Christopher Schmidt
Jan Truper
 
Adieu
Adieu  Identity Verified
Ukrainian to English
+ ...
TOPIC STARTER
... Mar 28, 2021

Mmm yeah I gathered that much. That's how I first quoted.

But after some first-hand experience, I also concluded that my current 50% rate was far too low.

Now, I seem to have an opportunity to right that wrong, and am curious if I should go 67%, 75%, or 80%

[Edited at 2021-03-28 11:01 GMT]


 
Adieu
Adieu  Identity Verified
Ukrainian to English
+ ...
TOPIC STARTER
The ship has sailed Mar 28, 2021

Sadek_A wrote:

I'm afraid "sustainable" & "MTPE" can NOT be in the same sentence together, as the latter is meant to remove you from the equation altogether. That eventual ZERO can in no way represent a sustainable value!
But, if you meant a 'hedge-rate' for what is to come as a result of helping MTPE, then charge north of $0.50 (Fifty Cents)/w. Mind you, it's not word-equivalents they're looking for anymore, that phase is almost complete, now they are perfecting, transferring the human translator's way of thinking into the machine.


As late as last year, I had a perfectly militant proud luddite "I don't do CATs" stance in most of my resumes.

Eff it, I give up.

Now I just want guidelines on how to milk this sinking ship while still viable.

PS also, frankly, I'm too lazy to go back to typing consecutive complete sentences.

[Edited at 2021-03-28 15:43 GMT]


Jorge Payan
Michael Newton
 
Sadek_A
Sadek_A  Identity Verified
Local time: 21:48
English to Arabic
+ ...
..... Mar 28, 2021

Adieu wrote:
I give up.
Ship has sailed.
How to milk this sinking ship while still viable.

Sinking is NOT Sunken.
Wise stakeholders, midsea, ditch water away not attract more in.
But, you are entitled to embracing your risk for your language pair(s).
May I ask what you plan to do for a living after sinking officially becomes sunken?


 
Anton Konashenok
Anton Konashenok  Identity Verified
Czech Republic
Local time: 18:48
French to English
+ ...
Disagree with Sadek Mar 28, 2021

Sadek_A wrote:
it's not word-equivalents they're looking for anymore, that phase is almost complete, now they are perfecting, transferring the human translator's way of thinking into the machine.


In my subject fields, MT has a long way to go with correct terminology. And replicating the human translator's way of thinking is currently well beyond the reach of even the most advanced AI. It can merely fake the human way of thinking at a mediocre level.


Baran Keki
Laura Kingdon
Angie Garbarino
SilenceOnTou (X)
Francesca Ciet
P.L.F. Persio
SandraV
 
Stepan Konev
Stepan Konev  Identity Verified
Russian Federation
Local time: 20:48
English to Russian
MTPE rate = translation rate Mar 28, 2021

I only accept MTPE jobs for the same rate as for translation jobs. And I don't explain my point to clients. This is just my pricing policy. Take it or leave it.

Matthias Brombach
Arabic & More
Laura Kingdon
Ines Radionovas-Lagoutte, PhD
SilenceOnTou (X)
Daryo
Vasaporn Chaiyakul
 
Sadek_A
Sadek_A  Identity Verified
Local time: 21:48
English to Arabic
+ ...
..... Mar 28, 2021

Anton Konashenok wrote:
Disagree with Sadek

Anton, shouldn't you disagree with Adieu first? He declared a sinking ship!
Anton Konashenok wrote:
In my subject fields, MT has a long way to go with correct terminology.

For the sake of us, your colleagues, current and future, becoming on the same page with your point, care to give us examples of said subject fields and how MT is failing with correct terminology in samples of those fields?
Anton Konashenok wrote:
And replicating the human translator's way of thinking is currently well beyond the reach of even the most advanced AI. It can merely fake the human way of thinking at a mediocre level.

It's not replication, as this would suggest the machine will later be able to adjust its thinking on its own like a human can do; however, it's imitation they're after.

You, human translator, get the MTPE text, locate the word with the incorrect use, send back to them, they cook an algorithm that imitates the way in which you were able to say "that is wrong use, now this is correct use", by linking the desired use with other variables within the same text to act like determinants.


 
Anton Konashenok
Anton Konashenok  Identity Verified
Czech Republic
Local time: 18:48
French to English
+ ...
to Sadek Mar 28, 2021

For the sake of us, your colleagues, current and future, becoming on the same page with your point, care to give us examples of said subject fields and how MT is failing with correct terminology in samples of those fields?

The main field in question, in which I have been asked to do PEMT many times, is medicine, or more specifically, clinical research.
Out of curiosity, I also checked MT output in aviation and several engineering subjects. The main problem with terminology is that one and the same word may have totally different translations in different subdomains of one broad subject field or in different contexts. MT is unable to distinguish between them. Actually, to be fair, inexperienced translators may be equally clueless.

It's not replication, as this would suggest the machine will later be able to adjust its thinking on its own like a human can do; however, it's imitation they're after.

Exactly my point. As a onetime university lecturer, I can't help comparing AI to a lazy student attempting to pass a complex exam, trying to shoehorn a non-trivial problem into a rigid framework of trivial reasoning based on an inadequate set of rules.
You, human translator, get the MTPE text, locate the word with the incorrect use, send back to them, they cook an algorithm that imitates the way in which you were able to say "that is wrong use, now this is correct use", by linking the desired use with other variables within the same text to act like determinants.

Not exactly. Cooking a new algorithm will take orders of magnitude more work than I spend correcting the errors. In the vast majority of cases, my response is fed back into the same adaptive faking algorithm. Furthermore, for every response by a top-flight specialist, the system will also receive a dozen responses from mediocre ones, so the system thus trained is condemned to eternal mediocrity anyway.


Adieu
Baran Keki
Alison Jenner
SilenceOnTou (X)
Inga Petkelyte
P.L.F. Persio
Zolboo Batbold
 
Adieu
Adieu  Identity Verified
Ukrainian to English
+ ...
TOPIC STARTER
True Mar 29, 2021

I literally watch their system acquire new quirks and un-learn previously known correct answers overnight.

And the deliciously hopeless two-tier review process makes it SO MUCH WORSE.


CroPro
 
Adieu
Adieu  Identity Verified
Ukrainian to English
+ ...
TOPIC STARTER
ANYWAYS Mar 29, 2021

I know it sucks.

That is why I am asking how much MORE to charge.


 
Sadek_A
Sadek_A  Identity Verified
Local time: 21:48
English to Arabic
+ ...
to Anton Mar 29, 2021

Anton Konashenok wrote:
MT is unable to distinguish between them.

Hence, my aforementioned
Sadek_A wrote:
imitates the way in which you were [capable of] linking the desired use with other variables within the same text to act like determinants

Anton Konashenok wrote:
can't help comparing AI to a lazy student attempting to pass a complex exam, trying to shoehorn a non-trivial problem into a rigid framework of trivial reasoning based on an inadequate set of rules.

The framework is actually tailored specifically for the problem, and running based on elimination of errors, one by one, until the matrix of links is complete.
The human brain has two separate processes: thinking & adjusting; through their "rigid" process of elimination they will imitate the 'thinking' part, where the machine will be able to deploy all the drawn links into reaching the successful distinguishment.
The 'adjusting' part, however, will prove much trickier and heavier to handle by them.
Anton Konashenok wrote:
Cooking a new algorithm will take orders of magnitude more work than I spend correcting the errors.

Not really. You didn't know where the error was, had to go through the whole document for full context, research external references for support, compare recurring instances within the document, sift through memorized vocabulary, and eventually eliminate all invalid possibilities one by one.
They, however, after your help, know where the error is and what it should be corrected to.
Let's take that unusual expression you used, shoehorn, for example:

-algorithm-
1. Declare variables shoehorn, into and AI
2. Verify_present_all shoehorn, into and AI
3. Replace shoehorn with shove

Before your help, shoehorn could've meant either shoespoon or shove. You located, eliminated and assigned, and so will the machine henceforth.

Anton Konashenok wrote:
for every response by a top-flight specialist, the system will also receive a dozen responses from mediocre ones

That's why they should not cheapen. Hire well and pay well, is what they should be doing.
Anton Konashenok wrote:
the system thus trained is condemned to eternal mediocrity anyway

Unless!....they are actually categorizing contributors beforehand, which could well be the case by now.


 
Kuochoe Nikoi-Kotei
Kuochoe Nikoi-Kotei  Identity Verified
Ghana
Local time: 17:48
Japanese to English
100% Mar 29, 2021

Adieu wrote:

I know it sucks.

That is why I am asking how much MORE to charge.

Charge your normal rate, then.


LatGerIta
 
Adieu
Adieu  Identity Verified
Ukrainian to English
+ ...
TOPIC STARTER
Alas Mar 29, 2021

Kuochoe Nikoi-Kotei wrote:

Adieu wrote:

I know it sucks.

That is why I am asking how much MORE to charge.

Charge your normal rate, then.


Fables instruct us to cook unwitting frogs via gradual temperature increases.

Too much of a bump, alas. Besides, that would ethically necessitate my full and undivided attention, while MTPE with some discount can still be brushed off as "meh, whatever, close enough".



[Edited at 2021-03-29 10:56 GMT]


slot bet
 
Pages in topic:   [1 2 3] >


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


What do we consider a sustainable MTPE rate?







Anycount & Translation Office 3000
Translation Office 3000

Translation Office 3000 is an advanced accounting tool for freelance translators and small agencies. TO3000 easily and seamlessly integrates with the business life of professional freelance translators.

More info »
Trados Studio 2022 Freelance
The leading translation software used by over 270,000 translators.

Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop and cloud solution, empowering you to work in the most efficient and cost-effective way.

More info »