Quote from: TranslateME on September 15, 2021, 06:05:14 PM

Quote from: j4rd4 on September 11, 2021, 02:58:27 PM

Complete shit.
"Founders" know shit about natural languages, so they suppose, that editing a machine translation would be easier than to create the real translation from the scratch.
There is no way to create a cake (a natural language act) from the shit (computer-generated garbage), even if it was the cake before.
To do a translation "online" is pretty naive notion of the translator's work. Haste makes waste: It is a highly expert professionals, which requires (rare) experience and (exhaustive) concentration. You completely ignore the fact that - if you ever get something like this up and running - mainly beggars, morons and people intent on doing harm will post their pseudo-translations and your AI will do nothing against that. In your naivety, you won't even be able to begin to find a way to filter out this drivel - even when someone (like I am right now) informs you of the problem.
Because it's not like pulling money out of lazy tourists.
You've invented an excuse for yet another useless cryptocurrency project, and you're going to harass other suckers with it until your laziness gets the better of you.
This is only about a hundredth of the problems you would have to overcome on the way to this naively set goal, so you greedy psychopaths better just give it up and go back to your original professions.
Maybe I'm wrong and you're not psychopaths. Maybe you are just naive children raised by social media spy networks, who are led to believe that the problem of mindfulness (attention to subtext) is merely a technical problem that can be solved by technical means. Either way, your project is doomed to fail because you consider translations to be mere data manipulation job.

Thank you for your feedback, I think the goals of TranslateMe Network might be very different to your view. Machine translation as a service is one of the fastest-growing technologies online, we started in 2017, there were only 7 well-known companies developing machine translation services. In 2021 we have over 28 companies which were recently published in a Nimdzi Atlas 2021, TranslateMe was featured for the first time this year.

Most of what you said above has actually nothing to do with TranslateMe Network, the time we have apparently wasted on been naive has yielded 14 language models, these are currently serving Millions of instant translation requests on our telegram translator application to thousands of users. We are about to partner with www.NVU.io he has over 450 000 users looking for an alternative solution to google. Many more clients will be looking for alternatives, TranslateMe Network is early in this field and we are extremely proud of the team and the results.

Maybe you are confusing the use of translators on our network with Machine Translation as a goal? We do not seek to replace human translation, this won't happen for at least 10 years, not even Google is trying to do this. We are trying to build a future for translators and show them that every word they translate has value beyond the once-off payment they receive for doing the work. What we are trying to do, if you don't mind.... is include translators in the value of machine translation by leveraging their data and assigning ownership for their data. I think for now let us prove this or at least allow us the benefit of the doubt. If we succeed we will not only be the best machine translation company but we would have succeeded in sharing this value with translators who helped build this network.

Kind Regards the future of machine translation!

What I wanted to say: So-called "machine translation" is not a translation at all. This MT can be used as some kind of "consecutive interpreting" tool for the text, but only for such reader, who does not understand the source language of the particular text. Second legitimate use case is when you want to say to somebody using a language, in which you do not feel comfortable while speaking/writing, and you both know that mistakes would happen. This is also some (new) kind of consecutive interpreting. The term translation has always been used in a completely different sense: It is the result of a creative attempt to interpret the hierarchy of meanings and desired functions of the source message in written form. Such interpretation, of course, always takes into account the perceived interests of the persons and groups concerned, so there is always a moral (and pedagogical) angle to any translation activity. This cannot be automated, just as you cannot expect patriotism or loyalty from, say, a computer system whose interface is slipped to you by an anonymous investor.
If you want to somehow mix computer-generated nonsense with the results of translators' work, or even confront translators with such computer data when they are trying hard to focus on the subtext of the original message, you will end up with superficial lego-like babble, which, after all, is already being generated in large quantities by fraudulent brokers with the help of underpaid low-IQ native speakers.
Do you realize that such computer-generated language is evaluated by you and your ilk according to the Turing test ("if an idiot can't tell difference, then it's ok")? The problem is, that the more faithful the imitation of a real person's utterance, the greater the deception, because it creates a false perception in the target groups of the existence of a non-existent person speaking to them.
More generally, artificial intelligence (deployed in interpersonal relationships, such as for generating - falsely imitating - natural language) will therefore not serve humanity, but to whomever who procures it as a weapon against others. This too has been going on for years.
You said, that there are a few top companies working on the whole thing, and that is true unfortunately. And in doing so they are also shamelessly destroying the last vestiges of the ability to understand and make personal connections in the coming generation.
As I've written before, you can't make a cake out of shit, even if the shit was a cake before.
To me you are just another of the many naive greedy boys who want to financially feast on something they have no idea about. You consider language to be data and language production to be data processing. You cannot help translators with your baby-software at all, because you have not even faintest idea what they actually do.
Another problem, which you completely avoided, is means of establishing trustworthiness of particular authors by multilevel hierarchy voting. You have simply several guys for several languages, but what if they are not that great? Take a look at their photos and compare them to your photo. Even though they may not be the best in the field, they have worked hard as translators over these texts (over every sentence and millions of words). You can see in their eyes that they have lived an interesting and long (professional) life. All you see in your eyes is selfishness, cunning and trying to make someone work for you. In the images of your engineers and technicians, you can see tiredness, boredom and struggle to concentrate. It's a surprisingly stark contrast even on the surface. Yet such a centralized notion of credibility and professional authority cannot succeed in the long or short term. You simply haven't addressed it at all, and even after my pointing it out, you don't get it - as I predicted. You console yourselves with the fact that the morons (my former colleagues and classmates) at IBM etc. have been working on this intensively (and in vain) for eight years, so you might as well get on with it yourself. There's obviously no shame in your neighborhood doing any kind of software. I've said enough. You know shit, so I'm not going to waste any more time on you. I hope I've at least warned some investors...