Hacker News Reader

Comment by carlosjobim | Hacker News Reader

Let's take the highest perspective possible:

What is the value of technology which allows people communicate clearly with other people of any language? That is what these large language models have achieved. We can now translate pretty much perfectly between all the languages in the world. The curse from the tower of Babel has been lifted.

There will be a time in the future, when people will not be able to comprehend that you couldn't exchange information regardless of personal language skills.

So what is the value of that? Economically, culturally, politically, spiritually?

View on HN

Language is a lot deeper than that. It's like if I say "we speak the same language", it means a lot more than just the ability to translate. It's talking about a shared past and worldview and hopefully future which I/we intend to invest in.

Then are you better off by not being able to communicate anything?

You could make the same argument about video conferencing: Yes, you can now talk to anyone anywhere anytime, and it's amazing. But somehow all big companies are convinced that in-person office work is more productive.

>The curse from the tower of Babel has been lifted.

It wasn't a curse. It was basically divine punishment for hubris. Maybe the reference is a bit on the nose.

Which languages couldn't we translate before? Not you, the individual. We, humanity?

Machine translation was horrible and completely unreliable before LLMs. And human translators are very expensive and slow in comparison.

LLM is for translation as computers were for calculating. Sure, you could do without them before. They used to have entire buildings with office workers whose job it was to compute.

Google translate worked great long before LLMs.

I disagree. It worked passably and was better than no translation. The depth, correctness, and nuance is much better with LLMs.

The only reason to think that is not knowing when Google switched to using LLMs. The radical change is well documented.

LLMs are not they only "AI"

It really didn't. There were many languages which it couldn't handle at all, just making completely garbled output. It wasn't possible to use Google Translate professionally.

I don't think you understand how off that statement is. It's also pretty ignorant considering Google Translate barely worked at all for many languages. So no, it didn't work great and even for the best possible language pair Google Translate is not in the same ballpark.

Not really long before, although I suppose it's relative. Google translate was pretty garbage until around 2016-2017 and then it started really improving

We could communicate with people before LLMs just fine though? We have hand gestures and some people learn multiple languages and google translate was pretty solid. I got by just fine in countries where I didn’t know the language because hand gestures work or someone speaks English.

What is the value of losing our uniqueness to a computer that lies and makes us all talk the same?

Incredible that we happen to be alive at the exact moment humanity peaked in its interlingual communication. With Google Translate and hand gestures there is no need to evolve it any further.

You can maybe order in a restaurant or ask the way with hand gestures. But surely you must be able to take a higher perspective than your own, and realize that there's enormous amounts of exchange between nations with differing language, and all of this relies on some form of translation. Hundreds of millions of people all over the world have to deal with language barriers.

Google Translate was far from solid, the quality of translations were so bad before LLMs that it simply wasn't an option for most languages. It would sometimes even translate numbers incorrectly.

LLMs are here and Google Translate is still bad (surely, if it was easy as just plugging the miraculous perfect llms into it, it would be perfect now?), I don't think people who think we've somehow solved translation actually understand how much it still deals extremely poorly with.

And as others have said, language is more than just "I understand these words, this other person understands my words" (in the most literal sense, ignoring nuance here), but try getting that across to someone who believes you can solve language with a technical solution :)

What argument are you making? LLM translating is available to anybody to try and use right now, and you can use services like Kagi Translate or DeepL to see the evidence for yourself that they make excellent translations. I honestly don't care what Google Translate does, because nobody who is serious about translation uses it.

> And as others have said, language is more than just "I understand these words, this other person understands my words" (in the most literal sense, ignoring nuance here), but try getting that across to someone who believes you can solve language with a technical solution :)

The kind of deeply understood communication you are demanding is usually impossible even between people who have the same native tongue, from the same town and even within the same family. And people can misunderstand each other just fine without the help of AI. However, is it better to understand nothing at all, then to not understand every nuance?

Reply on news.ycombinator.com