Image processing, machine learning and computer vision come to my mind. How do you translate “Convolutional Neural Networks”, “back propagation”, “loss function”, “noise-contrastive learning”, “ Generative Adversarial Networks” in German? It’s not “just a few words” but you will essentially end up speaking German with mixing it 30% with “english words that bring the actual science and innovation across”.

No meaningful papers in these domains are published in German...

You could translate these as "faltendes neuronales Netzwerk", "Zurückübertragung", "Verlustfunktion", "Rauschkontrastlernen" and "generative Gegnernetze".

Of course most people don't do that, because they learned the terminology in English and there's no particular reason to invent a completely parallel set of made-up terms. But if there were such a reason (e.g. if English literacy were uncommon among German computer scientists) it could be done.

I have never in my life heard any of these words. They are poorly translated and unpronounceable.

Yes, you would not make up these words just for the sake of speaking in German. We are not in the 30s any longer.

But no person in its right mind would speak German with 30% of the words “borrowed from english” because “all the other advantages of the German language outweigh borrowing so many english words”.

I love how people are downvoting me in their evening frustration about their lives - but yeah, if you miss like all the relevant vocabulary in a language why would you state that communicating in it is “smart of efficient”? There is no inherent reason. English has -for many fields- the best vocabulary. The english language does not get lost in four line sentences so easily like the German language and it’s not made up of 80 character words.

Have any of you ever read a German paper or textbook on a topic? Its awful. I love english textbooks I love english papers.

Right, but that was not the issue. The issue was a language suited for rules and relations on the one hand, and a language of humor and expressiveness on the other.

Of course, today everyone publishes in English. And of course, many English words are used in German academia.

But there is no inherent property of the English language, why it is better suited to communicating topics in machine learning.

Then why don’t I hear people discussing it in German?

Why don't you hear people discussing in German on Hacker News? Is it because German is unsuitable for informal, everyday discussions?

What?

No, but I suppose you miss all the vocabulary you need to efficiently communicate about ML in German... Or why is it?