Hacker Newsnew | past | comments | ask | show | jobs | submit | canjobear's commentslogin

> Not because the methods it displaced had stopped working, but because the money, the talent, and the prestige had moved elsewhere. The researchers who understood decision theory, Bayesian inference, and operations research didn’t lose their arguments. They lost their audience.

So, what's the problem with it?

It reads to me like Claude wrote the article too.


It's clearly in a different category from the "highbrow" examples like Solaris, just by virtue of being entertaining to a broad audience. In contrast Solaris is the kind of movie where there's a five minute unbroken scene that's just a guy driving in traffic and thinking about his life. (Like the author, I like them both!)

It's not obvious why it wouldn't, especially if it gets to read Poincaré and Riemann.

Knowing German would mostly be helpful for understanding the grammar of Old English. The three genders and four cases, participles prefixed with ge-, verbs like sindon (=sind). There are tons of cognates with German (like þurh = durch) but they're hard to recognize immediately unless you know the kinds of sound changes that are common.

I guess the concepts and some of the vocab are important (though I feel compelled to point out that þurh is cognate with through as well).

But Old English inflecting nouns, rather than relying on indefinite and definite articles, gives the language a very different quality to German. Also stuff like negative concord.


It goes all the way back to Indo-European. There wasn’t much French influence on English before the Norman invasion.

From what I’ve heard on “the history of English podcast”, after the Norman and invasion written English disappears completely for about a century. This is because the clergy and lawyers were the only literate people at the time, and they were all French. When it re-emerges, it doesn’t have much French in it yet, because only the common folk spoke English, and the norman upper class spoke French, and they didn’t interact that much. It actually took another 100 years or so for French words to percolate into the language.

What I learned from the podcast was that what really changed old English into Middle English was the viking invasion around 800. Danes and anglosaxons had different grammar and as a compromise a lot of the germanic cases on nouns, which allowed for arbitrary word ordering in a sentence, got discarded, and English developed the current emphasis on strict word ordering that we have today.


> There is a significant quality difference between English language output and other languages

?


They're saying LLMs do better when outputting English than other languages, an assertion I'm not really able to test but have heard elsewhere.

and this is somehow not related to the size and availability of corpora in English?

No, I'm quite sure that's why it's better.

OK but then that goes back to their other assertion that it gives a huge hint at what is going on behind the scenes, is that huge hint just "more data gives better results!" if so, that doesn't seem at all important since that is the absolutely central idea of an LLM. That is not behind the scenes at all, that is the introduction to the play as written by the author.

Not your fault obviously, but they have not yet described what that huge hint is, and I'm just at the edge of my seat with anticipation here.



That’s because engineering degrees were the only thing you could get from college during the Cultural Revolution.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: