In Denmark we have a bunch of weird ones:
When there isn't a problem: "There is no cow on the ice"/ Der er ingen ko på isen
When you're helping someone when it would be better they did it themselves you're doing them "a bears favor" / en bjørnetjeneste
When you want it both ways but cant: "You want to blow with flour in your mouth" / blæse med mel i munden. This always made more sense to me than the english, you cant have your cake and eat it too.
When something is complete gibberish, it "sounds like volapyk" / lyder som volapyk. Volapyk is an actual made-up language like esperanto. incedentaly the same expression also exists in Esperanto
I like the point about LLMs interpolating data while humans extrapolate. I think that's sums up a key difference in "learning". It's also an interesting point that we anthropomorphise ML models by using words such as learning or training, but I wonder if there are other better words to use. Fitting?
Denmark doesn't have unrestricted free speech. Censorship is against the constitution but hate speech, inciting violence or defamation is illegal and until 2017 blasphemy was illegal. There is some legal basis for punishing speech that is bringing the country into danger or is severely insulting another nation. I am not a lawyer so I don't know if it's applicable or if it will hold up in court.
Note that the government does not want to ban burning books. Just prevent/punish doing it as a provocation against another country.
I also want to stress that I am not defending the Danish government or the book burners. I just want to bring some needed nuance to the discussion
Which is also what the article is arguing. As soon as a majority of users are subject to this the websites can discriminate. If only safari does this it's still bad but not very bad
According to the article the linear algebra algorithm has a running time of O(log n)