probably because it justifies the sensationalized title, although the entire content can be summed up as "LLMs don't silo data, that's probably bad."
On the flip side, I thoroughly enjoy the fact that roleplaying in a videogame setting now counts as security research. Looking forward to the arxiv preprint "LLMs start playing really good drums if you pretend you're J.K. Simmons"
my experience is the opposite. Due to LaTeX's arcane scripting and the lack of interest people have in learning it beyond "it compiles on overleaf", I'm seeing a lot of accumulated superstition. People copying and pasting preambles with useless packages, unused newcommands. Worse, people sometimes use their group's newcommands without being aware of the native functions, e.g. \beginmymatrix replacing \begin{pmatrix}. Even if change is slow, any amount of Typst adoption is good.
I'm a bit out of the loop, but do we have some grasp on the size of these closed models? Is the trick still adding an order of magnitude to weights and training data or has something changed?
I had no idea that LLMs (or the transformer architecture) were within reach of complexity theory. But if transformers "can be" exponentially more succinct than RNNs, doesn't that mean we're approaching optimality?
It makes me wonder that despite the fast improvements in model capacity (and the claims) we're still using variations on a 9-year old architecture. How is it that we haven't been able to use LLMs to actually improve that?
The article, to me, comes across as focused on art as a job, especially sentences like "greater creativity and success in creative careers". There's a ring of self-help/pop business that just strikes me as artless.
> Nothing inherently requires art
Of course not: I've used academic in the precise sense of people deciding to go through the institutions of art, and coming out with a noticeable lack of tasteful intelligence. Is art education just a quest for social self-realization? is it a sinecure for the happy few? That I have no idea illustrates the point.
Certainly, I'm also aware of how difficult it is to implement open dialogue in practice. Perhaps my hope is that general education could help develop that sort of transversal insight that talented scientists use to naturally understand topics which they are not familiar with, by working with analogies and fundamental principles. I know that knowledge of the nitty gritty generally requires years of actually struggling with the thing, and this cannot be asked of any layman. Still, for example, I'm thinking of times when you deal with a topic that is nominally in the same field as yours, but that is so foreign that the only knowledge relevant to it is something barely above undergraduate, say Newton's laws or thermodynamics. Many scientists have managed to either take some lessons from other fields and bring them into theirs, or contribute despite their relative lack of education in that subfield.
I'd like to believe there is a sort of education that allows people not to understand details, but at least to be able to get the rough shape of the topic at hand and shape their ideas in a way that benefits the other party. Perhaps this is just a matter of language and shouldn't need so much more education than the basics and curiosity. Or perhaps it's a pipe dream.
As for the crackpots, well, I know some people spend time and energy with them, but it is hard to believe their true objective is learning or contributing. It is, fortunately, very obvious when you meet one in the wild.
Yes, that's a failing of science. Reading the early volumes of Nature from the 19th century shows how much more of an open dialogue it was back then: https://www.nature.com/nature/volumes
Though education was much more limited, so take "open" with a grain of salt.
I've always been wondering if the crime matters at all. A lot of people say that going through a trial, even as a victim, is a punishment; it's the process that counts (hence "der prozess"). Perhaps the lack of a stated crime reflects that we the general public never fully understand under which specific provisions someone gets locked up. And so the actual deed may have been done before K. was born, etc.
This whole thing strikes me as coming from the wrong direction. Tying artistic and financial success, trying to apply some cargo cult "problem" engineering mentality to art. I feel like these articles illustrate quite well why the academic plastic arts have become so irrelevant today that we could say they are not part of human culture at large, in the sense that they have vanishing influence on public discourse.
Kafka (who apparently had a great sense of humor) seems to really enjoy writing people who die from too much second guessing. In the trial, K. keep failing by attempting to outplay the system at every step, because he thinks that he can stay above it all (don't we all). It's what you might call an awfully credible idiot plot.
"dequantization" is a thing and it's a very legitimate part of quantum information research. It's useful to probe if something was truly quantum or just smokes and mirrors, because it helps us understand where the boundary between quantum and classical lies. Another dequantized result from the past days: https://arxiv.org/abs/2604.21908
A very simple addition that makes casual browsing much more fun is to add a menu with adjacent articles, as is done in this reconstruction of Littré's 19th century french dictionary: https://www.littre.org/ (see mots voisins)
Why put comté and gruyère in two different categories? I just realized that in France the categorization of cheeses is closer to how they are prepared:
- fresh
- soft
- hard but not cooked
- hard and cooked
and it results in entirely different groupings. This will surely make some people unhappy.
probably because it justifies the sensationalized title, although the entire content can be summed up as "LLMs don't silo data, that's probably bad."
On the flip side, I thoroughly enjoy the fact that roleplaying in a videogame setting now counts as security research. Looking forward to the arxiv preprint "LLMs start playing really good drums if you pretend you're J.K. Simmons"