Why I linkage
-
Archaeologists claim intellectual property rights to ancient Roman dodecahedra (\(\mathbb{M}\), via), dozens of which exist in multiple museums, and block online sales of replicas. According to Wikipedia there are “more than fifty possible explanations” for what these things might have been used for.
-
Breakthrough prize award livestream censored (\(\mathbb{M}\), see also, see also) after presenter Seth Rogan chastised billionaire audience members who “underwrote electing a man who, in the last week, single-handedly destroyed all of American science”.
-
Transactions of the Bitter Mathematical Society (\(\mathbb{M}\)) cited in a publication (see reference 10), presumably after someone ran “Amer.” through English–French–English translation and Cambridge University Press’s nonexistent copyeditors failed to catch it.
-
Someone please convince me that the following text taken from a Wikipedia Featured Article is not just an elaborate joke (\(\mathbb{M}\)):
Let \(V\) be the set of all twice differentiable real functions \(f:\mathbb R\to\mathbb R\) that satisfy the ordinary differential equation \(f''(x)+f(x)=0\). Then \(V\) is a two-dimensional real vector space, with two parameters corresponding to a pair of initial conditions for the differential equation. For any \(t\in\mathbb R\), let \(e_t:V\to\mathbb R\) be the evaluation functional, which associates to each \(f\in V\) the value \(e_t(f)=f(t)\) of the function \(f\) at the real point \(t\). Then, for each \(t\), the kernel of \(e_t\) is a one-dimensional linear subspace of \(V\). Hence \(t\mapsto\ker e_t\) defines a function from \(\mathbb R\to\mathbb P(V)\) from the real line to the real projective line. This function is periodic, and the quantity \(\pi\) can be characterized as the period of this map. This is notable in that the constant \(\pi\), rather than 2\(\pi\), appears naturally in this context.
Spoiler rot13: nyy vg ernyyl zrnaf vf gung gur pynff bs shapgvbaf n fva k+o unir ebbgf jvgu crevbq cv
-
Random regular graphs are neither almost always nor almost never optimal expanders (Ramanujan graphs), as one might expect and Alon and Sarnak famously bet on (\(\mathbb{M}\)). Instead, they are optimal expanders roughly 69% of the time. The result follows from the universality of the eigenvalue distribution for random matrices. This is in a recent preprint “Ramanujan property and edge universality of random regular graphs” by Jiaoyang Huang, Theo McKenzie, and Horng-Tzer Yau, arXiv:2412.20263, via a Quanta article which discusses the bet and the background for the problem.
-
A piecewise-linear isometrically immersed flat Klein bottle in Euclidean 3-space (\(\mathbb{M}\)). That is, a Klein bottle that (if only it didn’t self-intersect) could be folded from a flat sheet of paper. New arXiv preprint by Stepan Paul.
-
New preprint announcing a solutino to the \(X+Y\) sorting problem (\(\mathbb{M}\)) unfortunately turns out to be wrong, but at least posting it turned up the error. This is the problem of sorting the sums of all pairs of elements of two sets \(X\) and \(Y\) of size \(n\). It has long been known that you can do this with \(O(n^2)\) comparisons, but it has been an open problem since before 1975 whether you can do it with a comparison-based algorithm that runs in \(O(n^2)\) time, because the known solutions spend too much time figuring out which comparisons to try.
-
Trump administration compiling lists of Jewish university faculty and students (\(\mathbb{M}\), archived, via).
-
Wikipedia’s nonprofit status questioned by Trump-appointed US attorney, accusing it of “allowing foreign actors to manipulate information” (\(\mathbb{M}\), archived, via). Well, yes. If you define Wikipedia editing as manipulating information, then doing so is not limited to US-based Trump-loyalist propagandists. That’s a good thing. This should be a warning, though, that Wikipedia’s ownership and internet hosting being based in the US has become a liability rather than a strength. This is part of a recent but long line of US right attacks on Wikipedia: see also The Verge, “Trump DOJ goon threatens Wikipedia”, The Guardian, “Why is Elon Musk attacking Wikipedia? Because its very existence offends him”, Le Monde, “Why Elon Musk is calling for a boycott of Wikipedia”, The New Yorker, “Elon Musk Also Has a Problem with Wikipedia: Lately, Musk’s beef has merged with a general conviction on the right that the site is biased against conservatives”, etc.
-
Video of many new polyhedral realizations of higher-genus regular maps (\(\mathbb{M}\)) based on CodeParade’s newly-published paper “Polyhedral embeddings of triangular regular maps of genus \(g\), \(2\le g\le 14\), and neighborly spatial polyhedra”, MDPI Symmetry, with Jürgen Bokowski. Here “regular” means there are topological but not geometric symmetries between each two (vertex,edge,face) triples (flags).
-
On using social media as notes to oneself and the difficulty of searching those notes in Mastodon. This is one reason I post these linkage roundups: then I can and do grep the source files for my blog.
-
Reversing the fossilization of computer science conferences (\(\mathbb{M}\)). Bertrand Meyer in CACM on the problems caused by careerism in academia and the use of CS conferences as “a yearly exam and résumé building exercise” by their old-timers, leading to gatekeeping-motivated submission rules and conferences filled with papers reprising the greatest hits of the past, rather than fostering originality and creativity. He is optimistic that some minor changes of mindset can fix the problem but I’m not convinced. I think there’s a strong relation between the fossilization of existing conferences and the creation of new conferences, intended to be more innovative, that eventually become fossilized in the same way and need to be replaced by new conferences in the same way.
-
After being forced to listen to AI-generated audio voice over repetitive slides for over two hours for my state-mandated sexual harassment training (\(\mathbb{M}\)), a video by phoneticist Geoff Lindsey comes to mind: “What is that YouTube intonation?”. It’s about speech that gets the pronunciation of individual words and short word phrases correct (if a little precise and stilted) but that fails to properly emphasize or de-emphasize words and pauses to follow the higher-level structure of whatever it is saying. Once you start noticing it, it can become a very obvious tell of AI generation, but Lindsey argues that some human YouTubers are starting to follow the same patterns.
Normally I would just read the provided scripts and skip the voiceover, but that would take too little time; the state mandate is for at least two hours, not for coverage of the material. As one saving grace, my university’s new mandate of privacy-invading security software has led me to buy a cheap macbook as a burner machine to install the security software onto, freeing my other desktop machine for multitasking while I do things like this.
-
University of Waterloo withholds prestigious coding competition results over suspected AI cheating (\(\mathbb{M}\), archived, via, via2). In the comments, we hear that coding competitions at MIT are facing similar problems.