Dan Piponi

Disclaimer:
👽 My opinions are not my own. They're beamed to me by aliens

Current life:
⚖️ Considering new employment

Previous lives:
🍩 PhD in Riemann theta functions
🧬 Glaxo Group Research
🚀 Argonaut Games
💥 Mass Illusion, Esc, ILM
🎈 Google X, Verily, Google
🎮 Epic Games

Likes:
🚴 I like to bike
🏃 I like to run
🎛️ my musical tastes lie towards the electronic end of the spectrum
🚀 I like Andor *and* The Mandalorian
📈 I have an OEIS entry https://oeis.org/A387361

Public Key

npub1xs30e0pj7velktf5sxewnqf93tu2pdt3s6wtl6fug2tzgy8zxthsgfgrae

NIP-05 Address

dpiponi@mathstodon-xyz.mostr.pub

Profile Code

nprofile1qqsrgghuhse0xvlm956grvhfsyjc479qk4ccd89lay7y993yzr3r9mcpz4mhxue69uhhyetvv9ujuerfw36x7tnsw43q3fvkaz

Publishing to

relay.ditto.pub

Author Public Key

npub1xs30e0pj7velktf5sxewnqf93tu2pdt3s6wtl6fug2tzgy8zxthsgfgrae

Show more details

Last Notes

2026-04-07 17:47:01 UTC

If I owned a folding phone I think it'd be like living with the Sword of Damocles. Every open or close might be the one that slices the display in half. Even if you told me it could withstand 500,000 folds I'd still feel like that. Irrational, I know.

2026-02-21 22:53:40 UTC

I have no idea what this is.
I think ad recommenders must periodically suffer from some kind of model collapse. I'm being stalked by ads for things like operating theatre gear, hyperbaric chambers and devices like the attached.
I'm not gonna complain too much. It's much more interesting than it used to be and hope it doesn't revert to video poker too soon.
https://media.mathstodon.xyz/media_attachments/files/116/111/061/719/631/591/original/ad9582f7e6d41877.jpeg

2026-02-08 19:33:07 UTC

I didn't watch Caprica when it first appeared (well, I managed to get half way through the first episode) but it's a new landscape now and in some ways that makes it more interesting. It's basically about the risks of training an AI on a broad corpus and then hoping that with a bit of prompting you can specialize it for specific tasks.
Creepy thing is I've never seen Waymo ads before but I keep getting them in this show (streaming on Peacock) and they start "The robots, they're coming..." I'm just hoping it's a coincidence and you're all seeing the same ad.
https://media.mathstodon.xyz/media_attachments/files/116/036/673/287/613/562/original/9fa1c02d90bc537c.jpeg

2026-02-06 19:11:35 UTC

- reply

@nprofile…7jv6 Never in living memory has any government in the UK spewed out the constant stream of insanity we get from the White House. So yes, 100% serious.

2026-02-06 18:22:42 UTC

- reply

@nprofile…xuzd @nprofile…jxjq I think the likelihood of Farage in No. 10 is pretty high, but it'll still be a long long way from the unremitting insanity over here.

2026-02-06 17:55:00 UTC

- reply

@nprofile…xuzd Farage is a disgusting scumbag which barely makes him about 1 milliTrump.

2026-02-06 17:51:43 UTC

I hate when people talk about shared US and UK values. The British people would never tolerate garbage in Parliament the way Americans tolerate it in the White House.

2026-02-03 17:36:19 UTC

- reply

@nprofile…fh5j It must happen from time to time.
When I think back on my physics education you often had functions with singularities, but they were still meromorphic so if you map to a Riemann sphere everything is linear again, even at infinity.

2026-02-03 17:17:00 UTC

If you need to study some functions but have limited tools you can go a long way assuming everything is linear except at minima/maxima where you assume everything is quadratic.
Use this tool in physics and everything looks like a free field or a perturbation thereof.
Use this in statistics and ML and everything looks like a Gaussian distribution.
So when you see what seems like a deep connection between these domains I wonder if what you're seeing is just the shape of your tools.

2026-01-30 00:55:25 UTC

- reply

@nprofile…qa6s The Kobayashi Maru approach

2026-01-28 00:25:14 UTC

- reply

@nprofile…gu9p @nprofile…ck2y I personally take the view that if it's essential for survival we can't really ask why it's there without slipping into religion. So I'm not optimistic about discussion about why there is enough order out there for us to make a living.
But, one could ask how much more order there is than the bare minimum. Our ancestors didn't need general relativity for to climb down from the proverbial trees, nonetheless we do have it now. Should I be surprised that we found this extra regularity that wasn't required for survival? I don't know how to measure how surprised I should be. I certainly wouldn't expect that the only physics we have is precisely the bare minimum for survival - survival just gives the lower bound. But maybe it is surprising we're able to make accurate testable predictions about, say, black hole collisions.
But anyway, none of this changes my belief that pure mathematics was built on empirical principles much like physics.

2026-01-27 22:37:38 UTC

- reply

@nprofile…gu9p Mathematicians also add axioms as needed. ZFC didn't appear by magic ex nihilo. It's the result of a centuries long process of trying out different approaches until something was good enough to do most of what was needed - and the thing that was needed was very often motivated by physics even though mathematicians sometimes deny it.

2026-01-26 00:52:50 UTC

41% of people approve. I'll do my best to avoid those 41%.
https://media.mathstodon.xyz/media_attachments/files/115/958/659/885/432/786/original/c621f2868169a981.png

2026-01-24 22:43:01 UTC

Reminder: If you're in the Stanford Prison Experiment with two others holding someone down spraying pepper spray in their eyes it's probably a good moment to call for time out.

2026-01-22 17:48:45 UTC

In algorithms you can trade space for randomness. I don't know if there are some precise generic theorems about this as opposed to theorems about specific families of algorithms. But that's not why I'm here...
There's a similar trade you can make in games, TTRPGs in particular.
For example, you could go old school and track how long everyone's torches have been alight for as they explore their dungeon. But it's tedious to track that state.
There's another extreme where you can say a torch has a fixed probability of going out each turn. But that's unrealistic in a way that's not fun. You don't get the anticipation, or dread, of your torches slowly burning down.
Today I learnt about another approach to torches that uses a middle ground. You roll D12 every so often to see if the torch burns down. When you roll a 1 you now switch to rolling D10. When you roll a 1 you switch to D8 and so on. When you roll a 1 on D4 (say) the torch goes out. There are far fewer state updates. You can even store that state as a choice of die sitting in front of you.
There's probably also a connection here to DSP with the fixed probability case acting like a one pole filter.

2026-01-22 02:33:59 UTC

Is this answer correct:
https://tex.stackexchange.com/questions/32853/period-after-math-mode-equation
A "." inside an equation is surely not treated as punctuation although I've been doing that.
What is the correct way?
https://media.mathstodon.xyz/media_attachments/files/115/936/410/434/267/659/original/2a095c67d3ad8c55.png

2026-01-20 15:11:29 UTC

Random find of the day:
https://math.uchicago.edu/~may/REU2023/REUPapers/Soni.pdf
Not just about Brownian motion and complex analysis but "A course in Complex Analysis via Brownian Motion". Eg. Brownian motion in the complex plane visits every neighborhood infinitely often which allows us to deduce the fundamental theorem of algebra.

2026-01-19 16:07:40 UTC

- reply

@nprofile…chx4 You got me worried I'd made a mistake. It's minus n.

2026-01-18 01:06:56 UTC

Every stereotype I ever had about Republicans has turned out to be true. Apart from the ones that require DNA testing to confirm.

2026-01-16 21:21:03 UTC

- reply

@nprofile…t6k2 I was going to say something along those lines too. With hindsight history repeats itself because you get to cherry pick the pairs of bits you want to claim are isomorphic.

2026-01-16 20:39:01 UTC

If history really repeated itself then life would be a lot more predictable than it is.

2026-01-14 20:40:50 UTC

- reply

@nprofile…xhtk Americanism. Also note "savings" used as a singular noun which I think is also American.

2026-01-14 16:53:05 UTC

- reply

@nprofile…sjey @nprofile…5wht That was great! I think Carpenter's contribution to popular music is still underrated. His music became part of our shared background culture and so his name isn't always explicitly acknowledged.

2026-01-14 03:07:47 UTC

Driving home, I came to the end of an album clearly inspired by John Carpenter. Wanting more I started to ask Siri to play the soundtrack to Escape from New York. But I looked at the car in front and it had another soundtrack suggestion.

2026-01-12 00:13:25 UTC

One major event that took place in the history of D&D (or at least in D&D-adjacent history) took place in the steam tunnels of the University of Michigan.
https://en.wikipedia.org/wiki/James_Dallas_Egbert_III
Today is the day when I finally looked up "steam tunnel". I think this might not be a thing in the UK, or at least a thing known by this name.

2026-01-10 21:22:35 UTC

- reply

As I've mentioned before, I find the game Universal Paperclips fascinating:
https://universalpaperclips.fandom.com/wiki/Universal_Paperclips_Wiki
I think a stated goal of the creator was to construct a minimal addictive game. And one of the things that makes it addictive is (almost) monotonically increasing numbers. People love the feeling of making progress and love to see numbers indicating this.
I think this was a killer feature of D&D: your experience points, your level, your hit dice, the level of spell you have access to, the number of spells you have, your armour class, your wealth. This was lacking from other role-playing-ish games and was lacking from almost all of the history of wargames. Some wargames allowed for a tiny amount of improvement - like morale could increase for a unit after a successful battle and some units might be promoted to veterans.
But D&D offered long term fine-grained progress. Even if you didn't just go up a level you could see your XP go up after each battle inching you closer. It's exciting.
I think it's something that's easy to take for granted. The idea of role-playing was itself such an exciting one that you can miss that visible measurable progress was important.
I doesn't mean everyone plays for numbers today. I just think this feature was something that helped get D&D the critical mass it needed because it appealed to enough people.
And I have to admit, I see it in myself when playing retro JPRGs on my Miyoo Mini. At any moment in time one of the characters in your party seems close to the next level so you always feel like you need to fight one more battle before you stop for today...

2026-01-10 21:14:20 UTC

I finally finished Peterson's 2 volume "Playing at the World"
https://mitpress.mit.edu/9780262552318/playing-at-the-world-2e-volume-2/
(It was a challenge because each footnote would suggest other books I ended up reading...)
When I started reading about the creation of D&D I assumed that the key idea was the invention of the RPG and once invented it would just spread like wildfire. This is only partly true because Peterson talks a lot about role-playing-ish games that existed at about the same time, in the US, UK and Germany, but which didn't catch on: Western Gunfight, Hyboria, Armageddon, Midgard I & II as well as adjacent things like SCA and Coventry.
There are a couple of reasons why D&D took off. One was probably that it was fantasy based and there was a receptive audience primed by a first wave of Tolkien-mania.
But there's another one that I want to rephrase differently from Peterson...

2026-01-09 19:36:47 UTC

I'm still sure I was born on a timeline where C++ had no 'and' or 'or' keywords and that there's been a retcon.

2026-01-07 16:49:59 UTC

- reply

@nprofile…fh5j Typically people doing this stuff are working in some space of parameters for a family of distributions where Riemannian metrics make sense. But I can imagine some people don't...

2026-01-07 16:18:36 UTC

In Information geometry you study geodesics in spaces of probability distributions - paths with the property that samples drawn at time t+dt could most plausibly have come from the distribution at time t. It's what you use when you want to boil proverbial frogs without them noticing, or transition your population into Nazis.
https://en.wikipedia.org/wiki/Information_geometry

2026-01-06 17:57:21 UTC

- reply

BTW Those books marked me for life. I found material like this fascinating as a kid.
https://media.mathstodon.xyz/media_attachments/files/115/849/434/655/413/446/original/62e174c9b2759cae.png

2026-01-06 17:39:48 UTC

- reply

A curious thing. It seemed obvious to me that some of the shimmering effects are caused by saccades. Stare at something for a bit and the signal from the receptors will slowly get suppressed so that in effect the negative of the image is being superimposed on the image. Eg. you see the complementary image if you stare at an image and then look at a blank surface.
If your gaze shifts within an image then you expect to see the sum of the real image and a shifted negated version. For "high frequency" images like this one you'd predict that this would cause the shimmering we do in fact see.
Anyway, there is actually a paper about this particular piece that states cetegorically the effect isn't caused by saccades:
https://www.researchgate.net/publication/225797505_Saccades_in_Neal's_Square_o_f_Three

2026-01-06 17:27:40 UTC

When I was a kid by parents gave me a set of Childcraft books of which one was a book on art:
https://archive.org/details/lookagainchildcr0000chil
One of my favourite works in it was "Square of Three" by American artist Reginald Neal:
https://chrysler.emuseum.com/objects/22126/square-of-three-black-and-white
So I thought I'd write the shader. It's fun to stare at random parts for extended periods of time...
https://www.shadertoy.com/view/tcKfWG

2026-01-04 01:12:34 UTC

- reply

@nprofile…vhdw Ah, Humphreys, I used to own a physical copy of that book. I managed to read most of it which probably means I thought it was interesting. Can't remember much about it now except maybe I got some intuition about Dynkin diagrams from it.

2026-01-03 02:51:36 UTC

- reply

For example when playing Horizon Zero Dawn, Aloy almost feels like part of me, but Mario feels like some other character I'm trying to push around, always stepping too far and never quite in the right direction.

2026-01-03 02:20:47 UTC

Revisiting old video games can be a disappointment. But Super Mario 64 and Goldeneye are as good as I remembered.
But the controls and camera motion are as annoying as I remember - though still playable. Game designers have learnt a lot over the years.

2025-12-21 02:51:43 UTC

As we approach the New Year it's becoming clear that by the end of the year the number 17 will take the highest value it has ever achieved.

2025-12-20 18:05:04 UTC

One of the more beautiful books I've ever seen on sale.
I read the ebook recently and really couldn't see the appeal. Maybe it needed to have more physical presence.
https://shop.suntup.press/products/the-hellbound-heart-by-clive-barker-lettered-edition?utm_source=Suntup+Editions&utm_campaign=52efaad962-SCANNER_PREORDER_COPY_01&utm_medium=email&utm_term=0_447f8dd321-52efaad962-509116548&mc_cid=52efaad962&mc_eid=22e82ff232

2025-12-18 05:41:46 UTC

- reply

Seems successful. These old keypads are a bit finicky but it seems ok.
https://media.mathstodon.xyz/media_attachments/files/115/738/964/933/421/524/original/e4dd2eeaa76c5ff8.jpeg

2025-12-18 02:41:54 UTC

Contemplating merging calculators. One looks better than the other, and has a more reliable keypad, but has a faulty segment in the display. It means removing the keypads, which I've done once, but it's a bit delicate and I'm nervous of doing it again. I would like a near perfect Sinclair Cambridge Programmable.
https://media.mathstodon.xyz/media_attachments/files/115/738/244/906/627/315/original/a005eaefe31d93bc.jpeg

2025-12-16 19:50:24 UTC

- reply

It's still in Malacatán although it has moved a little. It's by this establishment that has corpses hanging...no wait...that says pinateria...that's a relief.
https://media.mathstodon.xyz/media_attachments/files/115/730/968/907/376/905/original/1e19693bfd888e23.jpeg

2025-12-15 20:14:33 UTC

I don't know about in the US but I don't recall the use of em and en dash being taught at school in the UK. We had lessons on the use of comma and semicolon and quotation marks and so on. Dashes seemed like part of the world of literature and journalism, places where typography matters, but not something ordinary people use. Surely this is why dashes are interpreted as signifiers for AI.
I started using dashes about 15 years ago when poorly formatted ebooks made them more visible. Iain Banks is hard to read when every dash becomes a hyphen. The fact that they were (and still are) an issue for ebooks suggests that even people who work in publishing are confused about them.

2025-12-15 18:22:18 UTC

- reply

@nprofile…lg9y I just want to say I value your reply more than just giving it a star.

2025-12-15 17:27:58 UTC

Browsing the Homeland Security account on X. The US is a vile country. I'm ashamed to live here.

2025-12-14 22:36:24 UTC

- reply

The corrected version was mailed out from Denmark and it's now installed. The "plaque" isn't actually visible from the outside so some disassembly was required.

2025-12-14 18:18:18 UTC

- reply

@nprofile…w448 Sometimes you get excited about an intellectual idea and see it everywhere, maybe when it's not really applicable. We've all done this, right? It's often just a passing phase. But when someone's in one of these phases there are some things they shouldn't work on. Like nobody in a fractal phase should be working on navigation software.

2025-12-14 00:35:33 UTC

I was about to start watching one of Villenueve's older films, Enemy, but I tried the trailer and it's the same yellow of the deserts of Dune and the same yellow of the dust clouds of Blade Runner++ and I didn't think I could face 2 more hours of movie jaundice.

2025-12-13 14:50:05 UTC

- reply

Past the volcanoes. Pacific Ocean here we come!
https://media.mathstodon.xyz/media_attachments/files/115/712/811/354/473/654/original/43d9c6bf8b265d7d.jpeg

2025-12-13 04:46:01 UTC

- reply

Now I'm surprised but it's popped up in a town called Totonicapán. It has travelled most of the way across Guatemala now. I wonder if it ends up at the Pacific coast.
https://media.mathstodon.xyz/media_attachments/files/115/710/431/672/437/343/original/2a6854d039225c12.jpeg

2025-12-12 20:33:07 UTC

- reply

@nprofile…9s72 Some bonus content for our fans! According to Google Street View this is the final place I got a ping from.
https://media.mathstodon.xyz/media_attachments/files/115/708/482/232/269/505/original/9c20c9a394d0d5b6.png

2025-12-12 01:40:41 UTC

- reply

No updates for 48 hours despite being in what seems like a well trafficked area. So I'm thinking Taggy (it has a name now) may have met its demise.
#AirTag
https://media.mathstodon.xyz/media_attachments/files/115/704/040/132/651/285/original/f50e1a6676ea8441.jpeg

2025-12-11 04:33:05 UTC

I printed something! That might not sound like much but judging by the documents the printer spat out, the last time I lost something in the print buffer was back in the 2010s. I did try to print something else, but like the dragons of Melniboné it's done its stuff and it's now back in "Deep Sleep", probably until some time in the 2030s.

2025-12-09 22:37:10 UTC

- reply

It spent a week opposite a toy store in Guatemala City but it's back on the move again, passing near Hobbitenango (the looking up of which is optional). #airtag
https://media.mathstodon.xyz/media_attachments/files/115/691/989/020/083/255/original/586b3c99523a5fbe.jpeg

2025-12-09 21:09:33 UTC

- reply

The role of mathematics here is a curious one too. It reminds me a bit of how mathematics is used in shader writing. Functions like sin and dot product and floor and relu are parts in a toolkit that can be thrown together creatively so see how they look and it doesn't matter that they're not being used the way they're "supposed to be" used as long as the result looks good.
In machine learning there may be an original theoretical foundation for why a method should be used but then it gets tweaked beyond recognition and no amount of a priori reasoning could have taken you to that point. We only know it's a good method because of empirical results. Which then makes me worry about what purpose the original theoretical foundation served.

2025-12-09 21:00:57 UTC

Modern ML research reminds me a lot of particle physics research though this is partly bias due to have only read one book on sociology of science. There is a lot of fitting of interpretation to empirical results (not to be confused with ML itself which is also the fitting of models) which is very different to other branches of CS.
It's interesting how many methods are used as a result of social forces as much as empirical results. Methods come in and out of vogue without much impact on quality. I've read a few papers doing ablation tests and there is a common pattern that much of common practice can be replaced with alternatives or even thrown out completely. I'm never sure how important things like dropout or batch norm or choice of activation function really are.
As a colleague just pointed out to me, some of practice may be shaped simply by the fact that researchers are in a hurry to get papers out and need to have a paragraph or two on interpretation of results. The paper gets accepted because the methods work but now the interpretation work is also published, a bit like a rider bill, even though it might not be justified.

2025-12-09 20:50:03 UTC

- reply

@nprofile…9s72 This has been my conclusion about most detectives!

2025-12-09 16:10:19 UTC

Apparently the hardest part of writing detective fiction is contriving a way for a detective to be present at a murder scene. Which is why so much detective fiction has implausible coincidences to frame the story: detectives who just happen to be invited to the country manor when a murder happens, detectives what happen to have a relative in trouble, detectives who just happen to be visiting the village fair as someone as shot, and so on.

2025-12-09 15:52:06 UTC

- reply

@nprofile…9s72 @nprofile…4x5t I enjoyed the ideas but not so much the text itself.
Tangential (though you have to read the book to know why): back in the old days before Arduinos I built a vampire robot using a 68000 micro processor and a hacked Nintendo Game Boy camera. It was a vampire because looking for the angles of a cross and running away was the easiest behaviour I could think of to code up.

2025-12-09 03:27:50 UTC

Is there any genuinely scary horror fiction? There's horror fiction that'll make me squirm for all kinds of reasons, but fear isn't one of them, unless it's fear that over the page I'll read something truly gross. But that's not what I think of as scary.
I don't think I've ever found written fiction creepy scary [1]. I might enjoy a creepy atmosphere, say, which is why I like Lovecraft, but I've never found his work even slightly scary.
[1] One exception: I was reading an old MR James (I think) story and something happened that's completely non-scary, except it was an almost trivial something from a recurring dream I'd had as a kid, and seeing it in writing did creep me out l, and I had to put the book down. But if I told you what it was you'd have no idea why I thought it was scary. (Like the terrifying shaky video scene in Signs that has no reason to be even slightly scary - but it didn't work for everyone.)

2025-12-09 03:05:10 UTC

It started with two power cuts and candlelight and running the fridge off a battery. (LED light too, I bought some great lamps after the previous power cut.)
And then we lost running water and I was filling up the largest containers I could carry from a friend's hose so we could flush the toilet.
But then winter arrived and our heating failed. Yes, it's a California winter but indoors it's still is a little chilly for sedentary activities.
So basically we've been camping indoors for the last several weeks. All an adventure. And while I was looking forward to testing out my new winter gear ahead of visiting New York for Xmas (can't wait!) I didn't expect to be breaking it in indoors.

2025-12-06 03:10:33 UTC

Good news: I found a ST:TNG episode I've never watched before. I didn't think I would ever again in my life have the experience of watching a fresh episode.
Bad news: Oh why oh why oh why is it a Wesley episode?

2025-12-02 18:36:59 UTC

- reply

It's moving around Guatemala City now. Apple Maps claims it's 29,905 miles away which is a reminder of why I think it's OK to ask Fermi questions in job interviews.
https://media.mathstodon.xyz/media_attachments/files/115/651/410/686/396/887/original/e3aafa4be0036a27.jpeg

2025-12-01 23:20:11 UTC

- reply

@nprofile…8eua It sounds more like something you might get at a bar, I like that kind of shot.

2025-12-01 22:56:06 UTC

- reply

Now it's in a warehouse in Mixco, a city next door to Guatemala City. I'm hoping it's sitting loose in the container and eventually ends up on another international trip.
I just remembered: this AirTag was likely originally in my carry-on baggage on my flight from London. When I went to collect it, it was no longer where I had placed it. I had to wait for everyone else to deplane so I could go searching for it and someone had moved it 20-30 rows forward in business class. I never got an explanation for that from the cabin crew.

2025-11-30 20:29:43 UTC

My lost AirTag sailed over the Gulf of Mexico and now it's in Guatemala. I wonder if it's about to take a road trip.
https://media.mathstodon.xyz/media_attachments/files/115/640/533/264/295/668/original/da4a6a2e8585ed02.jpeg

2025-11-29 19:13:59 UTC

If you're not seeing the problem, look closely.
Very embarrassing. They still haven't fixed that "chartered" error on the web site.
https://media.mathstodon.xyz/media_attachments/files/115/634/574/114/465/599/original/c33233ec7c782ddd.jpeg

2025-11-28 02:45:11 UTC

- reply

@nprofile…8v8d Coincidence but I just read the first reference I've ever seen to half elf-half non-human
https://media.mathstodon.xyz/media_attachments/files/115/625/027/569/491/285/original/512520ebbeccd812.jpeg

2025-11-28 02:25:29 UTC

- reply

@nprofile…8v8d I think that point was made even in one of the Dragonlance novels. Blame Lord Dunsany.

2025-11-27 22:38:20 UTC

- reply

@nprofile…4x5t @nprofile…8eua Entropy of mixing is confusing. If entropy is based on probabilities and probabilities are subjective, what does it mean when you think something has high entropy because it was mixed but then you learn that you can undo it?
(I'm sure there's a concrete answer here although I remember reading about physicists confused about MRI that can mix spin states and then undo them, confounding naive definitions of entropy.)

2025-11-27 21:31:22 UTC

- reply

@nprofile…8eua Also, while on the confectionery theme, when I was an undergrad our algebraic topology lecturer was Lickorish (who was one of those rare lecturers students look forward to) who gave his name to the Lickorish twist, a theorem about Dehn twists, and I wonder if what the the taffy machine is doing can be viewed as (some kind of slice through?) a Dehn twist.

2025-11-27 18:03:38 UTC

Not having grown up in the US I'm not 100% sure what does and doesn't count as "taffy" but it's a great subject for a paper. Connects to the topic of rational tangles that I touched on at my ICFP talk years ago.
https://arxiv.org/abs/2511.19461
https://media.mathstodon.xyz/media_attachments/files/115/622/972/891/326/648/original/0f7b256dcb4ad940.jpeg

2025-11-25 19:11:57 UTC

Ooops...on lego.com
https://media.mathstodon.xyz/media_attachments/files/115/611/904/617/965/000/original/2105c1ebe016a5c2.png

2025-11-25 14:55:36 UTC

- reply

As I was saying. No major media outlet seems interested in saying even the blindingly obvious.
https://www.theguardian.com/media/2025/nov/25/rutger-bregman-accuses-bbc-of-censoring-his-reith-lecture-on-trump

2025-11-21 04:41:01 UTC

While we're at it can the Moron-in-Chief sue all the other media for editing his ravings to hide the fact that the President of the US is dumber than a dead salmon?

2025-11-20 22:38:16 UTC

- reply

@nprofile…w8t5 Yes, one of the principal exports of the US is propaganda and the UK has always liked to import it.

2025-11-20 22:18:03 UTC

The US is the laughing stock of the world
https://www.cdc.gov/vaccine-safety/about/autism.html

2025-11-19 00:16:59 UTC

- reply

@nprofile…ztfm @nprofile…dand I'm a bit confused. In the finite game, seems to me Sam doesn't have to do anything strategically. Joe just guesses 0, 1, 2, ... and will eventually guess correctly in a finite number of guesses. But if they need to bound the number of guesses in advance then Sam needs to do something.

2025-11-18 16:44:12 UTC

I'm sure that it was within the last couple of years that I mentioned something here about a 1950s short story that took for granted that computers were infallible. And honestly, two or three years ago computers were still fairly reliable.
How low we've fallen in just a couple of years of AI going mainstream!

2025-11-17 19:25:51 UTC

- reply

@nprofile…ddsj I find myself doing similar things in word processors. Like in Google docs having to insert text inside some text, rather than next to it, so as to inherit its font, and then having to fix the mess afterwards - but still faster than the usual clunky way to select font.

2025-11-17 05:40:33 UTC

I mentioned recently I wish more software obeyed laws. Included in that is the wish that edge cases are handled consistently with correct handling of zero or one.
Example: I've been using Helix as my editor recently. It has a lot to commend it. But it handles edge cases badly. Two examples: you use 'x' to select a line but if the line is empty it selects the next line. You use 'w' to select a word but if it has 1 character it selects the next word too. This makes it hard to write macros - not that it supports proper macros yet. And it's really bad for multiball cursors because if some are on 1 character words and some are on 2 character words the behaviour varies.
More laws in software please! OK I admit there is a bizarre law in this case: the smallest possible instance of something doesn't count and selecting it selects the next thing too.

2025-11-16 22:48:03 UTC

- reply

@nprofile…9s72 And how long before they reached the...er...stage of eating things you're not supposed to eat?

2025-11-16 20:52:10 UTC

- reply

@nprofile…c2dj True, but there's an outlet at the base of the tank that can be opened with a screwdriver (for completely draining the tank) but there was only enough in there to fill a bucket to 3 or 4 inches.

2025-11-16 20:43:49 UTC

- reply

BTW I've been assuming that in the event of the apocalypse, and the water supply being cut off, we'd still have a hot water tank full of water. But it's an older house, and we hadn't got around to installing a backflow valve (it was on our list), and having a break in the water supply meant the contents of the tank was siphoned out of the house and onto our driveway.
So if you're prepping for the apocalypse make sure you have a suitable valve.

2025-11-16 16:06:39 UTC

- reply

I gotta say, it's challenging living without running water. At least I have access to showers at work. But curse the people who thought it was OK to pipe water using steel pipes buried under concrete.

2025-11-14 16:26:08 UTC

Got home to be greeted by darkness because of a power outage. Lucky power came back when it did because it enabled the Flume to start working and tell us about the massive water leak between the meter and the house. Seems we have our own mini black hole because none of the hundreds of wasted gallons is visible anywhere. Digging into the earthquake kit for water supplies...
Anyway, living the mini-post-apocalypse right now.

2025-11-13 20:00:04 UTC

- reply

@nprofile…578r I was already tempted to read some Warhammer stuff by Brian Craig because that's a pseudonynm for Brian Stableford whose work (and translations) I have enjoyed.

2025-11-13 19:49:00 UTC

- reply

@nprofile…578r I may have to read that if only because it was apparently published by Games Workshop and I've been on the lookout for their early publications.

2025-11-13 19:23:53 UTC

I spent my life waiting for the future to arrive and I think it finally has.

2025-11-12 15:33:08 UTC

À propos of nothing I just want to say how much I love this cover of Dragon magazine from 1982. Far from the usual sword and sorcery material it almost feels like a photograph from the world of D&D. I can't find any other work by the artist, Dean Morrissey, that I like.
https://media.mathstodon.xyz/media_attachments/files/115/537/437/359/053/757/original/caec9a12d3cf1edc.jpeg

2025-11-09 23:22:37 UTC

- reply

@nprofile…25az I'm just following in your footsteps. It was indeed the plastic mechanism that failed. I have to dig with my fingernails to open it now.

2025-11-09 23:02:55 UTC

- reply

@nprofile…25az Our previous kettle looks exactly like the Cuisinart but is made by Hamilton Beach. I bets it's the same parts! I've just resurrected it from the trash because it's better than the Oxo - well work around the failed opening mechanism for now.

2025-11-09 22:09:07 UTC

Living in a country where if you ask for a cup of tea in a restaurant they'll bring you a cup of lukewarm water and expect you to dunk a teabag in it I shouldn't be too surprised that I can't buy a decent electric kettle. This time I decided to get an expensive one rather than a cheap one that takes forever to boil and falls apart after a year. Well it seems well made, but produces hot water that tastes of plastic and silicone sealant. Why is this so hard?
(Yes, I could go old school and use a clay pot on open fire like my ancestors did but I'd like modern and convenient.)

2025-11-09 18:28:04 UTC

- reply

@nprofile…a9u6 Having said that, there may be some clever trick that overcomes this issue.

2025-11-09 18:27:21 UTC

- reply

@nprofile…a9u6 As mentioned in another answer, it's hard to see the discontinuities in the derivative that reveal the eigenvalues - even when you have billions of samples.

2025-11-08 05:17:36 UTC

I can't believe I didn't know this fact before today - it seems so fundamental.
If you have a real symmetric matrix A, and pick some unit vector x, then x.Ax (the Rayleigh quotient) is bounded by the smallest and largest eigenvalues. So what happens if I pick a random unit vector, how is x.Ax distributed?
Let's jump to the case of a Hermitian matrix A and a random unit vector on the complex sphere. How is T=conj(x).A.x distributed? Again it's standard that it's bounded between the smallest and largest eigenvalue.
We can start with a 2x2 matrix A. First surprise to me: T is distributed *uniformly* between the two eigenvalues.
The first plot below is the histogram of values of T for 1,000,000 vectors x and a randomly chosen Hermitian matrix A. The yellow lines mark the two eigenvalues.
Now try n=3. Remarkably the distribution is piecewise linear with the break points at the eigenvalues.
And let's jump to 5. We get the third plot, whose curve is given by a cubic B-spline with control points given by the eigenvalues.
Bizarrely you can prove you always get B-splines making this one of the few times I've seen B-splines in "nature", so to speak.
This is a purely mathematical result but I found it in this paper on wave transport through random media: https://arxiv.org/abs/2511.04602
My calculations (pretty much the same as the paper) are here: https://colab.research.google.com/drive/1g7ImBblYAsninp5-sr6uWQTVfTYurerE?usp=sharing
https://media.mathstodon.xyz/media_attachments/files/115/512/340/864/407/810/original/908f0decce0300a1.png
https://media.mathstodon.xyz/media_attachments/files/115/512/342/608/191/011/original/1b5efb64fdf6c8b6.png
https://media.mathstodon.xyz/media_attachments/files/115/512/346/950/504/579/original/51bb660ccfdf828f.png

2025-11-07 21:00:14 UTC

- reply

@nprofile…lg9y @nprofile…atf6 Probably not a lot if your matrices are large.
The reason I was thinking about this subject is that I was playing with a scenario where it does matter: I wanted a concurrent (single producer/many consumer) expandable array with stable elements so instead of copying to the newly allocated block you keep all the old blocks as pages. But then you need the page/offset computation to be cheap.

2025-11-07 18:01:07 UTC

- reply

@nprofile…lg9y @nprofile…atf6 And personality type :) I've a tendency to want to be a bit more careful with memory usage even when it's clear I have the resources to be a bit more profligate.

2025-11-07 17:45:50 UTC

- reply

@nprofile…atf6 @nprofile…lg9y I feel like a 30% or so "tax" on memory allocations is a bit steep, especially in applications with heavy usage of this kind of array. So just by feel I'd use something around the golden ratio for A. There's an argument that precisely this value is optimal though I don't think it applies for most use cases.

2025-11-07 16:43:14 UTC

A common strategy when implementing arrays that grow at the end is to allocate space for N bytes initially and then when more space is required resize to a block of size AN for some A, typically in the range (1, 2], and then keep reallocating as needed scaling up by A each time.
Let's say A=2. How much memory do we expect to actually occupy?
A good guess might be to say that typically at any moment in time we expect to be half way through the new region so that on average 3/4 of the array is used.
But Benford's law tells us that numbers (for some ill-defined class of number that probably includes array sizes) follow an improper distribution with the property that whatever base you pick, the mantissa M has a cumulative distribution given by log(M).
More precisely, if X is the size of a random array and S is the smallest power of 2 at least X, then we expect P(X/S) < z) = 1+log_2(z). Integrating we find
E[X/S] = 1/(2 log(2)) = 0.72...
slightly less than 3/4, but maybe close enough we don't need to worry about it.
https://en.wikipedia.org/wiki/Benford%27s_law
https://media.mathstodon.xyz/media_attachments/files/115/509/408/907/803/202/original/400963d4d137fe67.png

2025-10-26 00:11:50 UTC

I think the biggest surprise during my last trip to the UK was discovering you can buy sex toys in WH Smith now.