Weekend Editor

Retired physicist, after a career in machine learning & stats mostly for cancer drug discovery. Now blogging about stats in the news.

Avatar: convergence basins in the complex plane of Newton's algorithm searching for the cube roots of unity. (After a NYT column by @stevenstrogatz, long ago.)

Header: Quote from GK Chesterton, London Daily News, 1905-Aug-16 on epistemic humility and the ability to say "I am wrong" as the foundation of idealism.

#statistics #physics #r

Public Key

npub1g90kxj5pufgdlmp6v4ksekk4nf6ams90q6dv2ugcvxcmam7ys3yqmdff6u

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsyzhmrf2q7y5xlasax2mgvmt2e5awaczhsdxk9wyvxrvd7alzggjqpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds5e2e4s

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1g90kxj5pufgdlmp6v4ksekk4nf6ams90q6dv2ugcvxcmam7ys3yqmdff6u

Show more details

Last Notes

2026-04-06 20:21:22 GMT

Today's blog post:
DOOM can now be served and installed via DNS records, never touching a disk. Opinion among security experts seems divided between "OMG!" and "Yeah, that's been known for 40 years."
https://www.someweekendreading.blog/doom-dns/

2026-02-19 15:19:29 GMT

So *excruciatingly* polite and informative... up until the last sentence when the knife suddenly flashes.
Very well done.

2026-02-10 01:55:05 GMT

- reply

@nprofile…a3ey
Also, we've all been in the ogre's position, feeling like we only just *barely* understand the surface. Only after years of effort do we finally begin to understand anything deeply.
Anyone who says otherwise is egotistical, ignorant, or both.
Or, possibly, an ogre with even less ability for self-reflection than this one.

2026-02-10 00:29:12 GMT

- reply

@nprofile…wjmx
... and Common Lisp is a GPL with a way of extending its syntax for embedding DSLs.

2026-02-10 00:26:24 GMT

- reply

@nprofile…ur67
Hmm. We *do* keep re-inventing executable content in data formats, thinking it will finally be different this time.
I predicted last June this would be an avenue of attack, when some amiably crazy people made Doom run in a PDF's internal Javascript:
https://www.someweekendreading.blog/doom-pdf/

2026-02-05 16:13:58 GMT

- reply

@nprofile…tfwq
So, you making a stellarator?
'Cause, if so... blessings upon you, my son.
https://en.wikipedia.org/wiki/Stellarator

2026-02-02 15:45:58 GMT

- reply

@nprofile…xg4c
Can you point me at something to read that will show me how the former Nazis were still delusional?
I think I agree that this is probably the case, but I'm morbidly curious about seeing an example. (Feel free to advise me not to follow up on that curiosity for the sake of my sanity. :-)

2026-01-31 14:33:47 GMT

Numbers don't lie.
They do, however, occasionally burn.

2026-01-28 21:09:15 GMT

- reply

@nprofile…092r
I've seen them save everything ON THE DESKTOP, and then complain "the computer is full". :-(

2026-01-28 19:47:07 GMT

- reply

@nprofile…3n74
If the language under discussion is Intercal, then this makes perfect sense!
https://en.wikipedia.org/wiki/INTERCAL

2026-01-28 19:30:35 GMT

- reply

@nprofile…3n74
How many lines commented by £ and ¥? :-)

2026-01-28 16:47:13 GMT

- reply

@nprofile…vtpt @nprofile…c2nm
"... [they] have only have one idea: prompting chatbot. If they have a second idea it's prompting chatbot more."
They are well described in the 1845 Benjamin Disraeli (yes, *that* Disraeli) novel _Sibyl: Or, the Two Nations_, Book 4, Chapter 5:
"... he had only one idea, and that was wrong."

2026-01-28 16:36:06 GMT

- reply

@nprofile…gj0l
Rugby forum... so, we're thinking head trauma as the cause, then? :-)

2026-01-28 16:18:27 GMT

- reply

@nprofile…wjmx
Took me a minute to realize that by "G machine" you meant the graph reduction abstract machine. As in "spineless, tagless G machine."
See, Symbolics had a machine called the G machine, which was the L machine but redone in gate arrays (3610 & 3620, if memory serves). Just as fast or a bit faster, cheaper to make... and the marketing geniuses priced it *higher*.
Both called G machines, both novel architectures based on gate arrays, and both quite distinct things!

2026-01-28 16:14:31 GMT

- reply

@nprofile…gj0l
I *almost* bleeped over this as just incomprehensible.
But then... after 15 seconds of thought... it's *brilliant*.
It's not only funny on its own, but it leaves behind a free gift in the form of a puzzle: how in the world does one construct such a joke in the first place?!

2026-01-28 16:11:41 GMT

- reply

@nprofile…wjmx
Haskell nowadays is actually quite zippy, when used properly.
When used clumsily by a C/Java programmer, of course it's terrible. But that's because they won't take the time to learn to use the tools.
It's like, when I worked as a statistician for 20 years, I never *once* wrote a 'for' loop in R. Mapping functions across data structures was fast, but 'for' was interpreted, and slow.
Hence people always complained that "R is too slow", when really they just refused to use the tools the right way.

2026-01-28 16:05:28 GMT

- reply

@nprofile…wjmx
I still remember, at the 1982 Lisp & Functional Languages Symposium in Pittsburgh, seeing Turner give a demo of Miranda.
A quicksort in 3 lines! With performance data to show it really was n log n complexity!
(Of course, the coefficient in front of that n log n was huge, and thus people still preferred other languages.)

2026-01-28 13:36:53 GMT

- reply

@nprofile…tfwq
You *can* figure out LaTeX, and the details will eventually make sense. I mean, Knuth is not stupid!
But...
I've forgotten who told me this back in grad school, but it rang like a bell in my head with its strength of truth:
"To be able to use LaTeX is the mark of a scholar. To be able to use LaTeX WELL is the mark of a wasted life."

2026-01-28 13:29:11 GMT

- reply

@nprofile…en0v
Except... Symbolics monitors were bluish-white pixels on black. :-)
https://www.chai.uni-hamburg.de/~moeller/symbolics-info/

2026-01-28 00:21:41 GMT

- reply

@nprofile…ur67
I have just the web site for you to check up on CERN:
https://hasthelargehadroncolliderdestroyedtheworldyet.com/

2026-01-27 21:59:17 GMT

Yup. It's really cold, just like the good old days before global warning when we had serious winters.

2026-01-27 18:26:32 GMT

- reply

@nprofile…eyxl
Oh, good heavens! This is utterly and completely true.
I used to be regarded as "uncooperative" when I insisted on HUGO standard gene names for human genes, not the "folklore" names.
And don't get me started on how clever biologists are at creating yet another incompatible file format.
I once had a dataset delivered as over 100 binary Excel spreadsheets! (Fortunately, I could load them into R and make one omnibus .tsv file.)

2026-01-27 18:12:23 GMT

Honestly, it's also the hardest problem for the rest of us, too.

2026-01-25 16:05:30 GMT

- reply

@nprofile…e833
#CatsOfMastodon
In matters such as these, all cats are brothers.
https://media.mathstodon.xyz/media_attachments/files/115/956/585/519/094/340/original/de452cf278d5ec30.jpeg

2026-01-25 14:45:07 GMT

- reply

@nprofile…222n
I get your point.
I've been in a state of low-level panic since *Reagan*. But then, I'm old. No helping that.

2026-01-25 01:27:28 GMT

- reply

@nprofile…a570 @nprofile…vfwu
I've had polydactyl cats of whom I suspected something of this sort.
https://en.wikipedia.org/wiki/Polydactyl_cat

2026-01-24 22:46:27 GMT

- reply

@nprofile…vfwu
Thumbs! Thumbs! My kingdom for a pair of thumbs!

2026-01-24 22:41:03 GMT

It's ok to defend your mental health.
Especially if you already understand how bad the fascists are.

2026-01-24 16:51:14 GMT

- reply

@nprofile…ur67
Meryl Streep was *magnificent* in that movie!
That being said, here in southern New England, it's 8F as I type this. Wind chill, of course, is much colder.

2026-01-23 20:00:34 GMT

- reply

@nprofile…xfq3
I once had collaborators who wanted to measure expression, copy number, and mutation data on a wide variety of cancers and normal tissues.
Laudable, so far.
BUT: they insisted on processing them in a priority order, according to the priority of internal projects.
That meant: batch effects were convolved with organ effects, more or less impossible to tease apart.
Of course they blamed the uncooperative bioinformaticists (who had warned them about this in the beginning).

2026-01-23 15:16:29 GMT

- reply

@nprofile…9psq
Too bad that asking "From whom?" would have spoiled the joke.
I apologize for even thinking that.

2026-01-23 14:31:51 GMT

- reply

@nprofile…drt7
This one seems uniquely intense.
I think it's because, while the LLMs hallucinate, management hallucinates even harder about cost savings and potential layoffs.

2026-01-22 19:46:52 GMT

- reply

@nprofile…wsge
Odd, isn't it?
They insist on talking about LLMs *must* work... but they also insist on *not* talking about taxing the rich.

2026-01-22 19:05:34 GMT

Today's blog post:
Addendum re AI LLMs providing little benefit in business deployments. Price Waterhouse Coopers joins & agrees with MIT, CMU, Boston Consulting Group, the BBC, Reuters, The Economist, and Oxford Economics.
Evidence is piling up.
https://www.someweekendreading.blog/ai-update-2026/#addendum-2026-jan-22-price-waterhouse-coopers-agrees

2026-01-21 16:18:56 GMT

- reply

@nprofile…tfwq
I suspect that today's tech oligarchs would quickly leverage micropayments into macropayments.

2026-01-21 16:16:35 GMT

- reply

@nprofile…r6ml
We've surpassed the Gilded Age by a variety of measures, including the Gini coefficient. Corporate consolidation into monopolies is ongoing in multiple industries, obvious once you plot it. One paper, by economic historians, notes we are now worse than the Roman empire, with its aristocrats and slaves.
https://www.someweekendreading.blog/cornering-markets/

2026-01-21 15:53:15 GMT

"Profitable" is unspoken shorthand, expanding to "profitable for whom"?

2026-01-21 00:57:22 GMT

- reply

@nprofile…0deq
Tails are really fat, judging by the heights at the edges. That would incline me to guess a Cauchy distribution.
However, the central part appears to come to a point? That would incline me to a Laplace distribution.
No idea how to reconcile those, other than "Architects don't care about building resemblance to distributions." :-)

2026-01-20 16:52:27 GMT

- reply

@nprofile…ur67
Not sure I could have generalized from "pomegranate pips" to "things". :-)

2026-01-20 16:50:43 GMT

- reply

@nprofile…ur67
As a little kid, the moral I drew from Persephone (yeah, weird reading habits from unsupervised library wilding) was: "Don't take pomegranate pips from strange men."
As an adult, I realize that there may have been a few important bits to have learned from the story, that I missed as a kid. :-)

2026-01-19 04:01:16 GMT

Today's blog post:
Journal Club walk-through of a paper describing AI use and personality tests of users among students & gen'l public.
Conclusions: AI use is much less than you think, and heavy users score high on Dark Triad personality disorders.
https://www.someweekendreading.blog/who-uses-ai/

2026-01-17 22:42:31 GMT

- reply

@nprofile…600e
Love the "DRIVE SAFELY" sign.

2026-01-17 22:34:56 GMT

- reply

@nprofile…gent
We are (a) already obese, and (b) can't get GLP-1 receptor agonist drugs for obesity because our insurance won't pay for it.
"Cosmetic issues", don't you know. You have to become diabetic to get it prescribed and (mostly) paid for.

2026-01-17 16:21:35 GMT

- reply

@nprofile…l2f6
Here's a walk-through of an earlier paper (Ahlqvist, et al., 2024), reaching the same conclusion with a variety of pain killers and a variety of disabilities.
The key is to use sibling data so the comparison is genetics-aware. A genetics-ignorant comparison is useless, given the genetic component in autism, ADHD, and intellectual disability.
Exception: aspirin is *protective* against intellectual disability.
https://www.someweekendreading.blog/autism-acetaminophen/

2026-01-16 14:02:21 GMT

- reply

@nprofile…3fqs
Asking for advice:
What have Italians done to help recover from Berlusconi?

2026-01-16 13:35:56 GMT

- reply

@nprofile…l7vu
@nprofile…3nqu
A really weird addition is that US Republicans really seem to have declared war on weather forecasting, to the point where local armed groups destroy weather radars:
https://www.someweekendreading.blog/gop-war-on-weather/

2026-01-16 03:23:51 GMT

Today's blog post:
A chart that may reveal the priorities & values of US society, and thus where we need to do better.
https://www.someweekendreading.blog/soc-priorities-chart/

2026-01-15 20:25:16 GMT

- reply

@nprofile…4hp5
(This reply has nothing to do with the obvious humor you point out. I am, also humorously, taking you seriously when you probably didn't expect it. :-)
To be (more than) fair, MIT does have a remarkable humanities department. (Yes, "department": all humanities wedged into a single department. It's a niche interest at the Institute, though an intense niche.)
As a result, it takes a really *determined* professor to decide to teach there. They *really* want to spark an interest in the heads of the undergrads.
Heather Cox Richardson taught history.
Joe Haldeman taught writing.
Dirk Struik taught history AND mathematics.
I wasn't an undergrad in my time there, so I didn't have personal experience with "Course 21", as it was known. But EVERY SINGLE TIME I heard conversation about the quality of that small faculty, they were described as *excellent*.
But yeah, it's silly to put MIT at the top of the list for arts & humanities. You can GET a pretty good sideline in humanities there, but you have to seek it out.

2026-01-15 20:13:12 GMT

- reply

@nprofile…yjjy
Wrote about this just a couple days ago!
Oxford Economics notes the lack of productivity growth with LLMs, and hypothesizes that the LLMs are being used as camouflage for the layoffs that are the corporation's real goal.
Search for "Oxford" to get to that part:
https://www.someweekendreading.blog/ai-update-2026/

2026-01-15 14:19:20 GMT

- reply

@nprofile…zykx @nprofile…ds0l
One must recount the canonical stories of our tribe, in these circumstances:
https://www.hactrn.net/sra/vaxen.html

2026-01-14 21:09:48 GMT

- reply

@nprofile…e833
I don't suppose she knows how Ernst Röhm's story ended.
Just like Trump throwing a "Great Gatsby" party without knowing how that novel ended for Gatsby.

2026-01-13 21:52:05 GMT

- reply

@nprofile…vfwu
There's an awful lot to be said for retirement, if you can swing it.

2026-01-13 17:17:25 GMT

- reply

@nprofile…gent
What @nprofile…xq29 called his 2nd law of crime: "Punishable by a fine" means "legal for a price".
https://buttondown.com/kungfumonkey/archive/cons-heists-101-orientation/#:~:text=SIDEBAR%3A%20THE%20RULES%20OF%20CRIME

2026-01-12 16:21:35 GMT

- reply

@nprofile…n00n
That's exactly why, at our house last spring, we got MMR boosters with our COVID-19 boosters. In our 60s and 70s.
https://www.someweekendreading.blog/vax-vs-knuckleheads/

2026-01-12 14:35:26 GMT

- reply

@nprofile…jvst
Try Object COBOL instead.
https://www.wiley.com/en-us/An+Introduction+to+Object+COBOL-p-9780471183464

2026-01-12 03:26:37 GMT

Today's blog post:
A look at 7 or so studies of corporate AI use, and the miserably high failure rate.
Oxford Economics suggests AI is being used as corporate camouflage, to make layoffs look good.
https://www.someweekendreading.blog/ai-update-2026/

2026-01-11 15:26:12 GMT

- reply

@nprofile…dx40
Bonus points if you can make a tile of something like it that fills the plane.
Super extra bonus points if it's an aperiodic monotile, but that's for advanced players only. :-)

2026-01-10 20:12:44 GMT

- reply

@nprofile…6w3n
Doesn't destructuring-bind do something along these lines, if not exactly this?

2026-01-10 19:41:53 GMT

- reply

@nprofile…8qh7 @nprofile…m5ua @nprofile…am9t
You'll have to ask the Franklin Institute, because they're the ones labeling it as myth. I'm just trying to be honest about the possibly slightly doubtful nature of the story.
I rather imagine that the FI says it's (partly) mythical is because he said it privately to his daughter in a letter, not publicly advocating for it as the official symbol.
But that's just my guess.

2026-01-10 16:25:56 GMT

- reply

@nprofile…ur67
Agreed, though from a different personal history trajectory.
I was born into a Republican family that had been Republican since there *were* Republicans. My grandfather had a job at the courthouse when the mayor was Republican, and did not otherwise.
Also, it was what we'd now call a red state, and a small town -- semi-rural, really -- so "Republican" was just how most people were.
Sometime during Reagan, when I'd decamped to the Big Coastal City for grad school, it all stopped making sense to me. Maybe I finally came to my senses from broader experience, or maybe Reagan was just really that crazy. Either way, I became a Democrat.
Democratic candidates like Dukakis, well... they Just Made Sense to me.
Since then I've gotten rather further left, more like the DSA. Not that there are any candidates available in that slot, most of the time. (Ok, AOC & Mamdani. But I don't live in NYC.) And not that our first-past-the-post system would make them electable, most of the time.
Still, though our mutual origin trajectories differ, like you I think "Democrat" is as far right as I can vote. Even then, it takes some effort on my part just to do strategic voting for Democrats to monkeywrench the Republicans.

2026-01-10 13:59:47 GMT

- reply

@nprofile…m5ua @nprofile…am9t
Benjamin Franklin is often said to have preferred that the national symbol should have been the turkey, because it was useful.
Partly mythical, but he did say as much in a letter to his daughter:
"[The] Bald Eagle...is a Bird of bad moral Character. He does not get his Living honestly…[he] is too lazy to fish for himself. ... [The turkey is] a much more respectable Bird, and withal a true original Native of America...He is besides, though a little vain & silly, a Bird of Courage."
The Franklin Institute has the sources:
https://fi.edu/en/science-and-education/benjamin-franklin/national-bird

2026-01-08 23:46:04 GMT

- reply

@nprofile…ur67
Arnabu hurru ("rabbit hole", literally translated).

2026-01-08 23:14:08 GMT

Today's blog post:
Norwegian & French studies show COVID-19 mRNA vax lower ALL-CAUSE mortality, following people for 4 years.
Danish study shows updated JN.1 vax lowers 10/29 medical events in month after vaccination.
Also, mRNA vaccines INCREASE effectiveness of certain cancer immunotherapies.
https://www.someweekendreading.blog/covid-vax-mortality-again/

2026-01-08 14:19:47 GMT

- reply

@nprofile…ur67
Given your interest in cuneiform, you might want to browse the newly issued, free access (in PDF; hardcopy is $1500), comprehensive dictionary of Akkadian.
21 volumes, taking 90 years to complete. OTOH, nobody's spoken much Akkadian in several millennia, so maybe there wasn't a hurry. :-)
https://www.openculture.com/2026/01/dictionary-of-the-oldest-written-language-free-online.html

2026-01-07 19:53:18 GMT

- reply

@nprofile…4hp5
> "No fields in the Lagrangian that can't be varied!"
This does seem to be the general sentiment as far as fundamental physics goes.
But... when you let us humans introduce heuristics to do approximate calculations, all kinds of things turn out to be useful.
Effective field theory and mean field theory are good examples.
Energy levels in a nucleus are sort of hard ab initio, because the strong force resists perturbation theory. But mean field theories and things like the nuclear shell model (Mayer & Jensen, Nobel in 1963) engineers in a starting point. You start with the nuclear 2-body Hamiltonian, then add and subtract a 1-body potential 𝑈:
𝐻=𝑇+𝑉₂=[𝑇+𝑈]+[𝑉₂−𝑈]
Then you start exploring clever choices of 𝑈 that are calculationally tractable (at least numerically), will capture a few gross features (size of a nucleus) and zero out as many low-lying matrix elements of 𝑉₂−𝑈 as possible (to simplify perturbation theory).
The magic is that a simple harmonic oscillator potential for 𝑈 will do, and you can do perturbation on top of that. There are better ways, but the big surprise is that strongly interacting nucleons can be characterizsed at 0 order as otherwise free particles in a magic harmonic oscillator. Hence the stable nuclei with filled energy levels are called "magic number nuclei".
Anyway, the point here is that you can introduce this big static background potential and then perturb around it to get quite good results.
To your point: this is a calculational tactic, not fundamental physics.
https://en.wikipedia.org/wiki/Nuclear_shell_model

2026-01-07 18:51:07 GMT

- reply

@nprofile…wz83 @nprofile…pam6 @nprofile…ugpy
This was a famous annual ritual at MIT, during IAP (January "independent activities period").

2026-01-07 16:07:40 GMT

- reply

@nprofile…3e6m
Danny Kaye's got your back, in his 1955 film _The Court Jester_:
"The vessel with the pestle
holds the pellet with the poison
the chalice from the palace
holds the brew that is true."
https://www.youtube.com/watch?v=WzmnSyqv37A

2026-01-07 01:07:51 GMT

- reply

@nprofile…600e
Meanwhile, in the fact-based world...
The Ahlqvist, et al. study from 2024 used the ENTIRE NATION of Sweden as its trial population (let's hear it for universal health care and uniform electronic medical records systems).
Once you do a genetics-aware analysis by using sibling data, there is no effect for autism, ADHD, or cognitive disability from prenatal exposure to 5 pain relievers (including opioids!).
Exception: Aspirin is statistically significantly PROTECTIVE against cognitive disability.
I blogged about it here, last September:
https://www.someweekendreading.blog/autism-acetaminophen/

2026-01-04 00:27:55 GMT

- reply

@nprofile…ur67
Hmm... this looks interesting.
What's the experience of cleaning it like?

2026-01-03 01:19:31 GMT

- reply

@nprofile…vfwu
I pretty much parked at 60/40 (stock/bond ratio) for the last couple years of working, and had planned to keep it that way in retirement.
But... the Trump-revenant era made me want some more armor plating. And the Vanguard analysis of this year, which is REALLY interesting, persuaded me to go even further.
So now we're at 40% stocks / 60% bonds. About 60% of stocks and 50% of bonds are outside the US. There are strong tilts to small-cap and value stocks, to get away from the "Magnificent 7" big tech stocks that seem to be basically an AI/LLM bet.
That's pretty much where the Vanguard analysis left us.

2026-01-03 01:03:55 GMT

Today's blog post:
Vanguard recommends retirement portfolios get more conservative: 60% - 70% bonds!
I looked through their data, compared to another source and... agreed. It feels wrong to get that conservative, but that's what I did to start 2026.
https://www.someweekendreading.blog/weekend-portfolio-trump-3/

2025-12-31 17:27:46 GMT

- reply

@nprofile…25az @nprofile…atf6 @nprofile…00p2 @nprofile…ag4u
I had a similar reaction to topology when I first looked at it.
Just telling me that a topology is a pile of sets that behave a certain way under intersection & union was... uninformative, and seemed arbitrary.
But by the time I tried topology again, I'd learned that
(a) they were using open sets as a concept for "neighborhood",
(b) that topologies just codified what I already knew about open sets, and
(c) the whole preoccupation with open sets had to do with what was invariant under continuous functions.
If only I had noticed the FIRST time that we were just asking "hey, what DOESN'T change if we shove things through a continuous function?"

2025-12-31 16:20:29 GMT

- reply

@nprofile…j4rt
A look at the state of quantum factoring, and how some researchers proposed a speedup involving the use of a dog:
https://www.someweekendreading.blog/dog-factoring/

2025-12-31 16:12:27 GMT

Today's blog post:
Russian casualties in Ukraine now exceed 1.2 million.
A regression model on the data over time clearly suggests a hinge model, where the death
rate increases about 40% in April 2024.
https://www.someweekendreading.blog/ukraine-1200k/

2025-12-31 14:55:31 GMT

- reply

@nprofile…9jjk
When I first saw Kinkaku-ji, which is apparently what's in your picture, I was very puzzled.
On the one hand, Japanese were very eager to say it was quite ancient, built in 1399 or so.
On the other hand, it burned down in 1955 and was rebuilt as it stands today.
It puzzled me that they thought of it as the original temple, when it seemed clear to me that it's a modern copy (though apparently a very accurate one).
It's a Ship of Theseus problem: what matters is not so much the continuity of the exact bits of wood and metal, but rather the *cultural* continuity as a national treasure, a Zen temple, unbroken traditions, and so on.
It's sort of a Zen thing: the "real" Kinkaku-ji is not the physical one that you see, but the spiritual one that has long continuity in Japanese culture.

2025-12-31 14:38:37 GMT

- reply

@nprofile…8eua @nprofile…d9sd
An acute observation.

2025-12-31 14:29:07 GMT

- reply

@nprofile…n9c6
Also Riker: "When the train comes in, everybody rides."

2025-12-31 14:05:56 GMT

- reply

@nprofile…ygg9
It's 18 years old, though?
Sort of a historical snapshot of Haskell.

2025-12-30 01:24:11 GMT

Today's blog post:
Trump State Dept bans Calibri, a "wasteful DEIA" typeface (Biden admin selected for readability w/visual disabilities). Trumpers demand Times New Roman, in tradition of newspapers for rich conservatives.
I have "a modest proposal".
https://www.someweekendreading.blog/helpful-advice-typefaces/

2025-12-29 19:40:08 GMT

- reply

@nprofile…ueh4
I escaped tech in the 90s.
I finished a PhD and became an applied scientist, mostly doing applied math, a bit of logic programming, and later applied statistics & elementary machine learning. Though, to be fair, "machine learning" was mostly classical stuff like fitting logistic regression models.
So... not "tech", but tech-adjacent enough that sometimes managers had trouble telling the difference. Those managers were very much worth avoiding.
Example: One employer merged "Information Science" (statistics, applied math, data analysis) with "Information Technology" (computer support) and "Information Services" (librarians). They thought there must *clearly* be synergy there, since we all had "information" in our department names. More than once a manager asked me to fix his printer.
Most of that was in cancer drug research.
I retired in 2020.

2025-12-29 18:17:34 GMT

- reply

@nprofile…dzyf
Piet: source code is an image that usually looks like a Mondrian painting. The model is a 2 dimensional state machine with stack (PDA).
(No, this is not a serious suggestion. But it *does* change people's boundaries about what a programming language's notation can be.)
https://www.dangermouse.net/esoteric/piet.html

2025-12-28 09:22:59 GMT

- reply

@nprofile…fnwr
No criticism intended! Just happily playing with the puzzle.

2025-12-28 02:04:32 GMT

- reply

@nprofile…fnwr
Struggling to read it (looking at syllabaries and hoping for the best is seldom a recipe for success):
a ta ra * *
ta ra * nu ku
Where * means "beats the heck outta me".
Did I get even a single syllable right?

2025-12-27 21:46:12 GMT

- reply

@nprofile…fnwr
There's a scene in the webcomic _Gunnerkrigg Court_ where some robots are having a discreet conversation in the background while the main characters talk about the main plot line.
The robots are saying something to each other in binary. If you group it in 8-bit bytes and translate to ASCII, they are saying, "BLAH BLAH BLAH".
I love it when I'm not the weirdest one in the room. So many ways to learn to be weirder, for one thing.
https://www.gunnerkrigg.com/?p=1

2025-12-27 21:05:09 GMT

- reply

@nprofile…fnwr
Well, maybe more than a couple... but yes, it would be a joke for a very select audience.

2025-12-27 21:04:21 GMT

- reply

@nprofile…fnwr
Related: A friend is a classics prof.
One of his favorite moment in ancient Greek is to ask students about the grammar of a certain passage. When they analyze it in detail, they see that some genders don't match.
He tells them they *do* match, and they are puzzled.
He then reveals that what's weird about that passage is that a woman in the group is speaking. So rarely did anybody care about what women had to say, that it's grammattically odd.

2025-12-27 21:01:52 GMT

- reply

@nprofile…fnwr
It would be a complete hoot if the "Book of Wisdom" were written in cuneiform, but the text turned out to be the Ea-nāṣir complaint letter. :-)

2025-12-27 19:52:44 GMT

- reply

@nprofile…fnwr
Re Latin: I was chased through public schools by teachers intent on making sure I used classic Roman pronunciation, "lest people assume you're Catholic". (That was a thing, half a century ago in the upper Midwest. Yet another reason to migrate to a coastal city.)
Re cuneiform and "ancient truth": The only bit I've had to use in teaching is the Cyrus cylinder, which is about half remarkable good treatment of the conquered and half self-aggrandizing propaganda. All the other texts I've worked with (in translation) were 100% propaganda. (With a notable exception for the complaint to Ea-nāṣir, which is notably exceptional for many reasons. That one might actually *be* ancient truth, sadly enough.)

2025-12-27 18:41:14 GMT

- reply

@nprofile…fnwr
I'm aware of the clay tablets and even the wax tablets (sort of like "scratch paper"). I didn't know there was much in the way of papyrus with cuneiform, though.

2025-12-27 15:47:40 GMT

- reply

@nprofile…ygg9
I learned assembly on the CDC 6400/6500.
It was... different than everything else. :-)

2025-12-27 15:38:47 GMT

- reply

@nprofile…aya8
I once wrote a paper about an expert system language we called "Joshua", on which my co-authors added the subtitle:
Why Joshing is Better Than Conniving or Planning
It seemed funny at the time.
https://cdn.aaai.org/AAAI/1987/AAAI87-009.pdf

2025-12-27 15:14:56 GMT

- reply

@nprofile…2slu
Stein was a very conservative economist, of the American Enterprise Institute sort.
So while his saying is quite the _bon mot_, I'm always queasy about quoting him.
I guess even the other side can have a point, now and again.

2025-12-27 02:29:59 GMT

- reply

@nprofile…fnwr
Cuneiform is very cool, but it's generally found on baked clay tablets instead of books. A *book* of cuneiform would be anachronistic, like a modern book *about* cuneiform.
So the alternative might fly a bit better, with or without egg nog?

2025-12-25 20:36:54 GMT

- reply

@nprofile…tw26
It makes sense... if by "sense" you mean strict adherence to a slightly counter intuitive definition. (Each argument must be a proper list, except for the last argument which can be anything.)
Here's an SBCL example:
* (append '(1 2 3) 4)
(1 2 3 . 4)
https://www.lispworks.com/documentation/HyperSpec/Body/f_append.htm

2025-12-25 20:33:25 GMT

- reply

@nprofile…fnwr
Crows offer advice in exchange for pecans?
My cats won't *stop* offering advice until they get treats.
Though, to be honest, I think most of the "advice" is "you should give us treats".

2025-12-25 18:41:43 GMT

And Dickens' _A Christmas Carol_ is a story in which 3 sprits broke into a rich man's house and tormented him psychologically until he paid his workers better.
That's the MODERATE version of how most of us now view the rich.

2025-12-25 18:37:11 GMT

- reply

@nprofile…2whg
Then at a ~50mln ly distance, it looks like NGC 646 is just "in front of" PGC 6014, with a tail that happens to point that way.
Still quite pretty, though.

2025-12-25 18:29:35 GMT

- reply

@nprofile…2whg
So is it just a foreground/background coincidence, or does the tail of gas & stars tell of a previous interaction?
By the numbers you're quoting, they're about 50mln light years difference in radial distance from us.

2025-12-25 18:25:57 GMT

- reply

@nprofile…jr73
Probably the same idiot who decided that password vaults like Bitwarden should not be allowed to copy usernames and passwords into those fields, but require manual copying instead.

2025-12-25 18:23:44 GMT

- reply

@nprofile…a554
Link says "we do not provide services in your area."
I'm in the US, using a VPN hosted in New York.
https://media.mathstodon.xyz/media_attachments/files/115/781/597/663/879/398/original/fd048047d3110529.jpg