fl

@nprofile…wjmx Obvioulsy using Grok or any other LLM is also a good idea to learn assembly (comments about code, code generation, explanation about subtle points (lea vs. mov, addressing modes, modes and submodes of the x86-64 processors etc.).

2026-01-15 15:14:33 UTC

- reply

@nprofile…wjmx I have found this kind of translation (from C to Assembly) priceless to learn Assembly. And I think, if it is feasible, that translation from F* to Assembly might be a good idea.

2026-01-15 15:08:10 UTC

- reply

@nprofile…wjmx "Grok" is not my "budy". I think their product is of remarkable quality no more. I know the dangers that fly over democracy nowadays.

2026-01-15 14:55:10 UTC

- reply

@nprofile…wjmx I see the F* programs can be translated to C and Assembly. Do you know if the final product is cluttered with a lot of useless instructions or does it ressemble genuine C or assembly?

2026-01-14 19:56:14 UTC

- reply

@nprofile…t6k2 I'm very often buffled too by the accuracy of their answers concerning technology or softawares use.

2026-01-14 19:43:42 UTC

- reply

@nprofile…2jgt The problem is that in set theory for instance the only non logical symbol in the axioms is $\in$, and that it is built over classical first order logic with equality. So the normal way to define truth is to take a set therory theorem, or a first order logic with equality theorem. In the case "x = x" is a simple logical theorem that can be used to mean truth. You can obviously define a new symbol as Truth <-> x = x. But Anyway what is behind is x = x.

2026-01-14 14:41:20 UTC

- reply

@nprofile…w3sa @nprofile…rxrg @nprofile…xuzd Now they verify hundreds of topological theorems automatically thanks to LLM.

2026-01-14 14:36:52 UTC

- reply

@nprofile…f38c C standard recently dropped support for ones' complement representation. It still allows bytes to be a size more than 8 bits.

2026-01-14 14:24:13 UTC

Tremendous or Scarry.
https://mathstodon.xyz/@Jose_A_Alonso/115893690687477590

2026-01-13 20:49:28 UTC

The defense calculator was the 701. It's tremendous isn't it.
@nprofile…wjmx
https://www.google.com/search?q=Backus%E2%80%99s+1977+Turing+Award+lecture&oq=Backus%E2%80%99s+1977+Turing+Award+lecture&gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIJCAEQIRgKGKABMgkIAhAhGAoYoAEyCQgDECEYChigATIJCAQQIRgKGKAB0gEJMTEwMmowajE1qAIIsAIB8QXCRFZHAbdIWQ&sourceid=chrome&ie=UTF-8#fpstate=ive&vld=cid:b0406426,vid:yKXiQAqou04,st:0

2026-01-13 20:00:53 UTC

- reply

@nprofile…xuzd Thank you for the answer.

2026-01-13 12:44:40 UTC

- reply

@nprofile…xuzd What is a weeknote? A summary of your teaching during the past week?

2026-01-12 22:02:46 UTC

- reply

@nprofile…cjj5 C'est le plaisir de la géométrie classique: de merveilleux dessins partout.

2026-01-12 21:49:37 UTC

- reply

@nprofile…wjmx "But don’t assume that 'old' means 'simplistic'."
I don't. I only think this paper is peculiarly allusive and would deserve a full historical comment to understand exactly what was the historical background.

2026-01-12 21:29:32 UTC

- reply

@nprofile…wjmx "I recommend that you read Steele's "The Evolution of Lisp" (1996), instead of Grok."
We can read both. :-)

2026-01-12 09:51:02 UTC

- reply

@nprofile…wjmx By the way reading Lisp I manual is very instructive. Since you can see what had been invented by McCarthy and his staff by 1960.

2026-01-10 22:33:23 UTC

- reply

@nprofile…wjmx "But even these erudite, senior folk balk at my advice to show due regard for the "classics" in #computing by reading early seminal papers (Gödel, Church, Turing, Kleene, von Neumann, Shannon, etc.)"
The GOTO fight...

2026-01-10 21:58:38 UTC

- reply

@nprofile…t6k2 Try C-h k Esc-q to see what it means. C-h k gives the Emacs function a key is connected to.

2026-01-03 14:36:15 UTC

- reply

@nprofile…wjmx This is a rather interesting educational program. I understand why your frontend uses goto and call/return instructions. You don't want the students to get annoyed with syntax analysis. But there are two ways to remove the burden of syntax analysis: prefix or postfix notations.
There is a third way to create a minimal barely Turing-complete high-level language: no function calls (they are difficult to parse). The only constructs would be: "if", infinite "loop", and a break instruction This approach is preferable because "goto", "call", "return" are too close tp assembly.language and you want frontend and backend are very different.

2025-12-24 15:03:59 UTC

Les enseignants de mathématiques, ces capillotracteurs !
https://www.parents.fr/actualites/etre-parent/je-ne-comprends-pas-une-maman-demunie-face-aux-maths-de-sa-fille-de-6-ans-fait-reagir-des-milliers-de-parents-1145036

2025-12-22 14:31:54 UTC

Use Grok to add MSC2020 classification to your articles.
You would feel as if the great philosopher Dewey were by your side. Or maybe Peirce himself.

2025-12-22 14:27:10 UTC

According to Grok an article about the Delaunay-Rips filtration deserves the 55N31 MSC2020 class.
"55N31 (Algebraic topology → Homology and cohomology theories → Persistent homology and applications, topological data analysis)." (Grok)
And it might benefit from the secondary classes 55U10 and 62R40.
"55U10 (Applied homological algebra and category theory → Simplicial sets and complexes), for the focus on simplicial filtrations.
62R40 (Statistics → Statistics on algebraic and topological structures → Topological data analysis), if the article emphasizes data analysis or machine learning applications." (Grok)
But it concludes that:
"An article focused on theoretical properties, stability, or algorithms for Delaunay-Rips would most accurately be attributed to 55N31 as primary."

2025-12-22 14:07:49 UTC

Personally Grok inspires me to explore new mathematical articles I would otherwise never open.
Today an article about the Delaunay-Rips filtration.
You need some idea of what an article is about. And very often unfortunately it is as mysterious as America before the arrival of the Conquistacdors.
https://arxiv.org/abs/2512.17382v1

2025-12-21 15:03:03 UTC

- reply

@nprofile…ygg9 Are types useful?

2025-12-21 14:52:44 UTC

- reply

@nprofile…ygg9 Funny.
Personally I think Gödel's general recursive functions are underrated. He was inspired by Herbrand, who died so young. Otherwise there might have been a French school of logic.

2025-12-21 14:41:33 UTC

- reply

@nprofile…ygg9 But Lambda calculus is a formal system and then the only machine you need is a machine that replaces metavariables by expressions.

2025-12-20 15:03:35 UTC

"C’est simple : Elon Musk vient de récupérer [pour son compte en banque personnel] l’équivalent du PIB du Maroc, juste parce que trois juges ont décidé que la décision précédente était trop sévère."
https://www.frandroid.com/marques/tesla/2917397_140-milliards-de-dollars-elon-musk-gagne-son-bras-de-fer-et-le-plus-gros-cheque-de-lhistoire

2025-12-20 13:14:16 UTC

I've finished my first complete program in assembly and I've shown experimentally that the probability that GCD=1 when the inputs are chosen at random is close to 60%.

2025-12-20 13:04:46 UTC

- reply

@nprofile…ygg9 @nprofile…dnqz It's really curious. They had such a brilliant school of mathematics, and it didn't turn into a brillant school of computer science.

2025-11-23 16:32:23 UTC

- reply

@nprofile…7ru6 "justes, nuancées, personnalisées"
... et bienveillantes. 🙂

2025-11-21 11:18:46 UTC

- reply

@nprofile…7ru6 Il a eu combien ?

2025-11-20 14:57:51 UTC

Puisque Michèle Audin vient de disparaître et que le site "Images des mathématiques" est très présent sur Mastodon, je mets la liste des articles quee Michèle Audin a écrit pour eux.
https://images.math.cnrs.fr/auteurs/michele-audin/

2025-11-19 15:42:33 UTC

- reply

@nprofile…ygg9 Thank you.

2025-11-19 15:04:48 UTC

- reply

@nprofile…ygg9 All that is a bit touchy I agree.

2025-11-19 14:56:44 UTC

- reply

@nprofile…ygg9 I haven't had a look to State but I had a look the Monads in the documentation. There is Monad called NoIO. Meditate.
https://hackage.haskell.org/package/base-4.21.0.0/docs/GHC-GHCi.html#t:NoIO

2025-11-19 14:40:43 UTC

- reply

@nprofile…ygg9 "At the foundational level, all "modern computing" is still von Neumann implementation of Turing's A Machine. Is that where you're leading to?"
No not at all. I perfectly understand the interest of purely functional languages. But there are two ways to use Monad: one (without IO) is purely functional, the other one (with IO) is no longer so.

2025-11-19 14:34:00 UTC

- reply

@nprofile…ygg9 I wouldn't say that. Monad is a way to simulate successive values of a variable using the properties of lambda calculus application. It is still functional programming once you have translated the syntactical sugar into their lambda calculus counterpart. But IO is another beast. Namely the values of something (a file) can be modified outside from the function. When you are in the domain of the Monad, referential transparenty is still respected. When you are in the domain of the IO it is no longer the case.

2025-11-19 14:19:59 UTC

- reply

@nprofile…ygg9 "Most popular languages today are PP-OO hybrids or OO-FP hybrids. This may well be due to the conceptual difficulty and the pragmatic inconvenience of striving to be paradigmatically pure, when performing real-world tasks."
Haskell has found a solution to this problem. The language is not pure since there are I/O. But the system of types allows to make the difference between what is pure and what is not. The impure parts of a program are easily identified: there return type is IO().
Such an idea could be applied to the other mix of paradigms.

2025-11-11 20:12:01 UTC

Le Français Yann LeCun, figure de l’IA, va quitter le groupe Meta pour créer sa propre société
https://www.ouest-france.fr/high-tech/le-francais-yann-lecun-figure-de-lia-va-quitter-le-groupe-meta-pour-creer-sa-propre-societe-d904e814-bf14-11f0-a7a6-0300a242c945

2025-11-11 20:02:30 UTC

- reply

@nprofile…wj35 L'IA répond: "The first time a function is run in a Julia session, it needs to be compiled. This can lead to a noticeable delay"
et
"The key to achieving high performance in Julia is writing “type-stable” code. A function is type-stable if the type of its output can be inferred purely from the types of its inputs."

2025-11-11 19:45:28 UTC

- reply

@nprofile…wj35 Utilisez Grok. Vous pouvez lui poser toutes les questions que vous voulez. En peu de temps ça va revenir.

2025-11-10 12:42:09 UTC

- reply

@nprofile…qwt7 @nprofile…u6hy Hilbert is your friend. Take a fresh start.
https://www.gutenberg.org/files/17384/17384-pdf.pdf

2025-11-10 12:37:09 UTC

It seems that it takes $500 bilions to train chatGPT.

2025-11-09 13:20:10 UTC

- reply

@nprofile…67su
The answer of Grok to the former remark. It is so cute, so eager to do well...
"There are things at which it is very bad. When I ask it in what version of glibc appeared (let's say) a given macro it fails and answers bullshit. It is not good at erudition."

2025-11-09 13:19:24 UTC

- reply

@nprofile…67su There are things at which it is very bad. When I ask it in what version of glibc appeared (let's say) a given macro it fails and answers bullshit. It is not good at erudition.

2025-11-09 13:09:44 UTC

- reply

@nprofile…67su I'm aware that it needs to be tried out and possibly fixed. But there are many good ideas.
I had made my own version of the Makefile before asking the LLM. So I can compare the two. It is better than I.

2025-11-08 21:47:50 UTC

I've asked an LLM to make a complicated Makefile. And it succeeded easily. I'm so impressed. When I see that, in France, some people on the far-left expect this technology. to collapse...
I have no words...

2025-11-06 20:27:20 UTC

« Un labyrinthe dont on ne peut pas sortir » : des professeurs racontent le harcèlement par leurs pairs et leurs supérieurs hiérarchiques
https://www.lefigaro.fr/un-labyrinthe-dont-on-ne-peut-pas-sortir-des-professeurs-racontent-le-harcelement-par-leurs-pairs-et-leurs-superieurs-hierarchiques-20251106

2025-11-06 19:55:42 UTC

'Le milliardaire anti-écolo Bernard Arnault achète « Challenges » et deux journaux scientifiques [ "La Recherche" et "Sciences et Avenir" ]'
https://reporterre.net/Le-milliardaire-anti-ecolo-Bernard-Arnault-rachete-Challenges-et-deux-journaux-scientifiques

2025-11-05 15:42:16 UTC

- reply

@nprofile…u6hy Curious. Have you tried one with 1392637 a's ?

2025-11-02 12:35:05 UTC

- reply

@nprofile…p3nn Have you used a LLM?

2025-11-02 12:33:27 UTC

- reply

@nprofile…22u8 On the other hand he was terrible at quadratic equations.
Nobody's perfect.

2025-11-02 12:29:27 UTC

- reply

@nprofile…22u8 I preferred Monopoly. It was more balanced.

2025-11-02 12:27:58 UTC

- reply

@nprofile…22u8 My father never let me win at checkers. Never.

2025-11-01 13:34:49 UTC

I don't understand why it is said McCarty borrowed Lisp from Chruch while it is so clear that he borrowed Lisp from Gödel.

2025-11-01 13:30:26 UTC

Another question for your dear pupils: using your favorite LLM show that Gödel's general recusrive functions are as powerful as Turing's machine?

2025-11-01 13:27:41 UTC

One question for your dear pupoil: can you code Euclid's algorithm (the original one with substractions) using only the primitives allowed by Gödel's general recusrive functions theory?

2025-10-27 10:40:37 UTC

Jusqu'où l'humiliation !
https://x.com/Charlie_Hebdo_/status/1980665720607273180

2025-10-25 18:56:13 UTC

- reply

@nprofile…w448 Freud would answer are the only one to know. You have to associate to find the answer.

2025-10-24 16:57:07 UTC

- reply

@nprofile…w8t5 I try but making the difference between what is idiomatic and what is not when you are a foreigner...

2025-10-24 16:54:11 UTC

- reply

@nprofile…w8t5 Personaly I write the texts myself first and then ask the AI if they are in good and idiomatic English. I often find its suggestions very useful.

2025-08-28 11:24:03 UTC

- reply

@nprofile…ca6t Impressiionnant.

2025-08-27 20:53:38 UTC

- reply

@nprofile…m0kr Looks good. Thank you. It's in some typed lambda caluculs apprently.

2025-08-27 13:40:43 UTC

- reply

@nprofile…m2ru @nprofile…jc0y I would saiy it depends how old you are.

2025-08-27 13:06:43 UTC

Personnaly I think that category theory books are not enough formal. Too intuitive. In my opinion.

2025-08-27 12:31:04 UTC

- reply

@nprofile…zenm Do you have a mathematical reference? Just to see what it looks like in mathematical form.

2025-06-04 10:12:27 UTC

- reply

@nprofile…m2ru @nprofile…jnh6 What is curious in that kind of answer is that you also had bread and meet at home.

2025-05-22 10:14:54 UTC

- reply

@nprofile…xtcf A video that could be of interest for you @nprofile…jc0y and your lispers community. The personal life of McCarty with many beautiful videos about the computers in the USA in the 40s.

2025-05-20 10:02:19 UTC

- reply

@nprofile…5k3r I really like these charts. What book are they from?

2025-05-17 11:53:24 UTC

- reply

@nprofile…e5g0 "Depopulation" is not a problem it is the solution.

2025-05-13 19:18:37 UTC

- reply

@nprofile…59k6 We need an AI which is able to answer any question about the standard we ask, with the reasoning that leads to the answer and all the references to the standard.
Currently it doesn't exist: the answer of any AI about the C type system is pitiful.

2025-05-13 09:49:20 UTC

- reply

@nprofile…w4ge @nprofile…wn5k @nprofile…l0de It's not considered as a phonem is it?

2025-05-11 17:52:16 UTC

- reply

@nprofile…fa4k It is a real philosophical issue. In the street this afternoon I was looking to passerby of one gender or the other and at my great disbelief they were behaving as if they were of one of the two sexes.

2025-05-11 12:00:57 UTC

- reply

@nprofile…59k6 Try the chefs-d'oeuvre : The Hate (Mathieu Kassovitz) or The Missing Picture (Rithy Panh).

2025-05-11 11:48:07 UTC

- reply

@nprofile…fa4k Let's negotiate the sense! Every parent approves.

2025-05-10 20:04:46 UTC

- reply

@nprofile…hszg You meant "Hawking"?

2025-05-01 20:46:27 UTC

- reply

@nprofile…e5g0 Parsing is amazing.

2025-04-29 20:59:25 UTC

- reply

@nprofile…m2ru Ireally like the drop shadows of the photo., Dr. StrangeLove-style.

2025-04-29 20:27:27 UTC

- reply

@nprofile…m2ru yes it is true. Lisp I was prepared for the IBM 704. First line of the preface.
Thank you this photograph.
And for this information about the IBM 704.
https://bitsavers.org/pdf/mit/rle_lisp/LISP_I_Programmers_Manual_Mar60.pdf

2025-04-20 15:48:15 UTC

- reply

@nprofile…fa4k @nprofile…k56a You also have to explain what is the relationship between polarization and magnetism and electricty. At first thought there is none. Faraday's experience is a miracle. It was the experience that proves the nature of light is electro-magnetic. Quite a surprise!

2025-04-16 12:43:52 UTC

- reply

@nprofile…59k6 "I'm ashamed to admit it, but I don't think I've ever been able to genuinely motivate the definition of a topological space in an undergraduate course."
Normal. General topology didn't start historically with open sets but with neighborhoods defined in avery loosy manner by Weierstrass. It is only tenths of years later that mathematicains came up with the idea that open sets could replace neighborhoods in the definition and that it would be shorter.

2025-02-07 21:46:21 UTC

- reply

@nprofile…6kk0 Functionnal programming.

2025-02-03 12:41:00 UTC

- reply

@nprofile…6kk0 @nprofile…me4e I would like to instead connect these things to real life
Be careful the logic of real life is not classical logic. Classical logic only applies to eternity, to mathematics.

2025-02-03 12:37:55 UTC

- reply

@nprofile…6kk0 @nprofile…me4e "Yeah... I think we do need to teach how not to make these kind of mistakes, but I am doubtful that formal logic is the way to do it..."
I've no objection to that. I perfectly understand what you mean.

2025-01-25 20:52:36 UTC

- reply

@nprofile…fa4k @nprofile…kppt The witch hunt.

2025-01-25 20:37:30 UTC

- reply

@nprofile…5k3r You're going to live a ripe old age.

2025-01-12 14:41:38 UTC

- reply

@nprofile…xtcf "Part of what makes Lisp distinctive is that it is designed to evolve. As new abstractions become popular (object-oriented programming, for example), it always turns out to be easy to implement them in Lisp."
I wonder if we might implement static type checking (and patterns too.) in Common Lisp, turning it into a kind of ML.
@nprofile…jc0y

2025-01-06 21:45:25 UTC

- reply

@nprofile…3yjd I'd like to take this opportunity to republish Benoït Rittaud's video : "Maginfiques lograithmes" (a fine example of pedagogical video). It's in French but you can activate the translation. What concerns the number e and its relationship with logaritlms and hyperboles is at 13:20.

2025-01-05 18:32:55 UTC

- reply

@nprofile…fa4k It's not Adelard, it must be Abelard.

2024-12-27 12:57:58 UTC

- reply

@nprofile…qmju "This time, the paper was rejected for only being an epsilon improvement over the previous literature!"
LOL

2024-12-21 13:47:07 UTC

- reply

@nprofile…3yjd I have always the feeling that in physics we neglect the phenomena and experiments that are so important to understand what we are speaking about, in this case, the propagation of heat in solid bodies.

2024-12-21 13:42:52 UTC

- reply

@nprofile…3yjd What is interesting in Fourier's story is that he is building upon some experimental results discovered by Newton. The results are rarely described. It is unfortunate because in my opinion it is important for understanding Fourier's questions.

2024-12-13 12:58:40 UTC

- reply

@nprofile…qmju Perhaps physical libraries are a vestigial remnant of a pre-digital era, but I do have fond memories as a graduate student of randomly browsing books next to the ones I had been looking for, or articles after or preceding the one I was initially locating.
The late Jacques Roubaud called it "the good neighbor theorem". In a library the book you really need is always the book physically close to the one you are looking for.

2024-12-13 12:52:21 UTC

- reply

@nprofile…qmju In France may libraries, peculiarly scientific ones are throwing their books away.
Sad. Very sad.

2024-12-02 18:47:42 UTC

- reply

@nprofile…fa4k Now when I see that all of them are of category theory and spend their time in lean4. Well I prefer not to comment...

2024-12-02 18:34:29 UTC

- reply

@nprofile…fa4k There was a terrible anti-Bourbaki war at that time.
Set theory and structures have completely disappeared from high schools yet the has not improved significantly.

2024-12-01 17:39:36 UTC

- reply

@nprofile…qmju It reminds me Dr Eugenia Cheng who wrote in "The Joy of Abstraction" that the free monoid (the category) is an operation with no extra equation at all.

2024-11-24 19:22:30 UTC

Why Turing's machine are the universal basis for computers while the mathematicians have described several alternate models of effective computability : lambda calculus, general recursive functions, etc. Does it mean that in a certain way Turing's machines are more fundamental ? (While we know that they are just as potent as the other models.)