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 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.

2026-01-13 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.

2025-12-21 15:03:03 UTC

- reply

@nprofile…ygg9 Are types useful?

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 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 ?

2025-11-19 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 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 12:37:09 UTC

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

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.

2025-11-08 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...

2025-11-05 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 12:29:27 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?

2025-10-25 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.

2025-09-03 20:02:59 UTC

- reply

@nprofile…ca6t Il est pas bête.

2025-08-26 12:26:35 UTC

- reply

@nprofile…zkqj A perfumed candle ;-)

2025-08-25 19:01:44 UTC

- reply

@nprofile…mkep It is not free!

2025-08-25 11:03:42 UTC

- reply

@nprofile…vy0z I have always been a water fanatic. Its chemical properties are so impressive.

2025-08-25 10:56:50 UTC

- reply

@nprofile…qmju Why I don't understand with tools like grind is that they return that they have found a proof while I would expect that they return a pedestrian proof. Why woud I expect that they return a proof? Because I could check it myself. Often 'grind' is used for very simple proofs like propositional ones, or predictae calculus ones or simple arithmetcal ones and returning a simple proofs should be possible. Unfortunately what you find if you look at its work behind the scene is mere garbage. Something half human, half machine is missing in my opinion.

2025-08-24 15:50:50 UTC

I'm a bit anxious beacuse Intel is currently losing a billion dollars a month. And I'm very useded to their processors and would prefer not to have to learn a new assembly language.
Is it possible to save Intel?

2025-08-24 12:01:10 UTC

- reply

@nprofile…6kk0 What do you use now?

2025-08-23 21:02:46 UTC

- reply

@nprofile…hszg @nprofile…u9u3 You can mute the person and demute him two weaks later.

2025-08-21 20:19:02 UTC

- reply

@nprofile…hszg 3) is more a property of first-order theories than a property of ZFC itself I would say.

2025-08-20 16:05:01 UTC

- reply

@nprofile…ca6t leibniz pensait la même chose. Sa fin de vie n'a pas été une réussite.

2025-08-20 15:03:49 UTC

Is it true that we can't rely solely on the RDRANDOM instruction of the INTEL processors?
It is so crazy.
We cannot trust anybody any longer.

2025-08-19 16:01:50 UTC

- reply

@nprofile…ca6t Bref il en sait rien quoi.

2025-08-19 12:01:44 UTC

Is the a LLM database computed only once or is it updated regularly? Let's say every week.

2025-08-18 11:49:21 UTC

Karenia mikimotoi: more horrific than "Jaws" and "the birds" together.
https://fishingworld.com.au/news/no-respite-for-sas-algae-bloom/

2025-08-17 20:12:56 UTC

- reply

@nprofile…l7n9 Beautiful but wrong. There is undeniably a genetic component.

2025-08-17 19:48:13 UTC

- reply

@nprofile…hszg "But not understood theoretically."
We agree.
But it certainly says something about what linguistics is.

2025-06-28 19:09:45 UTC

- reply

@nprofile…hszg The next step is ML by Milner in 1973. I guess it can be described as a language based on a type system.

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.

2025-05-10 20:04:46 UTC

- reply

@nprofile…hszg You meant "Hawking"?