Gro-Tsen Profile picture
Mar 15, 2021 5 tweets 2 min read Read on X
Le poussinet est en colère (et je suis d'accord avec lui).
D'autres sont en colère aussi (et je suis toujours d'accord):
D'autres sont plutôt sarcastiques:

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Gro-Tsen

Gro-Tsen Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @gro_tsen

Jul 22
Among the very many harms that LLMs are causing to mathematics is the fact that any idiot armed with a ChatGPT/Claude/whatever “pro” subscription thinks they can produce interesting math with their stupid queries. Or even that the output is necessarily correct.
It used to be super easy to detect math crackpot writing. Now they write stuff that is superficially extremely plausible, highly technical, and sometimes (but not always) correct, so it's become very hard to detect, even though it's just as idiotic.
So now we have a gazillion idiots asking ChatGPT, Claude or whatever, to improve or expand upon the refutation of the Jacobian conjecture, and thinking it makes them smart.

Hint: if last week you didn't know what the Jacobian conjecture was, there's no way you can add value.
Read 6 tweets
Jul 20
Let me expand upon this 🔽 and say something about AIs, mathematics, proofs, conjectures, problem-solving, goalposts, paperclips, genii and, more importantly, Campbell/Goodhart's law. 🧵⤵️ •1/38
So, first, the reason we are interested in proofs, in mathematics, isn't so much that we want to be “really sure” something is true (yes, there is that, but it's far from the only reason, and probably not even the main one): it's that they give us understanding. •2/38
THE PURPOSE OF MATHEMATICS IS NOT TO PROVE THINGS OR TO SOLVE PROBLEMS: it is to gain a better understanding of the mathematical universe and how it works (either per se → pure math, or to use this understanding elsewhere → applied math). •3/38
Read 38 tweets
May 28
A linguistic point about French vs English:

I recently retrieved an item from an Amazon locker, and the locker thanked me with “merci d'utiliser Amazon locker”.

I would argue this is a bad translation of “thank you for using Amazon locker”. Let me explain. 🧵⤵️ •1/17
Just to be clear, there is nothing grammatically wrong here: “merci d'utiliser Amazon locker” is indeed correct, and if I were to back-translate it to English, “thank you for using Amazon locker” would be reasonable. My issue is with the choice of tense. •2/17
(I mean the choice of tense of what is in French an infinitive “utiliser” and in English a gerund “using”. The über literal translation of the French phrase is “thank-you of to-use Amazon locker”. I'm not discussing why French uses an infinitive and English a gerund here.) •3/17
Read 17 tweets
May 10
🔽 This is highly alarming, so let me try to explain what it's about in a manner understandable to laypeople:

‣ reCAPTCHA is a service to prevent bots from accessing Web sites. You probably know it as “click on all squares containing bicycles”.

… •1/5
‣ reCAPTCHA has become owned by Google. It is less and less efficient at preventing bots from getting through.

‣ So Google decided to change (retroactively!): to prove that you're a human, you will now need to use a Google account on a phone with Google services.

… •2/5
‣ Because of course all human beings have a phone with a Google account, right? If you don't, you're probably not human.

‣ By doing so, Google is establishing itself as the gatekeeper of a huge chunk of the Web, and will collect tons of info about what you browse.

… •3/5
Read 5 tweets
Aug 15, 2025
😩 Chaque année, le 15 août, paraît le «classement de Shanghaï», des services de comm' imbéciles se gargarisent de leur place, et chaque année je dois réexpliquer pourquoi ce classement ne veut rien dire et qu'il est même nuisible d'en parler et de le regarder. •1/21 🧵⤵️ Capture d'écran d'un post sur Bluesky de l'Université Paris-Saclay, qui commence: «Classement de Shanghai 2025 : l’Université Paris-Saclay 1ère université française, toujours dans le top 20 mondial!»
Commençons par l'évidence: le but de ce classement est de mesurer les critères que les autorités chinoises considèrent comme utiles pour piloter le fonctionnement de leur système universitaire. Ce n'est pas exactement neutre: … •2/21
… ce que je veux dire par là, c'est que même si la Chine ne se place pas elle-même en premier de ce classement, l'attention portée à ce classement promeut les critères que la Chine veut voir utilisés pour évaluer le monde universitaire. (À commencer par la taille.) •3/21
Read 21 tweets
May 30, 2025
Probably one of the most monumental achievements of human mathematics is the Classification of Finite Simple Groups (“CFSG”), a proof whose current incarnation is about 15 000 pages long spread across many articles by different authors. What is this all about? 🧵⤵️ •1/92
So, I'm not going to define what a “group” is formally (see for an attempt at an explanation), but I like to describe it informally as “a form of abstract symmetry” that an object (mathematical or otherwise) can have. •2/92
Most people are at least aware of geometrical symmetries: “in what ways can you rotate a cube by keeping its overall position fixed?” (e.g., rotate by 90° around a face, or 120° around a diagonal) is an example of a (finite) group (here with 24 elements or “order 24”), … •3/92
Read 92 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(