Elektrine
Log in Register
Paige Chat Timeline Gallery Friends Email Drive DNS Private DNS Domains VPN Kairo Nerve
Remote

Josselin Poiret

@jpoiret@types.pl
mastodon 4.8.0-alpha.2+glitch
  • Open on types.pl

types and categories and algebra and geometry and free software

i use guix btw

avatar: cover art of Love is a Stream by Jefre Cantu-Ledesma

background picture by @VojtechStep@mathstodon.xyz

147 Followers
117 Following
15 Posts
Joined February 24, 2023
stalk me at:
https://jpoiret.xyz
gpg:
39248CD841C63CC336DCAF2F505E40B916171A8A
Open post
Josselin Poiret @jpoiret@types.pl
· 2mo ago

so, i've started hate-reading "the proof in the code" (my good friend Anna lent it to me), and the technical explanations about constructivism are so inaccurate it's funny:

"For example, mathematicians often try to answer questions like “Prove that every even number greater than 2 can be written as the sum of two prime numbers.” Called the Goldbach conjecture, this question is one of the most famous open problems in mathematics. From a constructivist point of view, the only way to answer the question would be to construct specific examples that prove or disprove it. Either compute every even number as a sum of two primes—which is impossible because there are infinitely many—or compute an example of an even number that violates the rule."

oh yes we can't prove goldbach constructively because we'd need to compute infinitely many numbers!!!!

1/4

51
7
24
1
Open post
Josselin Poiret @jpoiret@types.pl
· 2mo ago

Timothy Gowers on the Leiden declaration: https://gowers.wordpress.com/2026/07/26/thoughts-about-the-leiden-declaration/

basically "yes there are some issues with LLMs but they're getting better so these will soon be resolved, also access to LLMs might be a source of inequality but let's just really ignore that part because I am a spineless AI-booster".
Also "AI companies are doing this out of goodwill and are TOTALLY interested in providing benefits to maths reaearch, surely not trying to sell a product".

Combined with 0 discussion of ethical and environmental concerns, this was unfortunately expected but still disappointing.

Thoughts about the Leiden Declaration
Gowers's Weblog

Thoughts about the Leiden Declaration

Last September I went to a workshop at the Lorentz Centre in Leiden to discuss mathematics and AI with historians, philosophers, computer scientists, AI researchers, and mathematicians of several d…

34
2
19
1
Open post
Josselin Poiret @jpoiret@types.pl
· 2mo ago

Why can't the conference dinner just be a pizza in a park?

15
2
4
0
Open post
Josselin Poiret @jpoiret@types.pl
· 7mo ago

I'm starting to think that no generalist newspapers have proper tech journalists (from the Guardian https://www.theguardian.com/technology/2026/feb/26/how-to-replace-amazon-google-x-meta-apple-alternatives).

theguardian.com
10
0
1
0
Open post
Josselin Poiret @jpoiret@types.pl
· 8mo ago

a comedy in 2 acts

(can you believe searching within volumes is a paid feature?)

8
4
4
0
Open post
Josselin Poiret @jpoiret@types.pl
· 6mo ago
Replying to
@sophiehuiberts@mathstodon.xyz zref-clever is working well enough for me, and with zref-titleref you also get nameref behavior!
4
11
2
0
Open post
Josselin Poiret @jpoiret@types.pl
· 6mo ago

typo theory

4
0
1
0
Open post
Josselin Poiret @jpoiret@types.pl
· 5mo ago
Replying to
@juliengossa@social.sciences.re beaucoup de sportifs ne vivent pas non plus de leur pratique, et une médaille représente souvent un gain financier conséquent. Je ne pense pas qu'on puisse dire que c'est "bien pire", c'est juste comparable.
2
1
0
0
Open post
Josselin Poiret @jpoiret@types.pl
· 5mo ago

me when the stochastic machine reinforces my beliefs

2
0
1
0
Open post
Josselin Poiret @jpoiret@types.pl
· 5mo ago
Replying to
@mc@mathstodon.xyz unless the accompanying paper is well-written and has interesting insights, I would say no
1
0
0
0
Open post
Josselin Poiret @jpoiret@types.pl
· 6mo ago
Replying to
@sophiehuiberts@mathstodon.xyz I guess you are numbering your lemmas with the same counter as your theorems? Most packages that do “clever” things with references in TeX use the “reference counter name” \@currentcounter to infer additional information, but if you share counters between environments that won't be enough. There is a section specifically about how to make it work with lemmas (with code) in the zref-clever manual section 10.2, look for the “shared counter” paragraph. The idea is that you locally change the inferred meaning of the theorem counter in lemma environments.
1
1
0
0
Open post
Josselin Poiret @jpoiret@types.pl
· 6mo ago
Replying to
@sophiehuiberts@mathstodon.xyz small mistake on my part, \@currentcounter is LaTeX, not TeX
0
0
0
0
Open post
Josselin Poiret @jpoiret@types.pl
· 5mo ago
Replying to
@carloangiuli @jonmsterling count me in the "funext isn't natural" camp 😈
0
13
0
0
Open post
Josselin Poiret @jpoiret@types.pl
· 5mo ago
Replying to
@jonmsterling @carloangiuli oops, i guess it's getting late and i misread the word! In any case, I'm very much intensionally-minded: while you can't distinguish extensionally equal functions internally, when i prove that two functions are equal I do mean to say that they are implemented in the same way. Otherwise I would be proving that they are extensionally equal. If you add funext, this distinction isn't possible anymore, which mean you trade nuance for convenience. Don't get me wrong, convenience can be very nice, but I also like having this precise tool for specific applications!
0
11
0
0
Open post
Josselin Poiret @jpoiret@types.pl
· 5mo ago
Replying to
@JacquesC2 i don't think the author version would include any proof whatsoever then
0
1
0
0
Back
313k7r1n3
Elektrine

Tor hidden service

elekhj7afj4qnrr4yd3bkzslsyo5jgfxw3orgjkhlcxifueodybyiiad.onion

I2P eepsite

j6b6cyk6gjmepjih7jjadxgxvvf3lzzujljuu2v4biemzpg3naya.b32.i2p

Platform

  • Email
  • Chat
  • Timeline
  • VPN
  • DNS

Company

  • About
  • Contact
  • FAQ
  • Lite (no JS)

Legal

  • Terms of Service
  • Privacy Policy
  • Transparency Report
  • Report Abuse
  • Warrant Canary
  • VPN Policy

Support

  • support@elektrine.com
  • Report Security Issue
Mail client setup IMAP mail.elektrine.com:993 POP3 mail.elektrine.com:995 SMTP mail.elektrine.com:465
© 2026 Elektrine. All rights reserved. Server: 20:05:59 UTC