Florent Storme (@FlorentStorme) profile picture

Docteur en Physique. Data Scientist NLP/time series.

Ile-de-France, France
Joined December 2018
Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)Photo by Florent Storme (@FlorentStorme)
Florent Storme retweeted
« Je n’ai plus mes règles depuis trois ans » : Océane Mahé se confie sur la gestion de son cycle menstruel sur le Tour de France femmes ➡️ l.leparisien.fr/3a8T
20
9
5
39
141,356
Florent Storme retweeted
Claude Science is incredible. I gave it some sequencing data, and in 8 hours it did a full analysis, generated figures, wrote a paper, submitted it for publication, got rejected, revised and resubmitted, got rejected again, it is now applying for positions in industry
102
663
59
9,011
498,953
Florent Storme retweeted
My grandpa used to say: “Every piece of software dies twice. One time when somebody uses it for the last time. A second time, when it’s rewritten in Rust.”
Flow is now written in Rust 🦀 The migration from OCaml to Rust brought up to 2x faster type-checking on large codebases.
52
87
7
1,737
64,428
Lack of electrical power.
Stupid exercise: Assume AGI and even ASI is imminent. Now, imagine it winds up not changing the world nearly as much as anyone thought, and the reason seems very stupid, but in retrospect, makes sense. What is the reason?
1
53
Florent Storme retweeted
i'm sorry mom
22
1,586
106
20,011
898,043
Florent Storme retweeted
Filed in "photos which are terrifying if you know what they are, nonthreatening otherwise."
1,751
3,620
803
76,207
3,135,696
Florent Storme retweeted
Software engineers don’t just find cushy $200k+ jobs, they create them! 1. Be responsible for boring but critical code 2. Over the years, overcomplicate it 3. Be the only one who understands it 4. Costly to remove code; hiring replacement takes too long Job security guaranteed.
145
209
51
4,831
463,830
Florent Storme retweeted
37
1,466
43
25,008
481,507
Florent Storme retweeted
artificial intelligence
6
84
4
1,230
34,157
Florent Storme retweeted
cloud was a mistake
3
37
3
726
21,784
Florent Storme retweeted
i’ll never learn
22
261
8
3,501
172,153
Florent Storme retweeted
> still writing the same code since 2012. > public class QuantumFactoryBuilderFactorySingletonProxyAdapter > makes 300k/year
40
214
37
5,126
298,706
N'empêche Mark Cavendish pensait mettre un terme à sa carrière il y a 4 ans, il a connu la dépression, de grosses chutes mais a persévéré et a réussi à aller chercher sa 35ème victoire sur le Tour à 39 ans et ça c'est beau 🫶
3
42
3
987
47,963
Florent Storme retweeted
> paid 60k/year, worked at his company for 25 years > IDE takes 30 minutes to open > compiles for 2 hours just to segfault > std::vector<int>::iterator::pointer::element_type > just one more typedef bro > 60 lines stacktrace for a missing ";" > debug build: 4GB
70
234
34
5,591
376,163
As per recent estimates, about 20.3% of pet cats and 18.1% of owners were seropositive for T. gondii. 6/10
8
39
5
631
201,139
C'est parce que le bon développeur il a créé un ticket Jira avant de debug.
Tu vois, y'a le mauvais développeur, y voit un truc qui bugue y corrige. Le bon développeur lui, y voit un truc qui bugue, y corrige, mais c'est un bon développeur quoi.
38
🌟 Evidemment ! Mathieu van der Poel décroche le titre de 𝗖𝗛𝗔𝗠𝗣𝗜𝗢𝗡 𝗗𝗨 𝗠𝗢𝗡𝗗𝗘 de gravel à Louvain #Leuven2024 #LesRP
3
52
4
557
40,802
Florent Storme retweeted
Sur #ParisTours, Christophe Laporte (Visma-Lease a Bike) a repris Mads Pedersen (Lidl-Trek) avec Mathias Vacek (Lidl-Trek). Le Danois avait attaqué à 70km et semble désormais dans le dur. Les visages des coureurs sont marqués par la boue, projetée par les chemins de vigne.
3
5
3
206
65,874
What happens if your CPU gets something wrong? If it wakes up one day and decides 2+2=5? Well, most of us will never have to worry about that. But if you work at a company the size of Google, you do, which is why this paper on "mercurial cores" is so fascinating. What the authors report--and supposedly this is common knowledge at the hyperscalers--is that a couple cores per several thousand machines are "mercurial." Due to subtle manufacturing defects or old age, they give wrong answers for certain instructions. These can cause all sorts of impossible-to-diagnose issues. Some rare problems at Google that were traced back to bad CPUs include: - Mutexes not working, causing application crashes - Silent data corruption - Garbage collectors targeting live memory, causing application crashes - Kernel state corruption causing kernel panics What makes CPUs go bad? It's very hard to tell. The authors posit that issues are becoming more frequent as CPUs get more complex, but there aren't solid numbers behind that. There are certainly strong relationships between frequency, temperature, voltage, and bad CPU behavior--most mercurial CPUs only cause problems under very specific conditions, but those conditions vary from CPU to CPU. Age is another source of problems, as older CPUs are more likely to exhibit problems. Bad CPUs are an especially serious problem because they're very hard to detect. If cosmic rays flip bits in storage or on the network, that can be detected through error coding. But there's no analogy for a CPU that allows cheap online verification of its correctness. Instead, the best detection techniques involve monitoring for symptoms. If a core exhibits exceptionally high rates of process crashes or kernel panics relative to its fellows, that's a strong indication something is wrong with it. For the most critical applications, the authors propose triple modular redundancy--redoing each of its computations on three cores and majority-voting a reliable result. More than anything, this paper is a call to action--letting everyone know that CPUs can fail. So now, if you ever find a bug you can't diagnose, you can blame the CPU! 🙂
155
1,277
154
8,435
760,424