
Head of department @gobelins_paris https://xitter.cf/t.co/lJVOhWyWXz | https://xitter.cf/t.co/MN8YYx1DXY | @Gephi plugins | Java dev ❤️
Paris, France
Joined April 2009
- Tweets22.4K
- Following2.8K
- Followers3.3K
- Likes9.9K
Clement Levallois / [email protected] retweeted
Le Web a découvert cette semaine qu’une partie de ses correctifs PHP passait par le travail bénévole d’Ondřej Surý, qui prépare des paquets pour Debian et Ubuntu.
Dans un nouvel entretien à la BBC, il raconte avoir reçu plus de 10 000 dollars de dons en quelques jours, soit environ une année habituelle de contributions.
bbc.com/news/articles/cw30ej…
Clement Levallois / [email protected] retweeted
So basically, today PHP got two critical bugs that let anyone run code on your server without even logging in, which is bad, and then the details leaked before the patch was even out, which is worse, and it works on WordPress and MediaWiki and Magento and Joomla and Laravel, and Censys says like 1.5 million servers are running a version with no fix, so the NSA was like "take your system offline," and the FBI was like "take your system offline," and CISA and the UK and CERT-EU were all like "take your system offline," and then the Prime Minister of Japan, the actual Prime Minister, posted at 5:55 in the morning telling 48.8 million people to take their system offline, so everyone took their system offline, like the Dutch cybersecurity agency took its own website offline, and the Dutch parliament went offline, and the Port of Rotterdam went offline, and Finland and France and Denmark started pulling government sites, and Wikipedia went down, and Archive.org went down, and Asana and GOG and Something Awful went down, and the University of Vienna's webmail went down for some reason, and that's just PHP. Because GitLab also dropped a CVSS 10.0, which is the highest score they give out, for a bug that's been there since 2014, so that's every self-hosted GitLab for twelve years, and then LiteLLM 1.107.0 got hijacked and shipped a credential stealer and a remote access trojan in every single wheel and the Docker images, including latest, and you didn't even have to import it, just installing it was enough, and it's the second time this year, and then Railway paused every deployment on the entire platform because someone got into the environment variables, so now everyone has to rotate every secret they've ever stored there, and TiDB Cloud said someone was walking around inside with stolen credentials for basically a month and an extortion group says they took 10.6 terabytes, and Zeabur got breached for the second time in two months and is just shutting down, like the whole company, and Node.js patched remote code execution and a TLS bypass, and CISA says Plex is being actively exploited, and someone took over 300 Tor exit relays to steal SSH passwords, and SafePay ransomware has hit over 600 critical infrastructure orgs, and India's biggest bank filed a data loss incident, and Japan has been in an official state of emergency in cyberspace since October 9, which honestly explains the 5:55 AM post. Oh, and eMudhra, a certificate authority that's issued over 60 million certificates, whose entire job is proving websites are who they say they are, was supposed to be using private keys that are random numbers over a hundred digits long, and one of them was 4. Just 4. And the others were like 66 and 237. Anyway, it's Wednesday.
Too bad I don't write code by hand anymore. Structured concurrency targeted to get out of preview in JDK 28 (in March 2027) is such a beautiful thing, I'd love to use it myself rather than via an agent #java
openjdk.org/jeps/543
Clement Levallois / [email protected] retweeted
On the small companion website, you can view higher-resolution images, access data from the book, and interact with a visualization of the archive network (using @Gephi Lite, thanks to @ouestware): intellectualcooperation.org/…
Clement Levallois / [email protected] retweeted
🌟Testing @Gephi Lite v1.1! (2026)
Try it now!
👉Learn more: gephi.wordpress.com/2026/10/…
cc @ouestware
#Gephi #NetworkScience #DataVisualization #SocialNetworkAnalysis
Clement Levallois / [email protected] retweeted
Beaucoup ne se rendent pas compte de ce qui se passe réellement avec l’IA 🤯
Hugo Duminil-Copin (médaille Fields 2022) : « La résolution de 350 problèmes majeurs par OpenAI, c'est comme si des camions m'avaient écrasé ! Tous les problèmes que je mentionnais dans mes exposés, articles et demandes de financement, ont été résolus. »
Clement Levallois / [email protected] retweeted
We just released Gephi Lite v1.1 🚀
The @ouestware team worked hard on geographic projection, color palettes and layout experience. Give it a try and have a look at their report on our blog: gephi.wordpress.com/2026/10/…
Clement Levallois / [email protected] retweeted
Announcing a New HNR Bibliography: Over 1,300 Publications on Historical Network Analysis historicalnetworkresearch.or…
(full report here: doi.org/10.5281/zenodo.23070…)
I tend to disagree. Up to fairly advanced software engineering tasks, AI lowers difficulty *very much*, in my practical experience.
Past a level of complexity, yes indeed difficulty reappears. It is compounded by the need to master new, unchartered AI agentic workflows.
I think the "difficulty" of software engineering is essentially constant no matter what abstraction level you move to, because human cognition adapts to new tools until it can fully utilize itself.
Tools are only affordances, not a magic wand that makes work disappear.
Great software engineering was immensely challenging before. It is still immensely challenging now, despite very different workflows.
Night time project: a dataviz of French elites on Wikipedia:
nocodefunctions.com/entourag…
The crawl is ongoing, you can see new bios arriving...
#socialnetwork #madewithJEV cc @CompleteSkeptic
Clement Levallois / [email protected] retweeted
🤯 Ce week-end, j'ai utilisé les données ouvertes du @education_gouv pour construire un outil de visualisation des résultats de la certification #PIX... c'est riche d'enseignements, surtout quand on s'intéresse à la jeunesse et au numérique !
On a le score par établissements depuis 2025 (et les paliers de 0 à 3 pour les années antérieures).
La première carte représente la distribution géographique France entière.
La deuxième carte représente la distribution géographique du top 50 des lycées publics "Genéral et Technologique" avec les 5 premiers que sont :
1- Lycée Louis-le-Grand
2- Lycée Clemenceau (Nantes)
3- Lycée Franco-Allemand de Buc (LFA)
4- Lycée Hoche
5- Lycée naval (Brest)
Analyse détaillée sur mon blog à venir demain. Spoiler : 98.75% de certifiés, mais des compétences numériques inégales (c'est le titre).
Dans l'outil, vous pouvez trier par académie, par ville, établir un top 10, un top 50, comparer à la moyenne nationale, etc.
Si l'accès à l'outil vous intéresse, laissez un commentaire ci-dessous. Je posterai le lien en commentaires et vous aurez ainsi une notification pour y accéder.
cc @datagouvfr, @EdouardGeffray
Clement Levallois / [email protected] retweeted
Take me somewhere expensive…
Clement Levallois / [email protected] retweeted
Playing DOOM on a Stirling engine powered ESP32 "computer".
[📹 PickentCode]
Clement Levallois / [email protected] retweeted
LLMs can now talk to each other without words.
Chinese researchers open-sourced a new paradigm that lets LLMs communicate without generating a single word.
It’s called Cache-to-Cache (C2C) communication.
right now, when multiple ai agents work together, they are forced to translate their internal "thoughts" into human text tokens just to pass a message. this loses rich semantic meaning and causes massive token-by-token latency.
So, instead of spitting out words, c2c uses a neural network to directly project and fuse the source model's "kv-cache" right into the target model. it is pure, direct semantic communication.. they even added a learnable gating mechanism to select exactly which layers benefit most from the cache transfer.
the benchmark results are actually crazy:
- avoids all intermediate text generation latency
- accuracy jumps by up to 14.2% compared to individual models
- beats traditional text-based agent communication by over 5%
- delivers a massive 2.5x speedup in overall speed
we are literally watching llms bypass human language to build their own silent, high-speed neural network..
Clement Levallois / [email protected] retweeted
Three.js conf time! Look how amazing this speaker badge is!
It has games!!
Clement Levallois / [email protected] retweeted
Learning that AI might kill us all doesn't sound as scary when a cute boy band sings it you 🥰
Clement Levallois / [email protected] retweeted
My whole TL today.
























