Guillaume Pitel's Avatar

Guillaume Pitel

@pitzeglide.bsky.social

Computer science & machine learning, Vélotaf & mobilités. CTO Babbar. Search engines, information retrieval, semantics, probabilistic data structures,... Also science !

145 Followers  |  846 Following  |  10 Posts  |  Joined: 13.11.2024  |  2.512

Latest posts by pitzeglide.bsky.social on Bluesky

Post image

En Italie aussi, ils concourent dans les infrastructures cyclables ubuesques 🤪

15.06.2025 17:35 — 👍 38    🔁 3    💬 5    📌 0
Post image

En région parisienne, les flux importants d'usagers à vélo ne sont plus réservés à la capitale. Ici, en proche banlieue, la coulée verte offre une infrastructure sécurisée et agréable qui favorise ce mode pour les trajets quotidiens.

24.03.2025 18:26 — 👍 25    🔁 2    💬 0    📌 0

Always. Every time. Some idiot will pop up with "I guess you hate poor people then!"

But then you check their feed and they are far from poor.

21.03.2025 17:53 — 👍 24    🔁 9    💬 3    📌 0
Preview
Hungary Bans LGBTQ Pride Events, Approves Facial Recognition to Track Attendees Thousands protested the ban outside Hungary’s parliament, blocking traffic and defying police orders to disperse.

if you think this can't happen here, think again. truthout.org/articles/hun...

20.03.2025 13:20 — 👍 268    🔁 135    💬 5    📌 14
Preview
Search LibGen, the Pirated-Books Database That Meta Used to Train AI Millions of books and scientific papers are captured in the collection’s current iteration.

Meta used at least 16 of my books, and numerous articles, to help train the AI it will use to make billions.

Authors, search your name here:

www.theatlantic.com/technology/a...

20.03.2025 12:16 — 👍 3458    🔁 2098    💬 285    📌 2132
Preview
The Unbelievable Scale of AI’s Pirated-Books Problem Meta pirated millions of books to train its AI. Search through them here.

Meta considered licensing books to train AI—but opted instead to pirate LibGen, a database that currently contains more than 7.5 million books and 81 million research papers, Alex Reisner writes.

20.03.2025 12:55 — 👍 929    🔁 468    💬 34    📌 191
Preview
Comité d'évaluation du carrefour Pompadour La transformation du carrefour Pompadour reste un sujet sensible. Mais est-ce vraiment à cause de l'aménagement cyclable ?

De 150 à 2100 cyclistes par jour : l'aménagement du carrefour Pompadour a nettement amélioré les conditions de déplacement à vélo dans le Val-de-Marne !

➡️ il faut désormais rendre plus sûres les axes le reliant pour que les Val-de-marnais·es prennent massivement leurs vélos !

20.03.2025 18:36 — 👍 32    🔁 12    💬 3    📌 0
Preview
Paris: Paul Varry dreamed of a cycling revolution. Then an SUV crushed him Paul Varry's death has triggered a debate in Paris over the dramatic expansion of cycle infrastructure.

Le 15 octobre, Paul a été écrasé, à l’arrêt et sur une piste cyclable, par le conducteur d’un SUV. Il est mort.

Les médias français en ont parlé pendant quelques jours, et puis… plus rien.

Cet article de la BBC est le premier que je lis à ce sujet depuis des mois.

23.02.2025 17:04 — 👍 135    🔁 64    💬 3    📌 1
Post image 11.02.2025 12:00 — 👍 32    🔁 10    💬 0    📌 0
Preview
Deep dive into the architecture of the Babbar Crawler: The Crawl Policy At Babbar, we try to crawl the web like a search engine. Of course, we don't have the same computing resources or bandwidth as Google, Bing or other

Deep dive into Babbar's crawl policy central.yourtext.guru/deep-dive-in... : in this article I explain how we maintain a target crawl rate while prioritizing some websites/URLs, all while trying to be gentle with web servers. I explain the metrics and indicators used to boost/unboost pages.

04.02.2025 10:15 — 👍 0    🔁 0    💬 0    📌 0

For anyone who’s written a scientific paper, this makes absolutely no sense unless your goal is to censor science.

03.02.2025 17:14 — 👍 447    🔁 162    💬 9    📌 1
Post image

This is what the government did with 120K+ Japanese Americans in 1942.

I know. I was there in those camps.

31.01.2025 19:11 — 👍 100559    🔁 32201    💬 3530    📌 1555
Preview
Flavors of overfitting Contextual overfitting and the Soviet Tank Problem

Many still believe in overfitting. Here’s a post trying to pin down what they mean and reach some common ground. www.argmin.net/p/flavors-of...

31.01.2025 15:29 — 👍 32    🔁 5    💬 3    📌 1

Le bashing Français en IA me fatigue 🫠
Lucie n’est pas l’alpha et l’oméga de nos capacités dans l'intelligence artificielle 🙅

Tout le monde a oublié qu'en France, nous avons entre autres :
- Mistral AI
- Kyutai
- Photoroom
- H
- LightOn
- …

On a de quoi être fiers, bordel ! 🐓

29.01.2025 18:24 — 👍 6    🔁 4    💬 0    📌 0
Post image Post image Post image Post image

“In 150 characters or less” by @nikitagill.bsky.social

21.01.2025 20:57 — 👍 3801    🔁 1043    💬 41    📌 56

D'un autre côté on peut le voir comme un effort pour ne pas être biaisé par leur entourage et avoir un échantillon plus représentatif !

16.01.2025 09:58 — 👍 0    🔁 0    💬 1    📌 0
Preview
Cycliste morte percutée à Rouen: une association d'usagers du vélo appelle à un rassemblement ce samedi L'association d'usagers Sabine Rouen Vélo appelle à un recueillement samedi 18 janvier en fin de matinée sur l'esplanade du musée des Beaux-arts. Le collectif dénonce aussi les dangers auxquels les cy...

"Cycliste morte percutée à Rouen: une association d'usagers du vélo appelle à un rassemblement ce samedi" @bfmtv.com www.bfmtv.com/normandie/cy... w/ itw @grima-guillaume.bsky.social

16.01.2025 09:48 — 👍 2    🔁 1    💬 0    📌 0
Post image

Ha! So, we have a comment board, and user victual_brother meme-ified a joke of mine

15.01.2025 15:14 — 👍 306    🔁 29    💬 20    📌 0
Preview
Backlink acquisition methodology For SEO, backlinks are a historical topic. We often talk about the 3 pillars of SEO, mentioning authority…

New Year, new article.
Today it's about Backlink acquisition, and I have to warn you, it's quite a long topic.

Feel free to go and read the other articles of the blog, many very interesting topics from my coworkers!

central.yourtext.guru/backlink-acq...

15.01.2025 11:03 — 👍 1    🔁 1    💬 0    📌 0
Preview
Main Content Extraction from Web Pages When browsing websites, not all displayed content is relevant to a search engine or a user. Some sections…

Aurélien, who just started a PhD with us, shares insights on extracting main content from web pages. He compares heuristic, ML, and visual approaches, assessing their strengths and challenges.
central.yourtext.guru/main-content...

06.01.2025 17:01 — 👍 0    🔁 1    💬 0    📌 0
Preview
Asynchronous crawling methods: focus on Reactor / Spring What is “asynchronous” programming? “Asynchronous” programming, by definition, contrasts with classic imperative or functional programming, which is inherently…

Final article of the festive break is by Maxime.
Max explores asynchronous programming and discusses its impact on web crawling. It's about Java's Reactor/Spring WebFlux and how reactive programming tackles back-pressure and resource management.
central.yourtext.guru/asynchronous...

06.01.2025 17:05 — 👍 2    🔁 1    💬 0    📌 0
Preview
GPUs: The Basics to Understand Their Performance Today, I’m going to talk to you about GPUs (Graphics Processing Units). Of course, you’re all familiar with…

It’s been quite a while since I last wrote a technical outreach post, but working on the new yourtextguru blog has renewed my enthusiasm.

Today's piece is about how GPUs work and why they are so well-suited for neural networks (yes, AI! 😅)

central.yourtext.guru/gpus-the-bas...

13.01.2025 14:42 — 👍 5    🔁 3    💬 0    📌 0

In following articles I'll describe how we store the augmented web graph, how we compute the metrics, how we decide which page to crawl, how we compress our data.

14.01.2025 09:26 — 👍 0    🔁 0    💬 0    📌 0
Preview
How does a high performance web crawler work? The Babbar case In this article, we are going to introduce the high level architecture of the Babbar SEO crawler, Barkrowler.…

First article of a series about the Babbar crawler : central.yourtext.guru/how-does-a-h... I wrote about the architecture of a webscale continuous crawler.

14.01.2025 09:26 — 👍 0    🔁 0    💬 1    📌 0

Talking about my favorite piece of tech is always a great pleasure ! Crawling the web from a full scale crawler's point of view.

10.01.2025 10:06 — 👍 2    🔁 1    💬 0    📌 0
Preview
Mullenweg’s WordPress Pause Triggers Unexpected Complications Matt Mullenweg's pause of WordPress services is causing unintended consequences as community grapples with the fallout from his decision. via @martinibuster.bsky.social #wpnews #WordPress

Matt Mullenweg's pause of WordPress services is causing unintended consequences as community grapples with the fallout from his decision. via @martinibuster.bsky.social

#wpnews #WordPress

20.12.2024 19:51 — 👍 2    🔁 1    💬 0    📌 0
Preview
OpenAI Search Leader Departs After Less Than a Year Shivakumar Venkataraman, a longtime Google search advertising executive who joined OpenAI earlier this year to help lead the development of search and artificial intelligence for enterprise customers,...

Former Google search ads exec, who joined OpenAI earlier this year to lead development of search and AI for enterprise, has left the co.

Scoop: www.theinformation.com/briefings/op...

20.12.2024 19:46 — 👍 3    🔁 2    💬 0    📌 0
Preview
SearchGPT vs. Google vs. Bing: Search Results Review To compare the search results in ChatGPT to Bing and Google, Dan looked at a number of queries to understand the possibility and limitations of SearchGPT

Before we get into the holidays, we want to take a moment to celebrate the top contributors for November.

Top Pageviews

🏆 @taylordanrw.bsky.social - SearchGPT vs. Google vs. Bing: Search Results Review www.searchenginejournal.com/chatgpt-sear...

20.12.2024 19:55 — 👍 7    🔁 3    💬 1    📌 0
Video thumbnail

Coloniser Mars, si c'était possible... serait-ce vraiment une si bonne idée ? 🤔
Et s'il était temps de repenser nos imaginaires sur la colonisation spatiale ?

Teaser de ma dernière vidéo ("Mars, la planète B ?") avec @sebastiencarassou.bsky.social

👉 Vidéo complète : 👀 youtu.be/N8G-KGqn2Lg 👀

20.12.2024 14:38 — 👍 76    🔁 22    💬 5    📌 2

Bonjour les éditeurs de livres français ! Vous savez où me joindre !

20.12.2024 15:53 — 👍 29    🔁 2    💬 1    📌 2

@pitzeglide is following 20 prominent accounts