ex- ESA, NASA and ECB. @creativedlab graduate for making @SpaceDAO_ai for humanity. Building @celestical_eu to buy aws and azure for 1 dollar
Joined May 2012
- Tweets3.2K
- Following776
- Followers743
- Likes3.2K
Dr. Red Boumghar retweeted
On y est, ça commence. En Chine, les chiens robots sont désormais pompiers.
Unitree et DEEP Robotics déploient des robots quadrupèdes dans les casernes chinoises. En opération réelle.
Le Unitree B2 "Fire Rescue" est déjà en service à Qingdao. Le DEEP Robotics X30 a été déployé dans le Hunan.
Ce qu'ils font :
→ Entrer dans des bâtiments en feu à la place des pompiers humains
→ Monter des escaliers (marches jusqu'à 40 cm)
→ Projeter de l'eau ou de la mousse à 60 mètres avec un débit de 40 litres/seconde
→ Détecter les gaz toxiques et mesurer les températures en temps réel
→ Transmettre de la vidéo live au poste de commandement via 5G
→ Porter jusqu'à 85 kg d'équipement
Autonomie de plus de 10 km. Vitesse max : 5 m/s. Batteries interchangeables en quelques secondes, même sous la pluie.
Quelqu'un pour faire pareil en France ?
Dr. Red Boumghar retweeted
Réflexion plus politique : en soi, une preuve d'âge sécurisée en ligne est une bonne chose *si* et seulement si son usage est consenti. C'est là le problème : on nous la balance dans les gencives en nous forçant à l'utiliser là où elle enlève des droits au lieu d'en apporter.
Dr. Red Boumghar retweeted
🚨 Element[.]io is experiencing a worldwide outage due to "legal reasons."
Element is a free and open-source instant messaging client based on the Matrix protocol. It provides secure, end-to-end encrypted communication for individuals, teams, and organizations.
Dr. Red Boumghar retweeted
🇫🇷 La start-up française Otrera développe des petits réacteurs modulaires (SMR) à neutrons rapides de 4ᵉ génération capables de recycler les déchets nucléaires existants.
🏭 Elle investit 45 millions d’euros pour construire une usine de 7 000 m² en Normandie, avec début des travaux en 2027 et production des composants critiques dès 2029.
📅 Le premier démonstrateur de ce réacteur 10 fois plus petit et moins cher qu’un EPR est prévu pour 2032, s’inscrivant dans la stratégie nationale de relance des SMR.
💼 Le projet devrait créer 200 emplois directs sur le site et jusqu’à 600 emplois au total, grâce à l’écosystème industriel historique de la Normandie en chaudronnerie et soudure nucléaire.
🏛️ La Région Normandie entre au capital d’Otrera et apporte 1,5 million d’euros de subvention.
lejournaldesentreprises.com/…
Dr. Red Boumghar retweeted
This is more than cotton, it's a movement. 🌍 A movement for transparency, sustainability, fairness, accountability. Join us. Follow us in the journey from farm to fabric. octave-space.com
#FutureOfFashion #CottonEthics #BlockchainForGood #OCTAVE
Dr. Red Boumghar retweeted
it's time to drop three new #opensource robotic hands! this time with tactile sensors! Tweak it, 3D print it, and use them in your robotics and physical AI research! Here are some wild examples ↓↓↓
Dr. Red Boumghar retweeted
someone built a $96 3D-PRINTED MANPADS rocket that recalculates its mid-air trajectory using a $5 sensor and piano wire
its called Project Canard
it integrates with distributed camera nodes to triangulate airborne targets and update flight paths in real-time
it proves the barrier to advanced hardware has completely collapsed, moving precision weapons from defense labs to consumer garages
the entire launcher and interceptor frame is 3D printed in PLA and runs off a standard off-the-shelf ESP32 microcontroller
it even spins up a local Wi-Fi network so you can monitor live telemetry and arm the system directly from your laptop
Dr. Red Boumghar retweeted
Pendant que beaucoup de monde continue à crier au drame écologique de l'IA en croyant qu'un prompt consomme des litres et des litres d'eau, les modèles qui tournent entièrement en local se développent à grande vitesse et atteignent déjà des performances très intéressantes.
Today we release LFM2.5-1.2B-Thinking, a reasoning model that runs entirely on-device. What needed a data center two years ago now runs on any phone with 900 MB of memory.
> Trained specifically for concise reasoning
> Generates internal thinking traces before producing answers
> Enables systematic problem-solving at edge-scale latency
> Shines on tool use, math, and instruction following
Dr. Red Boumghar retweeted
Okay team. Here it is. A 28 m resolution global map of the planet Mars, viewable on Google Earth. The starter version with just 271,790,899,200 pixels. One of the most beautiful things I have ever made. Please, enjoy!
drive.google.com/file/d/19So…
Dr. Red Boumghar retweeted
A perfect liftoff and smooth flight to orbit for Vega-C carrying five satellites to orbit.
🚀 Flight #VV27 liftoff occured at 23:03 local time 25 July.
🛰️🛰️🛰️🛰️ CO3D will map our planet in 3D
🛰️MicroCarb will monitor carbon dioxide from space
ALT A night-time liftoff of Vega-C clearing the launch pad
ALT A long-exposure photo of spectators watching the Vega-C VV27 launch from a beach nearby.
Dr. Red Boumghar retweeted
The Flare Forge Week is coming!
Join us for an exciting week of daily livestreams as talented builders showcase powerful use cases on @FlareNetwork.
Learn how to build with FAssets, FDC, FTSO and Flare's tooling. 🛠️
Details:
21–25 July | 13:30 UTC daily
📍 Streaming here on X.
Participate and help shape the future of data-rich dApps on Flare.
Dr. Red Boumghar retweeted
BREAKING: Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all.
They just memorize patterns really well.
Here's what Apple discovered:
(hint: we're not as close to AGI as the hype suggests)
Dr. Red Boumghar retweeted
Amazing pitching session for our #EuropeanCloud at @joinstationf with 120 people gathered for the startups incubated at @EDHEC_BSchool 's entrepreneurs programme.
Dr. Red Boumghar retweeted
Limp: "very confident" we can get Blue Moon Mk 1 lander on the Moon this year.
Dr. Red Boumghar retweeted
You probably know I'm doing volunteer Starlink repairs, which might look great. But I can't do many repairs; I still have my daytime job, other projects, and research activity. I decided it was much more important to share knowledge.
We currently have dozens of service centers and rework labs here in Ukraine.
Today, I would like to tell you more about this.
My friend runs the first in Ukraine (most likely in the world) and the biggest service center specializing in Starlink.
He said this wouldn't have been possible without my research, articles, and support.
I would say that all of this is impossible without the people doing all these repairs and working around the clock. Deepest respect to them all.
Currently, they are doing almost 100 Starlink repairs per week!
Dr. Red Boumghar retweeted
Satellites will one day run their own AI agents for collision avoidance, but there is a long list of problems to be solved: data accuracy, standardization, reliability, integration with existing systems, international cooperation, etc. Let's get to work! #SpaceDAO #AI #Blockchain
Dr. Red Boumghar retweeted
Attending #WebSummit next week? Connect to @redotics, founder and CEO of @parametry_ai
Let's meetup for a pastel de nata + awesome #space governance chat! ☕ 🥧 🚀 😀 A lot is happening at @SpaceDAO_ai and Red can tell you all about it. See you in #Lisbon! #OrbitalMissions #DAO
Dr. Red Boumghar retweeted
On our side in the last six month we've on boarded the first space sustainability NGO (🥰 @Cosmos4Humanity ) and presented at the Space Traffic Management conference 2024 to White House and DoD representatives. Decentralization is seen as a good solution for global collaboration.
4/8: What progress have we made?
We’ve supported ecosystem growth in Decentralized Science and Space with @gitcoin Grants, verified concepts of space marketplaces with @CopernicSpace, and ran our first iteration for incubating Space DAOs with @SpaceDAO_ai and @SpaceroxDAO
Dr. Red Boumghar retweeted
1/ How quickly are state-of-the-art AI models growing?
The amount of compute used in AI training is a critical driver of progress in AI. Our analysis of over 300 machine learning systems reveals that the amount of compute used in training is consistently being scaled up at 4-5x/year.
Highlights in thread 🧵
