Saviez-vous que ChatGPT peut concevoir des circuits imprimés ?
HARDWARE
-
Jet-Nemotron and PostNAS Optimize LLM Efficiency-Accuracy Trade-offs
By
–
Jet-Nemotron & PostNAS: Redefining the Efficiency–Accuracy Frontier in LLMs @NVIDIA
’s latest research introduces Jet-Nemotron, a family of hybrid-architecture LMs powered by Post Neural Architecture Search (PostNAS). Instead of costly pretraining from scratch, PostNAS starts -
Lightweight Travel Video Setup for Tech Content Creation
By
–
I typically record a video before I travel so I can release it while I'm away. For the news video, I have a simple travel setup. I do my best to round everything up but the quality is not quite as good normally. Dji Osmo Pocket, some DJi mics and a laptop are really all I need.
-
xAI to Open Massive Colossus 2 Data Center Next Month
By
–
Le monde n’est pas prêt pour ce qui arrive… xAI va ouvrir le mois prochain un immense nouveau centre de données : COLOSSUS 2
-
Nvidia Jetson AGX Thor: Blackwell GPU Chip for Real-Time Robotics AI
By
–
Nvidia a présenté le Jetson AGX Thor, une nouvelle puce pour la robotique conçue pour les tâches d’IA en temps réel.
— VISION IA (@vision_ia) 31 août 2025
Prix : 3 499 $, équipée d’un GPU Blackwell avec 128 Go de mémoire, offrant plus de 2 000 TFLOPs FP4 de puissance IA pour seulement 130 watts. pic.twitter.com/na7qbHDuaFNvidia a présenté le Jetson AGX Thor, une nouvelle puce pour la robotique conçue pour les tâches d’IA en temps réel. Prix : 3 499 $, équipée d’un GPU Blackwell avec 128 Go de mémoire, offrant plus de 2 000 TFLOPs FP4 de puissance IA pour seulement 130 watts.
-
Cerebras OSS 120B: 25 cents per million tokens, 3000 TPS
By
–
"But Cerebras hardware costs millions" My man, it's 25c per M tokens for OSS 120B. Let us worry about the hardware bill, you enjoy the 3,000 TPS tokens.
-
AIoT: Artificial Intelligence of Things and Machine Learning for IoT
By
–
The Artificial Intelligence of Things (AIoT): https://
techtarget.com/iotagenda/defi
nition/Artificial-Intelligence-of-Things-AIoT
… New #AIoT book v/ @PacktDataML → Hands-On #AI for #IoT — Expert #MachineLearning & #DeepLearning techniques for developing smarter IoT systems [2nd Ed.]: http://
amzn.to/3YAu6hS 𝓚𝓮𝔂 𝓕𝓮𝓪𝓽𝓾𝓻𝓮𝓼: -
Week of breakthroughs: Starship, CRISPR, DeepMind and optimism
By
–
Traylor Swelce.
Starship's perfect flight.
Chan and Doudna do CRISPR for kids.
Google DeepMind drops the neobanana.
Abundance Institute x Base Power Company. + Electric Slide
++ Cyan Banister on Dialectic What a week for the optimists. -
Boston Dynamics Spot and MATLAB Transform High Dynamic Range Inspections
By
–
🤖 Explore how @BostonDynamics' Spot robot and MATLAB-based AI algorithms are transforming high dynamic range inspections
— MathWorks (@MathWorks) 29 août 2025
Full video 🔗 https://t.co/zFP1w44in9#AI #Robotics #MATLAB pic.twitter.com/1clrJ3UebdExplore how @BostonDynamics
' Spot robot and MATLAB-based AI algorithms are transforming high dynamic range inspections Full video https://
spr.ly/6010ffKSY #AI #Robotics #MATLAB -
MoE Performance: Discrete GPUs vs Unified Memory Architecture
By
–
MoE is strictly worse for folks using discrete consumer GPUs like a 3090 AFAICT. It seems pretty good for unified memory stuff like Macs though.