Generating high-quality code is the basis for code assistants but also for almost all Agentic-AI approaches That's why I'm very excited to see 2025 starting to be the year of high-performance code generation in *open-source* LLMs After our latest release 'OlympicCoder' beat
RESEARCH
-
Code Fundamentals: Why AI Development Relies on Programming
By
–
Agreed. And even if it mainly applies to coding, that's pretty important. AI is made of code.
-

COBOL Unlearning and AI Model Training Implications
By
–
“How could unlearning COBOL lead to this?” I’m crying
-

DGX Station: High-Performance AI Computer for Researchers
By
–
7. DGX Station: A high-performance AI computer for data scientists and researchers
— Rowan Cheung (@rowancheung) 20 mars 2025
20 petaflops of performance with 72 CPU cores, a chip-to-chip interface, high-bandwidth memory, and PCIe slots for GeForcepic.twitter.com/GbrREYVxGP7. DGX Station: A high-performance AI computer for data scientists and researchers 20 petaflops of performance with 72 CPU cores, a chip-to-chip interface, high-bandwidth memory, and PCIe slots for GeForce
-

Newton: Open-Source Physics Engine for Robotic Simulation
By
–
2. Newton: An open-source physics engine to simulate robotic movements in the real world — developed jointly by Nvidia, DeepMind, and Disney Research.
— Rowan Cheung (@rowancheung) 20 mars 2025
(This is the physics engine that powers the Blue robot!)pic.twitter.com/YuCEpOkvKN2. Newton: An open-source physics engine to simulate robotic movements in the real world — developed jointly by Nvidia, DeepMind, and Disney Research. (This is the physics engine that powers the Blue robot!)
-
ArXiv Paper Curation Platform Powered by Community Engagement
By
–
We’re curating the most viewed and liked papers on arXiv every day 👀
— alphaXiv (@askalphaxiv) 20 mars 2025
You can now browse through AI papers based on what our community of hundreds of thousands of researchers enjoy most
Crowdsourced from alphaXiv and our Chrome extension
Goodreads for arXiv🚀 pic.twitter.com/yEqKLZG6CiWe’re curating the most viewed and liked papers on arXiv every day You can now browse through AI papers based on what our community of hundreds of thousands of researchers enjoy most Crowdsourced from alphaXiv and our Chrome extension Goodreads for arXiv
-

GridMix: Novel Machine Learning Approach with Open Source Code
By
–
Paper: https://
openreview.net/forum?id=Fur0D
tynPX
…
Code: https://
github.com/LeapLabTHU/Gri
dMix
… -

CORAL Approach: Multi-Resolution Feature Grids for Spatial Modulation
By
–
It leverages the CORAL approach, encoding each INR with a multi-resolution feature grid for spatial modulation and integrating a novel spatial domain augmentation strategy to improve generalization across spatial domains.
-

GridMix Enhances Implicit Neural Representations for PDE Solving
By
–
This ICLR 2025 Oral paper from Tsinghua University presents GridMix, which enhances implicit neural representations (INRs) for solving PDEs.
-

AI21 Labs Discusses Reasoning Models and AI Planning Systems at GTC
By
–
Today at @nvidia GTC! Join AI21 Labs’ co-founder and co-CEO @origoshen as he takes the stage to discuss the transition from reasoning models to AI planning systems. If you’re attending, let us know in the comments!
