They never revealed the base model for Composer 1 as far as I can tell
RESEARCH
-
Best Practices For LLM Training
By
–
Best Practices For #LLM Training
by @Python_Dv #GenerativeAI #ArtificialIntelligence #MachineLearning #MI -
Revealing Rejected Paper Authors to Reviewers After Decisions
By
–
Here's a minimal tweak that comes at almost no cost: reveal the authors of rejected submissions to reviewers after decisions have been made. (Larger tweaks include single blind, the ICLR-style public reveal, etc.)
→ View original post on X — @thegautamkamath, 2026-03-20 20:30 UTC
-
MIT Cheetah: Revolutionary Quadruped Robot for High-Performance Locomotion
By
–
MIT Cheetah: The Agile Quadruped #Robot Redefining High-Performance Legged Locomotion
— Ronald van Loon (@Ronald_vanLoon) 20 mars 2026
by @IntEngineering
#Robotics #Engineering #ArtificialIntelligence #Innovation #Technology pic.twitter.com/5qPBmu1R3XMIT Cheetah: The Agile Quadruped #Robot Redefining High-Performance Legged Locomotion
by @IntEngineering #Robotics #Engineering #ArtificialIntelligence #Innovation #Technology -
NVIDIA Launches SOL-ExecBench, Revolutionary GPU Benchmark
By
–
We now have a benchmark that turns GPU performance into a "speed-of-light" score so you can see how much hardware headroom is left. Excited to see which tools top the leaderboard. — NVIDIA AI Developer (@NVIDIAAIDev) How close can you get to the speed of light? ⚡ Introducing SOL-ExecBench from NVIDIA — a benchmark for real-world GPU kernels that measures performance against hardware-grounded Speed-of-Light (SOL) bounds, not just software baselines. It includes 235 CUDA kernel optimization problems extracted from 124 production and emerging AI models, spanning forward and backward workloads across BF16, FP8, and NVFP4 on NVIDIA Blackwell GPUs. Dive in:
🏆 Leaderboard: research.nvidia.com/benchmarks/sol-execbench
🤗 Dataset: huggingface.co/datasets/nvidia/SOL-ExecBench
💻 Evaluator: github.com/nvidia/sol-execbench
📑 Paper: arxiv.org/abs/2603.19173 [Translated from EN to English] -
US Military Strategy: AI and Tech Race Against China
By
–
The Under Secretary of War, @USWREMichael, is literally telling us that this War is about the power competion between China and the U.S., And that AI, the technology race, and ability to build the Industrial base, will determine the outcome. https://t.co/HHgoDrCBsQ
— Nina Schick (@NinaDSchick) 20 mars 2026The Under Secretary of War, @USWREMichael
, is literally telling us that this War is about the power competion between China and the U.S., And that AI, the technology race, and ability to build the Industrial base, will determine the outcome. -
Jet lag decision: Karpathy or Terence Tao talks
By
–
Back in germany, jet lag hitting hard, havent slept for 36 hours. Cant decide what I want to watch to chill down, so many great videos up. New talk w/ Karpathy on No Prior or Terence Tao on Dwarkesh.
-
Hands-On Projects Essential for AI Learning Success
By
–
This is great for students. Building and breaking things is how you actually learn. We see the same with our courses at Towards AI, the hands-on projects are where everything clicks.
-
Opus vs Sonnet: Long-Horizon Planning and Context Window Capabilities
By
–
10x per generation on a task that requires actual long-horizon planning is something. Cowork personally surprises me a lot on what it can do in both good and bad ways. And TBH, sonnet cannot understand and follow large skill files nearly as good as opus. And Opus 1m context is
-
Misleading Data Visualization: Truncated Y-Axis Distortion
By
–
“Dramatically” Yeah, if you truncate the Y axis. Another dataviz crime.