How a small research team turns the theoretical potential of GPUs into real-world performance for AI systems – the story of the Together AI team.
Salesforce AI Research explains how it is restructuring language model training for the agentic era – and why old approaches no longer work.
Researchers have developed a specialized test for AI systems used in architecture, engineering, and construction. The results were quite sobering.
AI: Events
Hands That Think for Themselves: Sanctuary AI Robot Masters Object Manipulation Without Prior Training
Products
Sanctuary AI demonstrated how its hydraulic hand autonomously reorients objects in its grasp, without pre-programmed movements or data labeling.
AI: Events
One GPU Failure Shouldn't Bring Down the Entire System
Technical context • Infrastructure
The Mooncake and Volcano Engine teams have integrated an elastic expert parallelism mechanism into the SGLang framework, allowing it to withstand partial failures without requiring a restart.
Arcee AI has released Trinity-Large-Thinking, an open model with a 'thinking' function for complex agentic tasks, available under the Apache 2.0 license.
AI: Events
AMD at MLPerf Inference 6.0: A Million Tokens Per Second and a Debut in Video Generation
Technical context • Infrastructure
AMD has presented its MLPerf Inference 6.0 results, showcasing new performance records, the first video generation tests, and scaling up to the cluster level on the Instinct MI355X GPU.
AI: Events
Red Hat AI Achieves Top Results in MLPerf Inference v6.0 – Here's What's Behind It
Infrastructure
Red Hat AI has secured top spots in the latest round of the MLPerf Inference v6.0 benchmark, testing three models on both NVIDIA and AMD GPUs.
Lab
Cartilage as a Living Battery: What Electrical Signals Reveal About Joint Health
Biology & Neuroscience
Scientists have learned to “listen” to the electrical signals inside articular cartilage – and this opens the door to early arthritis diagnosis long before pain appears.