Tag
Researchers unlocked the full performance of Nvidia's CMP 170HX GPU, a restricted mining variant, by cracking the debug AES key and exploiting a firmware vulnerability, enabling its use for AI inference and computing tasks.
Latham & Watkins, the second-largest US law firm, is building an in-house AI stack with Nvidia hardware and open-weight models to fine-tune on proprietary legal data, reducing reliance on AI labs like OpenAI and Anthropic.
NousResearch introduces a one-click local model setup tool for NVIDIA hardware, making it easy to deploy AI agents on Windows and Linux systems.
This article provides a practical guide on running 8x NVIDIA RTX PRO 6000 Blackwell GPUs for AI workloads, emphasizing high-concurrency inference, model fleets, and 70B fine-tuning, while comparing performance to more expensive data center solutions.
H3 Max is a post-trained version of MiniMax H3 optimized for maximum speed, ranking #1 in human preference evaluations for video quality, prompt understanding, and aesthetics while generating videos up to 35x faster than the official endpoint.
The article showcases the Qwen 3.8 27b AI model's ability to generate a sophisticated JavaScript and WebGL project for realistic real-time ocean rendering in the browser.
Antirez reports benchmarking DS4 inference on the DGX Spark (GB10), noting 12 tokens/sec generation speed and high prefill performance, with plans to merge the codebase once mature.