NVIDIA has introduced PAIR, a tool designed to help users connect old hardware into a personal AI cluster. The initiative is aimed at leveraging existing devices to run AI workloads locally.
The project allows individuals to utilize their accumulated hardware resources for artificial intelligence tasks.
FlashAttention-4 addresses the performance limitations of Blackwell's 4-bit floating-point (FP4) tensor cores by introducing new paths for noncausal inference and causal training. The method, called Direct-P, maps scores directly to FP4 probabilities to bypass softmax conversion overheads that typically dominate once matrix products shrink.
NVIDIA has scheduled the release of its next-generation Vera Rubin GPUs for mid-2027. The company expects to sell $20 billion worth of Vera Rubin hardware in the current quarter, which would account for 20 percent of its data center revenue.
Perplexity has released Portable Computer, a local-first build of its agentic platform that runs the agent harness, orchestrator, planner, tool router, and post-trained models directly on NVIDIA DGX Spark hardware. This packaged system allows every task to begin on-device, carrying no per-token charge for work handled by local models, while escalating only specific steps to 15+ cloud models upon user approval.
NVIDIA has released CUDA Python 1.0, providing a unified foundation for GPU computing in Python. This release introduces stable APIs to the platform.
OpenAI has released initial performance results for Jalapeño, its first custom inference chip, demonstrating significant gains in speed and power efficiency compared to existing commercial systems. The company states that the new architecture allows for higher throughput and lower latency simultaneously, addressing a common trade-off in current hardware.
We use cookies to measure traffic and improve the site. You can accept or decline analytics cookies. Privacy policy