rtx

NVIDIA optimizes Google DeepMind's DiffusionGemma, an open model that generates text in parallel 256-token blocks, achieving up to 4x faster performance on local RTX GPUs, DGX Spark, and DGX Station systems.

NVIDIA announced RTX Spark PCs and a wave of updates to enable local AI agents across RTX and DGX ecosystems, including the OpenShell runtime coming to Windows, NemoClaw expansion, performance improvements, and integrations with Adobe and H Company.