65 results for “HP”
Hugging Face added GGUF support to Transformers, letting developers load llama.cpp checkpoints directly on Apple Silicon Macs. The integration reuses ggml Metal kernels and keeps the generation loop in PyTorch.
Tencent released WeKnora 0.8.0, an open-source knowledge platform that combines RAG, a ReAct agent, and Wiki Mode. The update adds a skill sandbox runtime, a tenant skill catalog, and cross-session memory.
Anthropic launched a redesigned Projects beta in Claude Code that replaces folder-style workspaces with a coordinator agent managing parallel cloud threads.
The era of vibe coding — where developers describe software in plain language and AI agents handle implementation — has moved from experiment. The consolidation around $20–25/month pricing tiers signals a maturing market where.
NVIDIA says TensorRT Edge-LLM ran Qwen3.6-27B through the MLPerf Edge Agentic workload in 24 minutes 36 seconds on one Jetson AGX Thor. The 6.4x gap over llama.cpp puts long-context agent inference on edge.
A US Special Operations analyst used an AI chatbot that hallucinated a report claiming a Chinese ship carried nuclear components. Airborne forces prepared an intercept before the error was caught.
Alibaba launched Qwen3.8-Omni-Flash on September 18 with a 1M-token omnimodal window for agentic audio-video work. The model is live on Qianwen and Alibaba Cloud Model Studio APIs.
NVIDIA says Groq 3 LPX's deterministic execution cuts voltage droop by over 60% on Vera Rubin, yielding a low-double-digit power reduction for the same inference workload.
Z.ai published a blog post describing custom inference infrastructure for its GLM models, framed as a step toward recursive self-improvement. No performance data or hardware details were disclosed.
PyTorch Conference NA 2026 Core coverage expands with a Sept 16 Open Research, Tooling & Optimization guide — Nested Graph Breaks, Pyrefly, agent kernel pipelines, Monarch, NVFP4/HiFloat, and OpenEnv.
Tencent's Hunyuan team released Hy-MT2 on 21 May 2026 — a new translation model family with instruction-following templates, three sizes up to a 30B-A3B MoE variant, and open weights plus IFMTBench. This is the next generation after the April Hy-MT1.5 phone quantizations.
NVIDIA took a minority stake in data-center site developer Cloverleaf Infrastructure to accelerate AI factory construction across the U.S. The partnership integrates NVIDIA's DSX platform early in site design to maximize compute output per megawatt.