154 results for “Intel”
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.
Google DeepMind launched Gemini 3.8 Live and 3.8 Live Extended Thinking on September 15, 2026, adding real-time visual grounding and background tool execution to voice agents across the Gemini API and Workspace.
Apple Machine Learning Research published DACA-GRPO, a plug-in for GRPO-style trainers on diffusion language models. On LLaDA-8B it reports gains of up to 5.6pp on math, 7.4pp on code, and large jumps on structured constraint tasks.
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.
OpenAI CEO Sam Altman ruled out a 2026 IPO in a Fortune interview, citing safety work and the Hugging Face hack. The delay gives the company more runway to address alignment before facing public-market scrutiny.
IBM and Confluent brought Granite Time Series models to Early Access on Confluent Cloud, letting teams run forecasting and anomaly detection on streaming data via Flink SQL — without a separate ML stack.
A dark web service called Nexus is selling scans of 153 million driver's licenses, with timestamps linking the data to car rentals and dispensary ID checks. The FBI has opened an investigation into IDScan.net, the verification vendor suspected as the source.
Children's hospitals use NVIDIA's open source AI to model young hearts in seconds, including a transplant-matching model with 94.5% accuracy. The tools could cut donor-match delays.
Amodei’s pace-the-frontier call won lab backing — then Trump and Nvidia said no slowdown onstage. Monday’s tape split the story: cyber names jumped while chip infrastructure sold off.
OpenAI says Perplexity trusts GPT-6 Astra to write communications, edit software, and monitor production with less oversight. The endorsement suggests Astra's computer-use skills are crossing a trust threshold.
Alibaba launched the public beta of Wan3.0, a video model that generates 30-second clips from text, images, video, audio, PDFs, and web pages. The longer duration and document-to-video pipeline aim to cut stitching work for creators and enterprise teams.
SenseTime fully open-sourced SenseNova U1, a native unified multimodal model that fuses vision and language without a vision encoder or VAE. The 8B-parameter release matches commercial models on infographic generation and adds native 4K output in its U1.5 update.