AI News Archive - Browse Page 19 of 160
Browse AI news articles covering LLMs, tools, research, and industry trends
QuickReduce FP4 delivers ~4.1× speedup over RCCL at TP=4 for large messages
Large language models run on billions of parameters, and inference almost always leans on tensor parallelism to split the workload across several...
CODEX Agent Adds AI‑Q Deep Research Skill from GitHub Repository
Agent harnesses such as Claude Code, Codex and LangChain Deep Agents excel at orchestrating sessions, chaining tools and running code in response to...
AI models learn chemistry; talent and collaborations offset location concerns
The space of possible small‑molecule drugs is staggering—estimates put it between 10²⁰ and 10⁶⁰ individual compounds.
Study Uses SHARP and New Error Framework to Assess PHRs in Health AI
Personal health records promise patients a clearer view of their own data, but the sheer volume and medical jargon can make the files hard to...
Google's Gemini 3.5 Flash, pricier, adds 11 Omniscience points hallucinations 61%
Google DeepMind rolled out Gemini 3.5 Flash this week, positioning it as the fastest model in its class—more than 280 output tokens per second,...
Google to roll out AI agents this week, adopting OpenClaw features
Google is finally putting AI assistants into everyday use. At I/O 2026 the company unveiled a suite of agents that can pull data, plan events, and...
NVIDIA introduces verified skill cards to govern AI agent capabilities
Why does this matter? Autonomous AI agents are getting smarter, but their growing capabilities come with a transparency problem.
Alibaba launches Qwen3.5‑LiveTranslate‑Flash: 60‑language translation in 2.8 s
Simultaneous interpretation has always been a tough AI problem. You’re asking a model to translate speech before the speaker even finishes a...
AI Agent in EOS Automates Validation, Error Correction for Full Lab Cycle
Why does this matter? Modern labs promise faster, safer, more reproducible experiments, yet getting them to run autonomously still feels like...
Real-Time Diffusion on Apple M3 Ultra: CoreML, Quantization, Neural Engine
Why does Apple’s newest silicon matter for generative art? While NVIDIA GPUs have become the default playground for diffusion‑based image synthesis,...
NVIDIA and Google Cloud let developers scale AI from prototype to production
Google I/O this year is turning its attention to “agentic” experiences, and the spotlight lands squarely on the partnership between NVIDIA and Google...
Google plans USD 180‑190 B capex—six‑fold rise—to power AI agents like Spark
Why does this matter now? Google is gearing up to spend roughly $180 billion to $190 billion on capital‑expenditure this year—about six times what it...
Roundtables: Inside the Musk v. Altman Trial – Audio Session Available
Elon Musk's lawsuit against OpenAI ended in defeat last week, with a jury finding no merit in his claim that CEO Sam Altman and President Greg...
Google AI agents expand on Alerts, moving beyond simple notifications
At Google I/O 2026 the company unveiled a new layer of “agentic” functionality built into Search.
Google launches three AI subscription tiers at USD 10, adds Gemini Omni and 3.5 Flash
Google reshaped its AI subscription lineup at I/O 2026, swapping daily prompt caps for a consumption‑based billing approach.
Google’s Gemini Spark expands AI access to user data, raising trust concerns
Google rolled out a suite of AI tools at I O 2026 that promise to streamline everyday tasks.
Evaluating AI Agents: Does the Engine Grasp Instructions and Reason Facts?
Evaluating an AI system isn’t a one‑size‑fits‑all task. When you run a model benchmark, you’re looking at a foundation model in isolation—testing how...
Google Search agents can act without user input, says VP Robby Stein
AI agents have moved from buzzword to product demo in every software briefing over the past year.
I/O 2026 unveils Gemini Omni for universal creation, Gemini 3.5 Flash debut
Google I/O 2026 put a new spin on the company’s AI roadmap. The headline? Gemini Omni, a model that “can create anything from any input,” starting...
EKS Hosts Multistage Multimodal Recommender; DLRM Personalizes Rankings
Why does this matter? Because building a recommender that can juggle images, tabular data, and real‑time user signals is no longer a research...