AI News Archive - Browse Page 22 of 219
Browse AI news articles covering LLMs, tools, research, and industry trends
FreeToken Serves 753B Model on Single GPU, Fixing Existing Engine Flaws
A workstation with a single GPU just ran GLM-5.2, a 753-billion-parameter model, at usable speed.
AI Agents Drive 14x Surge in Token Usage on OpenRouter
February 6, 2026, might mark a small milestone in AI history, the day humans stopped being the main users of their own technology.
DRAM Shortages Lift Nvidia AI Server Prices by 15 Percent
Servers built around Nvidia's Vera Rubin and Grace Blackwell chips are getting more expensive, and by a lot.
Vercel's Free 'Is Agentic' Tool Audits Sites With 100+ Ora Checks
Vercel launched Is Agentic this week, a free tool that scores whether AI agents can actually find, load, and use a given website.
Chinese AI Labs Use Gray Market Claude Tokens to Train Models
Anthropic checks phone numbers. It checks foreign credit cards and billing addresses.
A.I. Agent Adoption Grows More Slowly Than AI Itself
Gartner expects more than 40% of today's agentic AI projects to be scrapped before 2028.
OpenAI urges stronger AI safety rules after recent incidents
OpenAI wants California lawmakers to go back and toughen up a law the company fought against just months ago.
AI Labs Lack Plans to Contain Rogue Models, Documents Show
Guidelight AI Standards graded five frontier AI labs on a question none of them like to answer directly: what happens the moment one of their models...
AI Agents Succeed or Fail Based on Their "Skills
Give an AI agent a "skill," a short cheat sheet of steps and warnings for a task, and it often does better.
Ignoring Human Beliefs Leads AI to Predict Wrong Actions, Study Finds
Sora can render a coffee cup vanishing into a cabinet with perfect physics. What it can't tell you is what the person who owns that cup will do next,...
RayNeo's AI Glasses Summarize Conversations With a Nod
RayNeo just put out a pair of glasses that has no camera and no speaker, which is an odd choice for a company selling AI eyewear in 2025.
RadixAttention Speeds First Tokens by Reusing Cached KV States
A Llama-3 8B model holding a 100,000-token conversation in memory needs close to 12.8 GiB just for its KV cache, before a single additional request...
Anthropic: Less Than 0.1% of Customer Chats Are Sexual Roleplay
Anthropic's usage policy is blunt on this point: Claude should not produce sexually explicit content, describe sex acts, or engage in erotic...
Holtercare-Bench: A Benchmark for Long-Term Dynamic ECG Analysis
A Holter monitor strapped to a patient's chest for 24 or 48 hours produces tens of thousands of heartbeats worth of data, far more than the...
Anthropic deploys Claude Mythos 5 for cybersecurity, with humans approving fixes
Anthropic is putting Claude Mythos 5, its most capable model to date, to work finding security holes rather than just answering prompts.
Deepseek Targets Visual Agents With New Experimental Flash Model
Deepseek pushed out V4-Flash-Vision-Exp this week, an experimental multimodal model that bolts image understanding onto its existing text-only...
The AI Agent Harness Is Security's 'Natural Control Point
Three separate reports this summer, from OpenAI, Anthropic, and the UK AI Security Institute, described frontier agents that wandered past their...
Nvidia's Linear Math Simplifies Costly AI Model Handoffs
Nvidia researchers have found a way to skip one of the most expensive steps in agentic AI: making a model re-read an entire conversation from scratch...
US Drafts Letter Forcing Partners to Choose Sides in AI Race
The State Department has drafted a letter telling partner governments they can't hedge on artificial intelligence anymore, according to Reuters,...
Meta AI Glasses Sales Surge Raises Covert Recording Concerns
Meta sold more than 2 million pairs of its AI-powered smart glasses last year, and that number keeps climbing.