Apple's 2029 AI server plan reportedly taps Nvidia NVLink for M8 Ultra
Two reports describe Apple building a custom M-series Ultra AI server for 2029, with one adding that Apple may use Nvidia's NVLink Fusion interconnect.
AI & ML
Frontier models, research results, agents, and developer tooling.
Harvey Uses GPT-6 Astra to Turn Legal Context into Drafts
OpenAI says legal AI Harvey now uses GPT-6 Astra to produce more structured, context-aware documents, freeing lawyers to focus on strategy.
NVIDIA AI Day Singapore to Spotlight Regional AI and HPC Advances
NVIDIA’s two-day Singapore event will pair hands-on training and expert sessions with partner showcases of AI progress across Southeast Asia.
Higgsfield Uses GPT-6 Astra to Ship Video Ad Tools in a Day
Higgsfield's one-day release of new video tools shows how GPT-6 Astra can compress product development time for small-business-focused platforms.
Border Surveillance Spending Fails to Prevent Deaths, Investigation Finds
A new investigation shows that billions spent on border technology have not prevented tragedies like the death of a man who crossed in New Mexico.
Why a Border Surveillance Camera Didn't Save a Dying Migrant
A woman's death in sight of a border camera reveals the limits of AI-powered surveillance in real-world rescue scenarios.
AI Can Empower Blue-Collar Workers, Not Just Replace Them
A new essay argues that the middle-class conversation about AI overlooks how hands-on workers can use the technology to improve their daily work.
Airbnb expands access to OpenAI's GPT-6 Astra for engineering work
OpenAI says Airbnb is broadening engineering teams' access to GPT-6 Astra and other frontier models to speed up debugging, design, and shipping.
New Study Finds No AI-Driven Hiring Drop for Recent College Grads
A CESifo working paper using Census data shows summer 2026 unemployment for recent graduates was normal, contradicting an earlier Stanford study.
Proaction uses Codex to build custom fleet demos, lifts sales 60%
A non-technical co-founder uses OpenAI's Codex to turn sales calls into personalized demos, cutting engineering hours and boosting deal conversion.
Pentagon seeks $30M for AI 'standoff' lie detector
A Defense Department budget request would fund algorithmic polygraph upgrades and remote physiological sensing, despite a history of unreliable deception detection.
AI Prior-Authorization Pilot for Medicare Draws Outcry and Legal Scrutiny
A Trump administration AI pilot program for Medicare prior authorization has produced widespread delays, denials, and provider complaints, while facing legal and procedural challenges.
Runway's WorldPrompt Steers Real-Time World Models with Persistent Context
A new Runway approach uses persistent context and timed actions to control a generative world model producing video and audio in real time.
Claude Opus 5.5, GPT-6 Astra, and Jev Lead AI News Roundup
This week's AI news features new frontier models from Anthropic, OpenAI, Google, and Xiaomi, plus a new class of cheap decision models like TypeSafe's Jev.
OpenAI Expands ChatGPT Ads Into Southeast Asia and Taiwan
The expansion gives eligible advertisers in the region access to a network that now spans more than 60 countries.
OpenAI Releases MentalHealthBench for Safer AI Therapy Talk
An expert-informed OpenAI benchmark aims to measure how well AI models handle realistic mental health conversations for both helpfulness and safety.
How invideo uses GPT-6 Astra to speed up color grading
A new OpenAI case study reports that invideo achieved a threefold improvement in color grading and produced 50 custom effects in a single day using GPT-6 Astra.
New Jersey fines AI data center $1.1M for unpermitted gas generators
The state's largest-ever data center penalty highlights the environmental cost of powering AI workloads with off-grid diesel generators.