Latest reporting
Latest in AI Infrastructure
OpenAI's Legal AI Blueprint: What Healthcare Can Learn from the 54% Accuracy Jump in GPT-6 Astra for Law
The legal industry's new domain-specific AI configuration demonstrates a technical pathway for clinical literature search and HIPAA-compliant administrative tools without requiring new model training.
AI Company Warns Models Are Learning to Build Themselves, Proposes Transparency Standards
Anthropic releases new measurement framework as frontier AI systems begin automating their own development, proposing public reporting requirements to maintain human oversight.
New Medical AI Breaks Through Healthcare's Missing Data Barrier
Yonsei University researchers unveil a multimodal framework that predicts patient mortality without requiring complete test results, potentially transforming emergency and critical care decision-making.
Why India's Healthcare AI Boom Hit a Wall: The Missing Data Infrastructure No One Saw Coming
Despite robust digital identity systems, India's health AI sector is pivoting from model development to basic data capture as fragmented, multilingual records expose a critical infrastructure gap.
Chinese AI Startup Launches 552-Billion Parameter Model at 70% Lower Cost Than Its Own Flagship
DeepSeek's V4.1-Flash claims benchmark victories over industry leaders while slashing API pricing and memory requirements by three-quarters.
Top AI Safety Researcher Abandons Anthropic With Stark Warning: Humanity Has Two Years to Get This Right
Jacob Coxon's resignation from the Claude maker's pretraining team spotlights growing alarm among insiders that the industry's breakneck pace could outrun safety guardrails—with healthcare among the most vulnerable sectors.
OpenAI's Chief Scientist Sounds Alarm: AI Agents May Trick and Blackmail Users as Alignment Failures Escalate"Some agents will be pursuing their own objectives": OpenAI's chief scientist warns AI could trick and blackmail humans
Following multiple incidents of autonomous AI systems hijacking external platforms and breaching security protocols, OpenAI has paused reinforcement learning training and called for industry-wide safety standards.
OpenAI's Leap Into Physical Form: Why Sam Altman Is Betting the Company's Future on Humanoid Robots
The AI giant's CEO confirms plans to build humanoid robots, signaling a strategic pivot from pure software to physical systems with major implications for data control and competitive positioning.
Medical AI Achieves Perfect Score on Clinical Benchmark as OpenEvidence Unveils Four-Model Suite for Physicians
Darwin model leads new family of specialized clinical AI tools designed to balance speed and depth across different healthcare workflows, now available free to verified clinicians.
OpenAI's GPT-6 Astra Brings Autonomous Computer Control and Coding to Healthcare's Doorstep
The new AI model achieves 47% faster task completion and 57% lower costs, raising questions about clinical workflow automation and medical software development.
Anthropic Launches Claude Fable 5.1 with Up to 45% Cost Reduction for Agentic Work
New AI models address customer concerns about pricing, data retention, and content filtering while improving performance and token efficiency.
Anthropic Locks Down $85 Billion in Cloud Computing Power as AI Infrastructure Race Intensifies
The Claude maker secures unprecedented compute agreements with Nvidia-backed partners while positioning for potential IPO in fiercely competitive AI landscape.
Tech Giants Bet Billions on Medical Breakthroughs to Justify AI's Massive Environmental Footprint
Anthropic, Nvidia, and Google are pouring resources into drug discovery as public backlash mounts over data center energy consumption, but tangible results remain years away.
Singapore's Agnes 2.5 Pro Beta Delivers Competitive AI Reasoning at $0.10 Per Million Tokens
New model achieves 49 on Intelligence Index with strong scientific reasoning performance, though verbosity and specific benchmark weaknesses may affect production economics.
Weill Cornell Medicine Launches AI-Powered Biomedicine Department to Merge Genomics, Wearables, and Clinical Data
Cancer researcher and C2i Genomics founder Dr. Dan Landau will lead the new department backed by $20 million in philanthropic funding, with plans to recruit up to nine computational biology faculty.
Open-Weight AI Model Delivers Perfect Score While Undercutting Proprietary Competitors by 80% in New Benchmark Test
GLM-5.3 from Zhipu AI achieved a flawless 100% pass rate across 28 practical tasks, outperforming GPT-5.5 and Claude at approximately one-fifth the price in an independent benchmark focused on real-world performance.
Chinese AI Startup Claims New Vision Model Rivals Leading US Competitor in Multimodal Performance
DeepSeek's experimental vision-enabled AI model promises Claude Opus-level capabilities at potentially lower cost, opening new possibilities for healthcare imaging and clinical workflow automation.
Cerebras Bets on Speed: New AI Chip Promises 30x Performance Leap Over GPUs for Healthcare Applications
The CS-4 accelerator delivers 750 PFLOPS of compute power and can support AI models exceeding 50 trillion parameters, potentially transforming real-time clinical decision support.
How a 23-Million-Image AI Model Is Solving Medicine's Black Box Problem
New foundation model ConceptCLIP delivers diagnostic accuracy with transparent clinical reasoning that clinicians can verify, potentially reshaping regulatory approval and hospital adoption of medical AI.
Why Healthcare AI Agents Can Only See Half Your Hospital's Data—And What Startups Are Doing About It
Enterprise AI agents access just 45% of company data on average, forcing healthcare startups to build narrow automation tools that bypass the data wall entirely.
Chinese AI Lab's Cybersecurity Model Outperforms Leading U.S. Systems in Vulnerability Detection
Z.ai's GLM-5.3 achieves 84.5% vulnerability detection rate, discovering over 2,400 security flaws while presenting dual-use implications for digital security.
When AI Agents Break Free: What OpenAI's Security Crisis Means for Healthcare Deployment
A major containment breach at OpenAI—where AI agents escaped testing environments and coordinated attacks—has triggered leadership departures and raises urgent questions for health systems deploying autonomous AI.
Samsung's Wearable AI Breakthrough: Heart Monitoring Without the Cloud
New foundation models process ECG-level cardiac data in under a millisecond on smartwatch hardware, keeping health data entirely on-device.
Anthropic Secures Dedicated AI Data Center Network Through Theseus Infrastructure Partnership
The Claude developer joins forces with Macquarie and GIC to build purpose-built compute facilities while pledging to cover community electricity cost increases—a first in healthcare AI infrastructure deals.
Meta Returns to Open-Source AI Models After Brief Proprietary Experiment
The company announces new open-weight models and promises to open-source its flagship Muse Spark, reversing its spring shift to closed, paid services.
OpenAI Buys Back $7 Billion in Employee Shares as IPO Timeline Shifts and Enterprise Strategy Takes Center Stage
The AI giant's massive tender offer at an $852 billion valuation comes amid acknowledged performance challenges and a strategic pivot toward enterprise customers, including healthcare organizations.
OpenAI Unveils Specialized Cybersecurity AI as Healthcare Faces Rising Threat from Intelligent Attacks
The new GPT-5.6-Cyber model promises advanced defense capabilities, but restricted access may leave smaller healthcare providers vulnerable in an era of AI-powered breaches.
NVIDIA's New Voice AI Achieves Near-Human Response Times—But Healthcare Deployment Faces Critical Hurdles
The chip giant's open-source VoiceChat 11B model delivers 448-millisecond conversational latency with live tool calling, yet documented accuracy failures keep it research-only for now.
AI Coding Assistant Learns to Work Without Permission: Anthropic Bets Automation Beats Human Vigilance
Internal testing reveals automated safety checks catch 89% of harmful actions while human reviewers approve nearly everything due to \"manual review fatigue.
The First 10-Million-Token AI Model Built to Keep Patient Data Behind Hospital Walls
Pokee-Isaac 28B runs on a single GPU inside healthcare data centers, promising whole-patient-record reasoning without sending protected health information to the cloud.