Nvidia outlines building blocks for agentic artificial intelligence
Nvidia details how to design, deploy, and scale agentic artificial intelligence using NeMo, NIM, blueprints, and GPU infrastructure. The company also spotlights an enterprise artificial intelligence factory design, demos, and sample use cases.
Artificial intelligence in banking: The rise of autonomous finance
Banks are rapidly adopting Artificial Intelligence to improve customer experiences and streamline operations, with board-level sponsorship now widespread. The sector is moving toward agentic, self-driving money while preparing for stricter oversight under the European Union Artificial Intelligence Act.
Amazon memos flag fundamental shift in how Artificial Intelligence startups spend on cloud
Internal Amazon documents say Artificial Intelligence startups are prioritizing models, inference, and developer tools over traditional AWS services, delaying cloud adoption. AWS disputes the characterization and says startups still choose its platform.
This week in European research, funding and Artificial Intelligence
Science|Business spotlights a packed news cycle from 14-16 October, including a proposed whistleblowing channel for misuse of Artificial Intelligence in science and a new defence innovation roadmap. Coverage also tracks the state of Artificial Intelligence in 2025, Horizon Europe debates and startup-friendly company rules.
Meta’s Yann LeCun reportedly clashed with the company over new publication rules
Meta’s top Artificial Intelligence researcher Yann LeCun has pushed back against stricter publication reviews at FAIR, with a report saying he even considered stepping down in September.
Anthropic Claude models on Vertex Artificial Intelligence
Vertex Artificial Intelligence provides fully managed access to Anthropic’s Claude models with streaming, logging, and flexible pricing. The catalog spans Sonnet, Opus, and Haiku tiers geared for agents, coding, research, and high-volume experiences.
Artificial intelligence guides personalized treatment for heart patients
An international team led by the University of Zurich used Artificial Intelligence to refine risk assessment in non-ST-elevation acute coronary syndrome, proposing a new GRACE 3.0 score that could better guide invasive treatment. The analysis spans data from more than 600,000 patients and suggests many should be reclassified.
Big tech backs BECCS carbon removal at mills and biomass plants
Microsoft and other large companies are paying pulp and paper mills and biomass projects to capture carbon dioxide and store it underground, spurring rapid growth of BECCS but also raising questions about carbon accounting and pollution.
Artificial intelligence is changing how clinicians quantify pain
Clinicians are testing artificial intelligence to turn pain into a measurable vital sign, from facial analysis apps in care homes to monitors in the operating room. Early deployments report fewer sedatives, calmer patients, and faster assessments, but questions about bias and context remain.