AGI

Nokia launched its AI-RAN platform, an “AI-native” radio access network solution developed with NVIDIA. It promises significant spectral efficiency gains, potentially doubling operator capacity without new infrastructure. Offered as a software subscription, it signals a strategic shift from hardware to software development. While a competitor, Ericsson, already offers an AI-in-RAN solution, Nokia’s GPU-accelerated approach and partnership with NVIDIA aim for a leading market position.

Bluesight utilizes AWS cloud and AI services to build Prism, an AI layer unifying pharmacy and compliance data. The Prism Assistant for ControlCheck is now live in 20 health systems, streamlining controlled substance monitoring. A future 340B GPO compliance agent aims to automate complex drug purchase reviews. This initiative drastically reduces manual effort in hospital pharmacy operations, with significant improvements in report generation and analysis speed.

Cloudflare’s new web defaults, effective September 15, will automatically block AI agent and training crawlers in real-time on ad-supported pages. This move aims to differentiate bots from human users and establish clearer rules for the “agentic internet.” Publishers and AI developers must navigate these changes, with potential shifts towards negotiated access and new monetization models like “pay per use” as free web access for AI becomes less common.

Nvidia CEO Jensen Huang proposed an “AI token budget” as an engineer’s worth metric, alarming if annual AI token consumption is less than half their salary. This highlights a broader industry trend of shifting capital from human resources to AI token expenditure, with companies like Meta and Uber facing challenges with ROI on AI investments and high token costs. Experts suggest optimizing token usage through prompt caching, model selection, and batch processing, and reinvesting savings into augmenting, not replacing, human talent, especially entry-level positions, to ensure long-term AI success.

AWS GraphRAG revolutionizes drug R&D, cutting development cycles by 87% by unifying disparate databases into a knowledge graph. This solution leverages Amazon Neptune Analytics and Amazon Bedrock with NLP to transform isolated data into a searchable network. Queries in natural language yield accurate, verifiable answers mapped to literature and internal data, significantly accelerating research and knowledge retention.

NHS hospitals are set to adopt an AI-powered blood test to improve womb cancer diagnostics for women with heavy bleeding. Developed by PinPoint Data Science, the test analyzes blood markers to stratify risk into low, elevated, or high categories. This could spare up to one in five women from invasive procedures like transvaginal ultrasounds, streamlining care and freeing up GP capacity. The AI tool has shown high accuracy in trials and is part of a broader NHS push towards AI integration.

Insilico Medicine’s AI-identified drug candidate, rentosertib, targeting idiopathic pulmonary fibrosis (IPF), has advanced to Phase III trials. Rentosertib, which inhibits TNIK, demonstrated significant FVC improvement in a Phase IIa trial. The drug’s discovery and development pipeline, powered by Insilico’s Pharma.AI platform, involved AI-driven target prioritization and generative molecular engineering, achieving preclinical nomination in just 18 months. This progression validates the potential of AI in revolutionizing drug discovery by accelerating timelines and originating novel therapeutic opportunities.

Nvidia CEO Jensen Huang’s “token budget” metric highlights a corporate shift from human capital to AI token expenditure. While companies invest heavily in AI, initial results show many haven’t seen improved financial returns, with some even rehiring staff after AI-driven layoffs. This trend raises concerns about the true efficacy of AI-driven efficiency and its disproportionate impact on junior roles and lower-cost labor markets.

L’Oréal is revolutionizing beauty product development by integrating AI, quadrupling formulation speed. This predictive science allows for rapid testing of ingredient combinations, enabling repurposing of existing molecules for novel applications. Other consumer goods giants like Nestlé and Mondelez also leverage AI for recipe generation, ingredient efficacy, and supply chain resilience, significantly compressing development timelines while augmenting human expertise.

China is implementing new regulations for AI companions, targeting services designed for emotional interaction. These rules, effective July 15th, mandate anti-addiction systems, usage notifications, and real-time monitoring. Major platforms like Doubao and Qwen have proactively disabled companion features due to design challenges and potential non-compliance. The regulations aim to protect users, especially minors, from emotional manipulation and addiction, while also allowing Beijing to influence AI-generated content.