Amazon and Uber reassess AI investments for efficiency

Cornell researchers have developed a device that stores data electrically but reads it through tiny mechanical motion, which could lead to more energy-efficient hardware for AI and scientific computing. This ferroelectric nanoelectromechanical system uses electrical pulses to program material and small read signals to reveal stored values, demonstrating 200 distinguishable electromechanical states.

Meanwhile, Claroty has launched an AI security agent called Claire, designed to help security teams identify exposures and prioritize remediation in critical systems, including connected operational technology and industrial control environments. Claire is trained on data from over 20,000 sites and 60 countries.

Nurses spend up to 41% of their time on electronic health records (EHRs), and AI-driven ambient documentation can help reduce this burden by capturing structured data from bedside conversations. Continuous AI accuracy tuning and AI-enabled change-management frameworks also support safe AI adoption in nursing.

Amazon and Uber are reassessing their AI investments as the trend shifts towards efficiency, following a period of excessive AI use. This comes as AI coding tools become popular and companies weigh the benefits and costs of AI investments.

A new protocol called 'Think First, Verify Always' (TFVA) aims to reduce AI manipulation risk by requiring employees to form their own judgment before consulting AI and cross-check critical outputs. The protocol has shown to improve decision quality by 7.87 percentage points.

KAIST has launched an AI college with an undergraduate program offering over 50 AI-specialized courses, aiming to connect problem definition to AI modeling using real industrial data. The university plans to recruit private-sector experts as adjunct faculty and build a 'full-stack' AI education and research system.

An AI system in Japan is being used to prevent suicides at stations and buildings, assessing the likelihood of individuals jumping and alerting security guards. The system has helped save at least 2 lives.

However, AI systems are not without their flaws. A nurse at Erlanger Baroness hospital stole fentanyl, and AI-based drug diversion software failed to detect it, highlighting concerns about AI transparency and oversight.

The AI industry is facing growing resistance to new data centers and regulations, with communities luring AI projects with tax breaks and lawmakers passing bills to regulate AI.

Key Takeaways

• Cornell researchers develop energy-efficient AI hardware using ferroelectric nanoelectromechanical systems.
• Claroty launches AI security agent Claire for critical systems, trained on 20,000+ sites and 60 countries.
• Nurses spend 41% of time on EHRs, and AI-driven ambient documentation can help reduce this burden.
• Amazon and Uber reassess AI investments for efficiency as AI coding tools become popular.
• 'Think First, Verify Always' protocol reduces AI manipulation risk by 7.87 percentage points.
• KAIST launches AI college with 50+ specialized courses and plans to build a 'full-stack' AI education system.
• AI system in Japan prevents suicides at stations and buildings, saving at least 2 lives.
• AI-based drug diversion software fails to detect fentanyl theft at Tennessee hospital, highlighting transparency concerns.
• AI industry faces growing resistance to new data centers and regulations, with communities and lawmakers pushing for change.

Cornell Engineers Use Tiny Beams to Boost AI Efficiency

Cornell researchers developed a device that stores data electrically but reads it through tiny mechanical motion. This approach could lead to more energy-efficient hardware for AI and scientific computing. The device uses a ferroelectric nanoelectromechanical system to store and read data, reducing electrical disturbance and idle power. It demonstrated 200 distinguishable electromechanical states, allowing for precise analog computing. The team aims to develop larger arrays of the device for complex matrix operations.

Tiny Vibrating Beams Rethink AI Hardware

Cornell researchers created a computing device that stores information electrically but reads it through tiny mechanical motion. This unusual approach could lead to more energy-efficient AI hardware. The device, a ferroelectric nanoelectromechanical system, uses electrical pulses to program material and small read signals to reveal stored values. The team demonstrated 200 distinguishable states, enabling precise analog computing. The approach could also help study emerging ferroelectric materials and develop adaptive microsystems.

Claroty Launches AI Security Agent for Critical Systems

Claroty introduced Claire, an AI security agent for cyber-physical systems, including connected operational technology and industrial control environments. Claire is trained on data from over 20,000 sites and 60 countries, and helps security teams identify exposures and prioritize remediation. The system aims to reduce manual work in risk prioritization and compliance preparation. It integrates with Claroty's existing platform, including asset inventory and threat detection.

AI Adoption in Nursing Gets a Boost

Nurses spend up to 41% of their time on electronic health records (EHRs), according to the U.S. Department of Health and Human Services. AI-driven ambient documentation can help reduce nursing burden by capturing structured data from bedside conversations. Continuous AI accuracy tuning and AI-enabled change-management frameworks also support safe AI adoption. These insights come from health system deployments and can help address skepticism around AI in nursing.

Amazon and Uber Reassess AI Investments

Tokenmaxxing, a trend of excessive AI use, sparks debate on productivity and cost efficiency. Amazon and Uber are recalibrating their AI usage as the trend shifts towards efficiency. This comes as AI coding tools become popular and companies weigh the benefits and costs of AI investments.

New Protocol Reduces AI Manipulation Risk

A new protocol called 'Think First, Verify Always' (TFVA) can reduce AI manipulation risk. The protocol requires employees to form their own judgment before consulting AI and cross-check critical outputs. A randomized controlled trial showed TFVA improved decision quality by 7.87 percentage points. The protocol aims to bolster independent judgment and verification, reducing AI risk.

The Sales Automation Mirage

The sales landscape is shifting with AI-generated content, making personalization less meaningful. Relevance and contextual timing are becoming crucial. The traditional sales development representative (SDR) role may become obsolete as AI handles tasks. The future of sales may involve a unified architectural sales role with AI as a cognitive amplifier.

KAIST Launches AI College with Specialized Curriculum

KAIST introduced an AI college with an undergraduate program offering over 50 AI-specialized courses. The program aims to connect problem definition to AI modeling using real industrial data. The university plans to recruit private-sector experts as adjunct faculty and build a 'full-stack' AI education and research system.

AI Missed Fentanyl Theft at Tennessee Hospital

A nurse at Erlanger Baroness hospital stole fentanyl, and AI-based drug diversion software failed to detect it. The incident highlights concerns about AI transparency and oversight. Experts stress the need for transparency and public disclosure of AI errors.

AI Reality Check Arrives

The AI industry faces growing resistance to new data centers and regulations. Communities are luring AI projects with tax breaks, but lawmakers are passing bills to regulate AI. The industry is bracing for a patchwork of state-level regulations.

AI System Prevents Suicides at Stations and Buildings

About 40 stations and commercial buildings in Japan use an AI system to prevent suicide. The system assesses the likelihood of individuals jumping and alerts security guards. It has helped save at least 2 lives and analyzes security camera footage to detect suicidal signs.