Nvidia launches OpenShell as Google simplifies search

Key Takeaways

• Argonne National Laboratory uses AI and supercomputers to design new carbon materials for medicine, energy, and national security.
• The THOR AI framework, developed by the University of New Mexico and Los Alamos National Laboratory, solves complex physics calculations in seconds.
• NVIDIA released OpenShell, an open-source runtime, to enhance the security of autonomous AI agents through isolated sandboxes and policy controls, a technology also leveraged in its partnership with Cisco for enterprise AI agent security.
• DIQSEO.com and CYBRSPC.AI are expanding AI-powered digital marketing services in Texas, focusing on customer automation and CRM modernization.
• Michigan lawmakers are considering regulations for major AI companies and the use of AI for workplace surveillance.
• Students at Brigham Young University are debating the ethical use of AI tools like ChatGPT, addressing concerns about copyright and misinformation.
• Mistral AI launched Mistral Small 4, a 119 billion parameter multimodal AI model with a 256k context window.
• Google quietly removed its "What People Suggest" AI feature, which provided crowdsourced health advice, as part of a search page simplification.
• Fortinet's FortiOS 8.0 introduces AI governance and visibility controls, including AI agents for network troubleshooting and managing generative AI usage.
• Agentic AI is expected to transform sales by automating tasks such as account research and CRM logging, impacting how sales professionals interact with systems like those from Salesforce.

AI helps scientists design new carbon materials

Scientists at Argonne National Laboratory are using supercomputers and artificial intelligence to understand how carbon transforms under extreme heat and pressure. This research, published in the journal Carbon, could lead to new materials for medicine, energy storage, and national security. By simulating atom-by-atom transformations, researchers can predict and control how carbon forms advanced structures. These new materials could be stronger than steel and lighter than plastic, with tailored properties for specific uses. The work showcases the power of AI and supercomputing in scientific discovery.

AI helps scientists design new carbon materials

Scientists at Argonne National Laboratory are using supercomputers and artificial intelligence to understand how carbon transforms under extreme heat and pressure. This research, published in the journal Carbon, could lead to new materials for medicine, energy storage, and national security. By simulating atom-by-atom transformations, researchers can predict and control how carbon forms advanced structures. These new materials could be stronger than steel and lighter than plastic, with tailored properties for specific uses. The work showcases the power of AI and supercomputing in scientific discovery.

THOR AI solves complex physics problem in seconds

Researchers at the University of New Mexico and Los Alamos National Laboratory have developed THOR, a new AI framework that solves difficult physics calculations in seconds. This system uses tensor network mathematics and machine learning to predict how atoms behave in materials, a task that previously took weeks on supercomputers. THOR can compute key properties of materials much faster while maintaining accuracy. This breakthrough is expected to speed up discoveries in materials science, physics, and chemistry.

NVIDIA and Cisco secure enterprise AI agents

NVIDIA and Cisco have partnered to enhance the security of enterprise AI agents. NVIDIA's OpenShell provides a secure runtime environment with isolated sandboxes and policy controls for these agents. Cisco AI Defense then monitors and verifies the agents' actions within these boundaries. This collaboration aims to make autonomous AI agents safer for critical business workflows. The system ensures that agents operate within defined limits, protecting against unmanaged risks and unauthorized access.

NVIDIA OpenShell enables safer autonomous AI agents

NVIDIA has released OpenShell, an open-source runtime designed to make autonomous AI agents, called claws, safer to use. OpenShell provides isolated sandboxes and fine-grained policy controls, allowing agents to run securely without code changes. This system addresses the security risks associated with self-evolving agents by enforcing boundaries and permissions. Developers can now benefit from the productivity of claws while maintaining enterprise privacy and security. The NVIDIA Agent Toolkit offers a full deployment stack for building and optimizing these agents.

AI marketing firm expands to Fredericksburg Texas

DIQSEO.com, a digital marketing company, is expanding its partnership with AI automation firm CYBRSPC.AI. They are opening new offices in Austin and Fredericksburg, Texas, to support regional industries. The companies will offer AI-powered growth, customer automation, and CRM modernization for businesses. Their new AI intelligence layer aims to unify business operations and automate tasks. They will also focus on SEO for the new AI search era and provide AI chatbots that remember conversations.

Michigan considers AI regulations

Michigan lawmakers are discussing new rules for artificial intelligence, focusing on how AI companies operate and how AI can be used to monitor employees. One proposal aims to set guardrails for major AI companies, defining critical risks and setting financial thresholds for compliance. Another measure seeks to regulate the use of AI for workplace surveillance, addressing concerns about employee privacy. While some support these regulations to ensure safety and public interest, others worry they could stifle innovation and create inconsistencies.

BYU students debate AI ethics and use

Students at Brigham Young University are debating the ethical use of artificial intelligence tools like ChatGPT. Some students see AI as an opportunity with risks that can be managed through personal responsibility. Others are concerned about copyright issues, environmental impact, and the potential for AI to produce incorrect information confidently. Professors note that learning to use AI is crucial for students' future, despite the challenges universities face in adapting curricula to rapid technological changes.

Mistral AI launches unified multimodal AI model

Mistral AI has released Mistral Small 4, a new 119 billion parameter Mixture-of-Experts model. This model combines instruction following, reasoning, and multimodal capabilities into a single system. It supports a 256k context window for tasks like document analysis and coding. Mistral Small 4 also features configurable reasoning effort, allowing developers to balance latency and depth of analysis. The company claims significant improvements in completion time and throughput compared to previous models.

Tech layoffs: AI is not the only reason

Many tech companies are blaming recent layoffs on efficiency gains from artificial intelligence. However, the reality is more complex, with factors like over-hiring during the pandemic and investor pressure also playing significant roles. While AI is disrupting some jobs, particularly in programming and customer service, the scale of displacement is often overstated. Layoffs may also be used to fund future AI investments rather than being a direct result of AI replacing workers. The future likely involves job transformation rather than elimination, with AI skills commanding higher wages.

Google removes AI health advice feature

Google has quietly removed its "What People Suggest" AI feature, which provided crowdsourced health advice from users worldwide. The company stated the removal was part of a broader simplification of its search page and not related to the feature's quality or safety. Initially launched to offer perspectives from people with similar medical experiences, the feature faced scrutiny amid growing concerns about AI-generated health tips. Google continues to provide health information from various sources, including forums.

Fortinet FortiOS 8.0 adds AI governance to networking

Fortinet has released FortiOS 8.0, updating its Security Fabric operating system with new AI governance and visibility controls. The update includes features for managing generative AI usage, inspecting encrypted traffic, and enhancing data loss prevention. It also introduces AI agents for troubleshooting and configuration within firewall and SD-WAN environments. This release aims to simplify security architectures as organizations adopt AI and cloud technologies, ensuring secure networking across hybrid and multi-cloud setups.

Agentic AI is the future of sales

Agentic AI is poised to revolutionize the sales industry by automating many human-led tasks. Instead of managing tasks, sales professionals will oversee AI agents that handle account research, lead classification, scheduling, and CRM logging. This shift will enable dynamic sales strategies, real-time messaging adjustments, and continuous optimization of pipeline management and forecasting. Mastering this technology will be crucial for sales success in the coming decade, as AI agents are expected to become more effective than humans in certain sales functions.

Sources

• Nanodiamonds and beyond: designing carbon materials with artificial intelligence at exascale
• Nanodiamonds and beyond: designing carbon materials with artificial intelligence at exascale
• THOR AI solves a 100-year-old physics problem in seconds
• Securing Enterprise Agents with NVIDIA OpenShell and Cisco AI Defense
• Run Autonomous, Self-Evolving Agents More Safely with NVIDIA OpenShell
• AI Automation and Website Digital Marketing Company Expand to Fredericksburg Texas
• Michigan lawmakers weigh new rules for artificial intelligence
• BYU students debate ethical use of artificial intelligence
• Mistral AI Releases Mistral Small 4: A 119B-Parameter MoE Model that Unifies Instruct, Reasoning, and Multimodal Workloads
• Tech companies are blaming massive layoffs on AI. What’s really going on?
• Google scraps AI search feature that crowdsourced amateur medical advice
• Fortinet’s FortiOS 8.0 Pushes Secure Networking Toward AI Governance
• Agentic AI Is the Future of Sales: Here’s How to Get It Right