Large Language Models

Today's Stories

1. 1

   Lam Research breaks ground on AI chip lab in Oregon

   Lam Research has started construction on an R&D center in Tualatin, Oregon, as part of a $3 billion, five-year investment in its global lab network for advanced AI chips. The 120,000-square-foot facility will boost cleanroom space by over 50% and create more than 400 jobs, raising Lam's Oregon workforce to over 11,000 and its annual economic contribution to over $1.8 billion.

   The lab is expected to open in 2028 and will support process development for customers including Intel and Micron Technology, with a separate $1.7 billion investment commitment in Tualatin through 2041.

2. 2

   Visa expands AI cybersecurity tools to fix threats faster

   Visa announced enhancements to its cybersecurity portfolio, including the next evolution of its open-source Visa Vulnerability Agentic Harness (VVAH). The new release moves from discovery to remediation, with some resolutions shrinking from weeks to hours. Visa also expanded its Visa Consulting & Analytics Cybersecurity Advisory Practice with three new services. These help organizations assess risk, prioritize fixes, and strengthen resilience, addressing the challenge of deciding which vulnerabilities matter most before attackers exploit them.

   VVAH now supports Anthropic and OpenAI models alongside any other AI model through configuration. Since its open-source release in June 2026, it has been downloaded by tens of thousands of developers worldwide.

3. 3

   Visa Says Its AI ‘Harness’ Makes Anthropic Cheaper for Cyber Defense

   Visa says its AI 'harness' makes Anthropic cheaper to use for cyber defense. This could reduce the cost of using Anthropic's AI for security operations.

   How this affects Visa's cyber defense capabilities and broader AI adoption.

4. 4

   AMD ROCm 10 simplifies production AI for Instinct GPUs

   AMD introduced ROCm 10, the next version of its software stack for AMD Instinct GPUs, designed to streamline the path from development to production AI deployment. The update aims to reduce complexity and lower the barrier for enterprises running AI workloads on AMD hardware, potentially making AMD a more attractive option for businesses that currently rely on other platforms.

   AMD is positioning ROCm 10 as a simpler route to production, which could influence adoption among AI-focused companies evaluating hardware choices.

5. 5

   Home Depot Rolls Out AI Assistant to All U.S. Stores

   Home Depot is bringing its Magic Apron AI assistant to all U.S. stores, according to Yahoo Finance. This move could change how shoppers get help and how employees manage tasks, potentially setting a new standard for retail AI use.

   The impact on customer service and store operations as the AI tool becomes widely available nationwide.

6. 6

   Meta exec quits, warns AI will hit 1 in 5 corporate roles

   Clara Shih, former CEO of Salesforce AI and ex-Meta business AI leader, quit Meta this spring to start the New Work Foundation and consumer brand Dear CC. She says watching AI agents collapse a 10-step product development process into one or two people radicalized her. Shih predicts one in five corporate roles will be 'especially going to be challenged' by AI. She admits the optimistic story she told about AI freeing workers for higher-order tasks has 'primarily not been true,' citing customer support reductions from around 9,000 to 5,000 staff at Salesforce.

   The New Work Foundation offers free tools like Field Report, which shows AI exposure by major and occupation, and Game Plan, a mentoring app for rejected applicants. Shih's taxonomy, borrowed from MIT's David Autor, outlines three ways AI reshapes jobs, including case three where expert tasks get automated and job quality drops.