In the world of AI, Google's Big Sleep AI tool has successfully prevented the exploitation of a critical SQLite vulnerability, CVE-2025-6965, marking the first instance of AI proactively stopping cyber threats. Google DeepMind and Google Project Zero developed Big Sleep, which used threat intelligence to predict and thwart the attack. Meanwhile, Intel is shifting its focus from AI training, where Nvidia dominates, to edge AI, agentic AI, and foundry services, leading to job cuts in multiple states. Thinking Machines Lab, an AI startup led by Mira Murati, has secured $2 billion in funding from investors including Andreessen Horowitz, Nvidia, Accel, ServiceNow, Cisco, and AMD, and is valued at $12 billion despite launching recently. Oracle's co-founder, Larry Ellison, has become the world's second-richest person due to Oracle's AI investments and involvement in President Trump's Stargate initiative, which includes a $500 billion investment in AI infrastructure. Trade talks between the U.S. and South Korea could benefit from AI cooperation, addressing trade barriers and promoting innovation. Physical AI is also enhancing combat capabilities, with the market expected to reach $210 billion by 2034. China is leveraging undersea data centers to cool AI servers, aiming for low-carbon digital infrastructure, while Microsoft previously tested a similar project. Bob Ghafouri of Alvarez & Marsal has been recognized as a top AI consultant for helping companies implement AI solutions. However, a UN report indicates that only one South Korean researcher is among the top 100 AI researchers globally.
Key Takeaways
- Google's Big Sleep AI prevented the exploitation of SQLite vulnerability CVE-2025-6965, showcasing AI's potential in proactive cybersecurity.
- Intel is pivoting from AI training to edge AI, agentic AI, and foundry services due to Nvidia's dominance in the training market.
- Thinking Machines Lab, led by Mira Murati, raised $2 billion from investors like Andreessen Horowitz, Nvidia, and AMD to build safer AI systems.
- Larry Ellison's wealth increased by nearly $60 billion in 2025 due to Oracle's AI investments and participation in the $500 billion Stargate initiative.
- AI cooperation could strengthen U.S.-South Korea trade relations and address economic vulnerabilities.
- Physical AI is transforming warfare, with the market projected to reach $210 billion by 2034.
- China is utilizing undersea data centers to cool AI servers, aiming for low-carbon digital infrastructure.
- Bob Ghafouri of Alvarez & Marsal was named a top AI consultant for his work in implementing AI solutions for businesses.
- A UN report reveals that only one South Korean researcher is among the world's top 100 AI researchers.
- Google is improving its Timesketch platform with AI to speed up incident response.
Google AI Big Sleep stops SQLite vulnerability exploitation
Google's Big Sleep AI tool discovered and stopped the exploitation of a critical SQLite vulnerability called CVE-2025-6965. This marks the first time an AI agent has directly prevented active cyber threats. Big Sleep, developed by Google DeepMind and Google Project Zero, used threat intelligence to predict imminent attacks. This allowed Google to implement defensive measures and patch the vulnerability before it could be exploited. The AI's success shows the potential for AI to proactively defend against cyber threats.
Google AI finds critical security flaw in SQLite
Google's AI agent, Big Sleep, found a critical security flaw in SQLite, known as SQL CVE-2025-6965. This flaw was only known to hackers and could have been used to attack systems. Google used threat intelligence and Big Sleep to predict and stop the vulnerability from being exploited. Google DeepMind and Google Project Zero developed Big Sleep, which previously found another real-world security vulnerability last November. Google is also improving its Timesketch platform with AI to speed up incident response.
Google's Big Sleep AI prevents SQLite vulnerability exploitation
Google's AI, Big Sleep, stopped hackers from using a security flaw in SQLite before they could attack. The flaw, called CVE-2025-6965, could have let attackers inject harmful code. Big Sleep, created by Google DeepMind and Google Project Zero, found the flaw. Google believes this is the first time AI has directly stopped a real-world hacking attempt. Google is also working on ways to make AI agents secure, with human oversight and clear actions.
Intel shifts focus to edge and agentic AI
Intel CEO Lip-Bu Tan said it's too late for Intel to catch up in AI training. Intel will now focus on edge AI, agentic AI, and foundry services. Edge AI involves AI chips in devices like phones and cars. Agentic AI uses chips that make decisions without human help. Intel is cutting jobs in Oregon, Arizona, California, and Texas as part of this change. Nvidia's strong position in AI training led to Intel's decision.
Intel may leave AI training market
Intel is likely reducing its focus on the AI training market, where Nvidia has a strong lead. CEO Lip-Bu Tan said it might be too late for Intel to compete in AI training. Instead, Intel will focus on AI inference and agentic AI, which could be bigger opportunities. AI inference uses trained AI models, while agentic AI involves AI that acts independently. Intel may still develop rack-scale AI solutions, but its main focus is shifting.
Larry Ellison becomes world's second richest person
Larry Ellison, Oracle's co-founder, is now the world's second-richest person, surpassing Mark Zuckerberg. His wealth increased by nearly $60 billion in 2025, reaching $251 billion. This rise is due to Oracle's strong earnings and investments in AI. Oracle is involved in President Trump's Stargate initiative, investing $500 billion in AI infrastructure. Ellison also plans to focus his philanthropy on the Ellison Institute of Technology.
AI cooperation could save US-South Korea trade talks
Trade talks between the U.S. and South Korea face challenges due to tariffs and digital trade barriers. South Korea wants a "manufacturing renaissance partnership" with the U.S., focusing on AI and other technologies. However, disagreements over digital regulations and agricultural imports are obstacles. Cooperation in AI and economic security could strengthen the U.S.-South Korea alliance. This partnership could address shared economic vulnerabilities and promote innovation.
Physical AI enhances combat capabilities on battlefield
Physical AI, which uses sensors to control vehicles and robots, is changing warfare. AI systems can now identify targets, control drones, and make real-time decisions. These technologies give countries a strategic advantage and boost technological growth. Military technologies are also spreading to civilian industries, like logistics and healthcare. The Physical AI market is expected to grow significantly, reaching $210 billion by 2034.
Only one South Korean in top 100 AI researchers
A report from the United Nations shows that only one South Korean researcher is among the world's top 100 AI researchers. China leads the list with 57 AI experts, followed by the United States with 20.
Thinking Machines raises $2 billion for AI startup
Mira Murati's AI startup, Thinking Machines Lab, raised $2 billion in funding led by Andreessen Horowitz. The company is valued at $12 billion, even though it launched in February and has no products yet. Nvidia, Accel, ServiceNow, Cisco, and AMD also invested. Thinking Machines aims to build safer and more reliable AI systems. The company will soon share its first product, which will include open-source components.
Bob Ghafouri named top AI consultant and leader
Bob Ghafouri, a Managing Director at Alvarez & Marsal, has been named one of the top 25 AI consultants and leaders of 2025. He helps companies use AI to improve customer experience and operations. Ghafouri leads A&MPLIFY, which creates AI solutions using strategy, design, technology, and data. His work includes helping clients with vendor selection, model tuning, and risk management.
China uses undersea data centers for AI boom
China is using underwater data centers to cool the servers that power AI. These data centers use less energy because they are cooled by the ocean. One underwater data center is being built off the coast of Shanghai and will be powered by wind energy. China aims to lead in low-carbon digital infrastructure with this approach. Microsoft tested a similar project but has since stopped, while China is moving forward with large-scale deployments.
Sources
- Google's AI Tool Big Sleep Uncovered Critical SQLite 0-Day Vulnerability and Blocks Active Exploitation
- Google’s AI Agent Finds a Critical Security Flaw in SQLite
- Google AI "Big Sleep" Stops Exploitation of Critical SQLite Vulnerability Before Hackers Act
- Intel Officially Throws in Training Towel, Will Focus on Edge and Agentic AI
- Intel Might Be Quitting the AI Training Market for Good
- Larry Ellison surpasses Mark Zuckerberg to become world’s second-richest person after Oracle doubles down on AI investment
- How AI Cooperation Can Save the U.S.-ROK Trade Talks
- Artificial intelligence enters the battlefield – Physical AI is enhancing combat capa
- Only 1 S. Korean among world’s top 100 AI researchers
- Mira Murati's AI startup Thinking Machines raises $2 billion in a16z-led round
- Bob Ghafouri Named Among the Top 25 Artificial Intelligence Consultants and Leaders of 2025
- China Powers AI Boom with Undersea Data Centers