The afternoon briefing.
OpenAI reportedly considers legal action against Apple, Cerebras makes a huge Nasdaq debut, and Americans express strong opposition to AI data centers.
AI Partnership Tensions. OpenAI is reportedly preparing legal action against Apple, signaling a fraying relationship over their Siri integration. This comes after Apple's deal to use ChatGPT in iOS 18, with OpenAI allegedly dissatisfied with the partnership's terms. Meanwhile, Microsoft has begun canceling Claude Code licenses for its internal developers, suggesting a shift in its AI coding tool strategy. These developments highlight the complex and often contentious nature of strategic alliances in the rapidly evolving AI landscape.
Hardware & Investment Boom. The AI hardware sector continues to see significant investment and innovation, with Cerebras almost doubling its value in a Nasdaq debut to a $100 billion market cap. Nvidia is advancing agentic AI with its Vera Rubin platform, designed to handle the complex, non-deterministic workloads of AI agents. Cisco's CEO also reported a "networking supercycle" driven by strong AI demand, causing its stock to pop 13%. These events underscore robust growth and strategic hardware development across the AI industry.
Public Perception & Policy. A new Gallup poll reveals that 71% of Americans oppose building AI data centers near their homes, preferring nuclear power plants due to concerns over water, energy use, and pollution. This public sentiment is driving policy discussions around AI infrastructure and its environmental impact. Separately, Germany's Sovereign Tech Fund has backed KDE with €1.3M, indicating European efforts to foster open-source alternatives and reduce reliance on proprietary operating systems. These trends reflect growing societal scrutiny and governmental interest in shaping the future of AI and tech development.
AI's Practical Challenges. Real-world applications of AI are facing scrutiny as an Ontario audit found that doctors' AI notetakers are making up information, including therapy referrals and prescriptions. This highlights critical reliability issues in sensitive domains. Concerns also extend to privacy, with a robot dog reportedly sending data to China, raising questions about data security in connected devices. These incidents underscore the ongoing need for robust validation, ethical guidelines, and transparency in AI deployment.
OpenAI reportedly preparing legal action against Apple over Siri partnership
OpenAI has reportedly enlisted an outside law firm to explore legal options against Apple, following dissatisfaction with their Siri integration partnership. This potential legal action highlights tensions between the two tech giants despite their recent deal to integrate ChatGPT with iOS 18.
Cerebras almost doubles in Nasdaq debut, topping $100 billion market cap after blockbuster IPO
Cerebras, a chipmaker specializing in AI, saw its stock nearly double on its Nasdaq debut, achieving a market capitalization over $100 billion. This marks one of the most significant pure-play AI IPOs to date, capitalizing on the current bull market for silicon.
Microsoft starts canceling Claude Code licenses for internal developers
Microsoft is reportedly scaling back its internal use of Anthropic's Claude Code, canceling licenses for its developers. This move comes after the AI coding tool proved very popular internally, suggesting a re-evaluation of its AI strategy.
Americans would rather live next to a nuclear plant than an AI data center, Gallup poll finds
A new Gallup poll indicates that 71% of Americans oppose the construction of AI data centers near their homes, a higher percentage than those opposing nuclear power plants. Concerns cited include high water and energy consumption, pollution, and rising utility costs.
How the NVIDIA Vera Rubin Platform is Solving Agentic AI’s Scale-Up Problem
NVIDIA's Vera Rubin NVL72 platform is designed to address the scale-up challenges of agentic AI inference workloads. The platform handles non-deterministic trajectories and compounds end-to-end latency across hundreds of inference requests per session.
Intel is now making iPhone chips for Apple, per report
According to a new report from Ming-Chi Kuo, Intel has begun initial production on iPhone processors and other chips for Apple. This follows earlier reports of a deal between Apple and Intel for chip manufacturing.
AMD promises to bring improved, hardware-backed FSR 4 upscaling to older Radeon GPUs
AMD has committed to bringing its improved, hardware-backed FSR 4 upscaling technology to older Radeon GPUs. While FSR 4.1 on RDNA3 or RDNA2 GPUs may experience a performance hit, the update aims to enhance gaming visuals across a wider range of hardware.
Zero-day exploit completely defeats default Windows 11 BitLocker protections
A new zero-day exploit has been discovered that can completely defeat the default BitLocker protections in Windows 11. Microsoft is currently investigating the vulnerability, the exact mechanism of which remains unclear.
Another major Linux security issue uncovered - new Fragnesia flaw allows attackers to run malicious code as root
A new Linux security vulnerability, dubbed Fragnesia, has been uncovered, allowing attackers to run malicious code with root privileges. This flaw is related to the Dirty Frag family of exploits and enables privilege escalation at the kernel level.
Germany's Sovereign Tech Fund Backs KDE with €1.3M
Germany's Sovereign Tech Fund has provided €1.3 million in backing to KDE, an open-source desktop environment. This investment reflects Europe's growing recognition of the need for independent operating system alternatives.
AI is making me dumb
A personal reflection piece discusses the potential for AI tools to diminish human cognitive abilities and critical thinking. The author explores how over-reliance on AI might lead to a decline in individual intellectual capacity.
Man Finds Robot Dog Is Bad at Protecting His Chickens, But Might Be Good at Sending Data to China
A man discovered his robot dog was ineffective at protecting chickens and raised concerns about its potential to send data to China. The incident highlights ongoing debates about data privacy and the security of connected devices.
Your doctor’s AI notetaker may be making things up, Ontario audit finds
An audit in Ontario has revealed that AI notetakers used by doctors are prone to making up information, including incorrect therapy referrals and prescriptions. These findings raise serious concerns about the reliability and safety of AI in healthcare applications.
Granite Embedding Multilingual R2: Open Apache 2.0 Multilingual Embeddings with 32K Context
IBM has released Granite Embedding Multilingual R2, an open Apache 2.0 licensed model offering multilingual embeddings with a 32K context window. This model achieves the best sub-100M retrieval quality, enhancing capabilities for multilingual AI applications.