The evening briefing.
Today across AI and tech: Government restricts access to OpenAI's new GPT-5.6 Sol, Oracle's stock plunges amid AI financing concerns, and Apple faces supply chain leaks and price hikes.
Frontier AI Governance. The US government is reportedly vetting access to OpenAI's new GPT-5.6 Sol model, a restriction OpenAI calls unsustainable. This marks a significant shift in how advanced AI models are deployed and controlled, raising questions about who gets to use cutting-edge technology. The model reportedly outperforms Anthropic's Claude Mythos 5, but is being rolled out under government-controlled access rules, following a similar directive given to Anthropic.
AI Market Dynamics. The financial strain of AI development is becoming evident, with Oracle stock experiencing its worst week since 2001 due to surging spending and debt related to AI financing. Chip stocks, including Micron and Nvidia, faced a broad sell-off, reflecting mounting concerns over the rising cost of AI infrastructure. China's Zhipu is reportedly closing the gap on top U.S. AI models, shifting the competitive landscape. Nvidia CEO Jensen Huang dismissed fears of AI reducing jobs as "complete nonsense."
Policy and Ethics in Tech. Regulatory and ethical debates continue to shape the tech landscape. The New York Times has intensified its copyright infringement claims against Microsoft, alleging the company built a supercomputer to aid OpenAI's infringing activities. Simultaneously, Google expressed a desire for AI regulation, but on terms that would allow its current operations to continue unimpeded. In California, a proposed 3D printer surveillance scheme faces opposition, highlighting privacy concerns.
Tech Industry Shifts. The tech industry is also seeing significant shifts in talent and supply chains. OpenAI has notably poached Apple's Vision Pro and smart glasses chief, indicating a strategic move into hardware-adjacent AI applications. Concurrently, Apple is addressing a sensitive files leak from its supplier Tata Electronics, underscoring vulnerabilities in global supply chains. Furthermore, Apple has announced price hikes on Macs and iPads due to rising memory chip costs.
US government will decide who gets to use GPT-5.6
The US government is reportedly vetting access to OpenAI's new GPT-5.6 Sol model, a restriction OpenAI calls unsustainable. This intervention follows a similar directive given to Anthropic regarding its Fable 5 model.
China's Zhipu is closing in on top U.S. AI models
Zhipu's GLM 5.2 demonstrates the AI competition is shifting towards delivering the most intelligence per dollar. This development makes open source models a real contender against established players.
OpenAI poaches Apple Vision Pro and smart glasses chief
OpenAI has hired Paul Meade, who was previously in charge of Apple Vision Pro and Apple's smart glasses initiative. This move signals OpenAI's continued expansion into hardware-adjacent AI applications.
NYT slams Microsoft for building copyright-infringing supercomputer for OpenAI
The New York Times has intensified its copyright claims against Microsoft, alleging the company built a supercomputer to help OpenAI infringe copyrights. This legal action follows a recent SCOTUS ruling against Sony.
Google wants AI regulation, but on its own terms
Google has expressed its desire for AI regulation, but with the caveat that rules should allow the company to continue its current operations. This stance suggests a preference for self-serving regulatory frameworks.
Nvidia CEO Jensen Huang: 'People talk about AI reducing jobs — complete nonsense'
Nvidia CEO Jensen Huang dismissed fears that AI will lead to job reductions, calling such concerns "complete nonsense." He pushed back against long-held automation fears.
Anthropic and 19 organizations launch open source security body
Anthropic and 19 other organizations have launched an open source security body following the Fable 5 ban. This initiative aims to coordinate vulnerability disclosures in open-source projects.
Accelerating Gemini Nano models on Pixel with frozen Multi-Token Prediction
Google Research is accelerating Gemini Nano models on Pixel devices using a new technique called frozen Multi-Token Prediction. This advancement aims to improve on-device AI performance.
Oracle stock has worst week since 2001 as AI financing concerns escalate
Oracle's stock experienced its worst week since the 2001 dot-com bust, driven by investor concerns over surging spending, negative free cash flow, and a $130 billion debt pile related to AI financing.
Micron sinks 6%, wrapping a wild week of trading that saw big swings
Micron Technology's shares tumbled 6% on Friday amidst a global sell-off in chip stocks. This decline reflects broader market concerns over AI infrastructure costs.
Apple working with supplier Tata after sensitive files leak online
Apple is reportedly working with its Indian supplier, Tata Electronics, to address a significant data breach that exposed sensitive files earlier this month. The leak highlights supply chain security challenges.
Apple raises prices on Macs and iPads before price hike kicks in
Apple has raised the prices on Macs and iPads in response to the rising costs of memory chips. This makes existing discounts on Apple laptops significantly better deals.
We Can Still Stop California's 3D Printer Surveillance Scheme
The Electronic Frontier Foundation is campaigning to stop California's proposed 3D printer surveillance scheme. The legislation raises significant concerns about privacy and government oversight.
Corgi, Y Combinator-backed insurance tech startup, denies stealing open source product
Corgi, a Y Combinator-backed insurance tech startup, has denied accusations of stealing an open source product from Papermark. The controversy raises questions about "vibe coding" and intellectual property.