AI News (06/25) : Google computer vision, OpenAI's custom chip, and China's 'cyber weapon' stir AI landscape.
The AI landscape is shifting rapidly, with today's news highlighting major strides in model capabilities, hardware innovation, and the increasingly complex regulatory and geopolitical environment. From Google's groundbreaking integration of computer control into Gemini to OpenAI's bold move into custom silicon, the industry is pushing boundaries while governments grapple with the technology's profound implications.
1. Google bakes computer control directly into Gemini 3.5 Flash, letting the model see and operate your screen
Google has integrated "Computer Use" directly into Gemini 3.5 Flash, giving the model the unprecedented ability to see, understand, and interact with computers, browsers, and mobile devices autonomously. This capability, previously confined to a separate Gemini 2.5 model, significantly enhances Gemini's utility for developers building agents that can automate tasks across various digital environments. The model scores 78.4 on the challenging OSWorld benchmark, putting it on par with top-tier competitors like Sonnet 4.6 and just shy of GPT-5.5 and Opus 4.8. Google is also implementing robust safeguards against prompt injection attacks, offering adversarial training and enterprise-level protections. Why it matters: This move dramatically expands Gemini's potential, making sophisticated automation, software testing, and office tasks more accessible, while also underscoring the critical need for advanced security measures.
2. China claims to have developed AI ‘cyber nuclear weapon’
China has reportedly claimed the development of an AI "cyber nuclear weapon." While details are scarce, the announcement itself signals a significant declaration in the ongoing global AI arms race. Why it matters: This potentially marks a new and concerning phase in national security and cyber warfare, indicating a strategic shift towards AI-powered offensive capabilities with global implications.
3. OpenAI will initially only release ChatGPT 5.6 to government-approved customers
OpenAI plans to stagger the release of its latest model, ChatGPT 5.6, initially making it available only to federal government-approved customers during a preview period. CEO Sam Altman stated this isn't their preferred long-term approach, indicating a response to increasing government scrutiny. This decision follows a recent presidential executive order calling for voluntary federal review of powerful AI models and a federal directive that led Anthropic to disable access to two of its models. Why it matters: This illustrates the tightening grip of government regulation on frontier AI models, potentially slowing broader public access and creating complex new hurdles for AI developers.
4. OpenAI just made its biggest move against Nvidia — and it could make ChatGPT cheaper to run
OpenAI has unveiled Jalapeño, its first custom-built AI processor, developed with Broadcom. Unlike general-purpose GPUs, Jalapeño is an inference chip optimized specifically for generating responses from trained LLMs, aiming to make ChatGPT faster, cheaper, and more reliable. Early internal tests suggest better performance per watt than current leading AI accelerators. Why it matters: This strategic move by OpenAI to design its own hardware challenges Nvidia's dominance, could significantly reduce operating costs for ChatGPT, and signals a broader trend of AI companies vertically integrating their stack.
5. Next evolution of attack drones like Shaheds will see AI choosing their own targets: Ukraine's drone guru
A Ukrainian "drone guru" has predicted that the next generation of attack drones, such as Shaheds, will be equipped with AI capable of independently selecting their own targets. Why it matters: This forecast highlights a concerning trajectory towards increasingly autonomous lethal weapons systems, raising profound ethical, legal, and military implications for future conflicts.
6. IBM unveils tech for chip smaller than 1 nanometer in AI computing push
IBM has announced a technological breakthrough: the development of a chip smaller than 1 nanometer. This advancement is aimed at accelerating the capabilities of AI computing. Why it matters: This represents a significant leap in semiconductor technology, promising exponential increases in processing power and efficiency that could unlock entirely new frontiers for AI development.
7. US assures India over AI ‘kill switch’ as ‘Pax Silica’ expands in bid to counter China
The US has assured India that access to AI technology, once provided, would not be cut off, addressing concerns sparked by Anthropic's recent restrictions. This discussion occurred during the "Pax Silica" summit, a US-led initiative to build AI supply chains free from Chinese influence. The US is developing a review mechanism for powerful AI models. Why it matters: This highlights geopolitical strategies aimed at securing AI supply chains and controlling technology access, demonstrating the increasing importance of AI in international relations and economic policy.
8. Databricks’ former AI chief thinks he can cut AI’s power bill by 1,000x
Naveen Rao, former head of AI at Databricks, has launched Unconventional AI, a company aiming to reduce AI inference power consumption by an astounding 1,000 times using a novel oscillator-based computer architecture. Their first model, Un-0, an image-generation system, already performs on par with state-of-the-art diffusion models running on conventional hardware, albeit currently via simulation. Why it matters: This ambitious project directly confronts the looming energy crisis in AI, proposing a revolutionary computing paradigm that could fundamentally alter the sustainability and scalability of future AI systems.
9. Gov. Gavin Newsom releases first-in-the-nation tool to track AI-related job losses
California Governor Gavin Newsom has launched a pioneering tracker to monitor AI-related job losses, a tool developed with the University of California’s California Policy Lab. Initial data shows no statewide surge in unemployment in at-risk fields, though upticks were observed in the Bay Area and tech sectors since 2022. This initiative aims to inform proactive policy responses and support for the workforce. Why it matters: This proactive governmental step provides crucial data to understand and mitigate the societal impact of AI on employment, setting a precedent for how other regions might address AI-driven labor market changes.
10. The New York Times Amends Lawsuit Against OpenAI and Microsoft
The New York Times has amended its ongoing lawsuit against OpenAI and Microsoft. The specific details of these amendments were not provided. Why it matters: This continued legal battle remains a pivotal case for establishing intellectual property rights and fair use principles in the generative AI era, with potential broad implications for content creators, publishers, and AI developers.
To stay up to date on everything going on in AI, check out the tracker at the500feed.com
Sources
- Google bakes computer control directly into Gemini 3.5 Flash, letting the model see and operate your screen
- China claims to have developed AI ‘cyber nuclear weapon’
- OpenAI will initially only release ChatGPT 5.6 to government-approved customers
- OpenAI just made its biggest move against Nvidia — and it could make ChatGPT cheaper to run
- Next evolution of attack drones like Shaheds will see AI choosing their own targets: Ukraine's drone guru
- IBM unveils tech for chip smaller than 1 nanometer in AI computing push
- US assures India over AI ‘kill switch’ as ‘Pax Silica’ expands in bid to counter China
- Databricks’ former AI chief thinks he can cut AI’s power bill by 1,000x
- Gov. Gavin Newsom releases first-in-the-nation tool to track AI-related job losses
- The New York Times Amends Lawsuit Against OpenAI and Microsoft