Anthropic accuses Alibaba of 'industrial scale' AI theft; IBM breaks sub-1nm barrier
The tech and geopolitical news cycle is accelerating today, with three major stories reshaping the AI race, semiconductor roadmap, and humanitarian response landscape.
The tech and geopolitical news cycle is accelerating today, with three major stories reshaping the AI race, semiconductor roadmap, and humanitarian response landscape.
Anthropic vs. Alibaba: AI distillation at "industrial scale"
Anthropic has sent a letter to U.S. Senators Tim Scott and Elizabeth Warren accusing Alibaba-linked operators of carrying out what it calls the largest "distillation attack" yet against its Claude AI model [1]. The San Francisco-based company alleges that operators used thousands of fraudulent accounts to conduct nearly 29 million exchanges with Claude, extracting responses to train a weaker model in a practice known as distillation [1]. Anthropic says the campaign targeted Claude's most valuable capabilities, including long-horizon reasoning and decision-making, and warned that such attacks turn "hundreds of billions of dollars in American investment and R&D into a massive subsidy for our geopolitical competitors" [1]. The letter also cited U.S. Department of Defense claims that Alibaba, Baidu, and BYD have ties to the Chinese military — allegations the firms deny — and noted that Alibaba has sued the U.S. government to be removed from a Pentagon blacklist [1].
IBM claims first sub-1nm chip breakthrough
In hardware, IBM unveiled a prototype chip design it says is equivalent to roughly 0.7nm — potentially the first known technology below the 1nm threshold [2]. The company's new "NanoStack" architecture layers transistor sheets vertically, which IBM compared to building a 100-storey skyscraper rather than a city of houses [2]. IBM claims the prototype delivered 50% better performance and 70% greater energy efficiency than its own 2nm chip in tests, and could eventually pack nearly 100 billion transistors onto a fingernail-sized surface [2]. Production remains years away, and rivals including Samsung and Intel are pursuing their own 3D stacking approaches [2].
Gemini 3.5 Flash gains native computer control
Google is also pushing agentic AI forward by integrating "Computer Use" directly into Gemini 3.5 Flash, allowing the model to see and operate screens across browsers, mobile devices, and desktops [3]. The feature, previously limited to a separate Gemini 2.5 model, scored 78.4 on the OSWorld benchmark — ahead of GPT-5.4 mini and matching Sonnet 4.6, though trailing Anthropic's Opus 4.8 [3]. Google is offering enterprise safeguards including user confirmation for sensitive actions and automatic stops when indirect prompt injections are detected [3].
World responds to deadly Venezuela earthquakes
Away from Silicon Valley, back-to-back earthquakes of magnitude 7.2 and 7.5 struck Venezuela, killing at least 164 people and injuring more than 970, according to Acting President Delcy Rodríguez [4]. The U.S. Geological Survey warned of probable high casualties and widespread disaster [4]. The Simon Bolívar International Airport in Caracas is closed due to damage, and a state of emergency has been declared [4]. Offers of aid have come from across Latin America, the U.S., China, India, Spain, Italy, Pakistan, and Türkiye, with El Salvador pledging 300 rescuers and 50 tonnes of supplies [4].
Together, these stories underscore a single theme: the boundary between technology and geopolitics — and between innovation and crisis response — is disappearing faster than ever.