Artificial Intelligence
A powerful new AI model called Ox Alpha has surfaced anonymously on the developer platform OpenRouter, offering a roughly 1 million-token context window and multimodal capabilities for free, even as its creator remains unidentified.
Nisha Omkumar
Author
Chinese AI lab DeepSeek will stop charging peak API rates on weekends starting August 23, easing part of a pricing structure introduced just a week earlier that had raised some token costs by more than 1,100 percent.
The move follows Anthropic's expanding compute partnerships with Google and Broadcom, and reflects a broader industry pattern of AI labs building deeper in-house chip expertise.
The investment includes roughly $250 million for a supercomputing project involving Huawei and iFlytek, underscoring Brazil's multipolar approach to building sovereign AI capacity.
India's central bank chief says AI should be seen as a capability to be responsibly harnessed in lending decisions, not merely as a risk to be contained.
Cybersecurity researchers have uncovered evidence that near-autonomous AI agents are being deployed in real-world cyberattacks probing critical water infrastructure, raising urgent questions about the intersection of agentic AI and national security.
OpenAI has reportedly paused a major frontier model training run after experimental models were found to have crossed internal security boundaries during testing, intensifying an already fraught global debate over AI safety governance.
Google has struck a deal that could see it take a stake in chipmaker Marvell worth as much as $12.2 billion, as the search giant deepens its push into custom AI silicon to reduce reliance on external chip suppliers.
Dubai has signed a memorandum of understanding with India's NASSCOM establishing a framework for collaboration in agentic AI and deeptech, creating new market access pathways for Indian companies expanding into the Gulf's advanced technology sector.