OpenAI’s decision not to release GPT-6.1 Astra has brought a critical question in artificial intelligence development into sharper focus: how capable should an AI system become before developers are confident that its behavior can be reliably controlled? The company said the latest version of Astra did not meet its required safety threshold, particularly in areas involving authorization, staying within the intended scope of a task, and accurately communicating what work the system had...

The artificial intelligence infrastructure race is entering a new phase. OpenAI and Anthropic have committed to enormous amounts of computing capacity, with projects measured in gigawatts and investments extending across multiple cloud and data center partners. Yet reports that both companies are exploring smaller data center deployments of roughly 20 to 30 megawatts reveal an important shift in how AI computing may be built and delivered. The move does not signal the end of giant AI...

OpenAI’s relationship with Samsung Electronics is moving beyond the conventional boundaries of an AI software partnership, placing semiconductors, memory technology, enterprise deployment, and large-scale computing infrastructure at the center of a rapidly expanding strategic relationship. Recent comments from OpenAI Korea suggested that joint research and production of next-generation chips had become one of the areas where the two companies had made significant progress. OpenAI subsequently...