Why a decade of doomsday warnings failed to slow the AI race | Technology
• OpenAI released its latest AI model, Astra, despite previous warnings about the existential risks of artificial intelligence. • The company paused some AI training development for two weeks following a 2026 security breach at Hugging Face. • Astra was launched with limited cybersecurity capabilities to mitigate potential threats to human safety.
theguardian.com





