Brain-Guided Language Models Move Beyond Representational Alignment for

OpenAI pauses development of Astra AI model over critical cybersecurity capability concerns