Anthropic Redirects 150 Engineers to Security Efforts

Anthropic stated it temporarily reassigned roughly 150 product engineers to security, reliability, and privacy work. Researchers also rotated out of pretraining as part of efforts to improve alignment and security practices. The company describes itself as an AI safety and research organization focused on reliable and steerable systems. In response, researcher Maksym Andriushchenko posted that he would not be surprised if 99 percent of researchers and engineers eventually work on safety-related topics.