OpenAI disbands its robotics team as its general models start running robots (fortune.com)

OpenAI has disbanded its standalone robotics team, integrating its researchers into broader teams focused on general AI models. This reorganization signals a strategic pivot towards leveraging general-purpose artificial intelligence, like GPT-6 Astra, for robotics applications. OpenAI claims that advancements in their foundational models can now manage robotic tasks previously requiring specialized training, viewing robotics as an extension of general intelligence rather than a distinct challenge. Despite this shift, OpenAI remains committed to its long-term vision of developing general-purpose robots, with CEO Sam Altman aiming for a humanoid robot demonstration by 2027. The implications of this move are significant for the AI/ML community, as it suggests a growing consensus that advanced AI models can directly control robots, potentially reshaping the robotics landscape. This challenges conventional wisdom among robotics startups that emphasize the need for specialized AI "brains" to facilitate robot functionality. Recent experiments indicate that general models can outperform specialized counterparts, hinting at a transformative era where the lines between AI and robotics are blurred. However, experts caution that while general models show promise, the deployment experience and specific training of existing robotics may still provide competitive advantages in the evolving market.