IT之家 11 月 14 日消息,《财富》昨日(11 月 13 日)发布博文,爆料称 OpenAI 已解散独立机器人团队,并将相关人员调至公司其他部门。OpenAI 向该媒体确认了此次组织调整,只是强调本次调整并非退出机器人领域,机器人仍是其长期愿景的核心,重组旨在加快研发。
在架构调整维度,机器人研究人员并入核心研究团队,向首席研究官马克 · 陈(Mark Chen)汇报;应用机器人与硬件团队则转入规模更大的应用部门,该部门由公司总裁格雷格 · 布罗克曼(Greg Brockman)负责。IT之家附上相关截图如下;d
对于本次架构调整的根源,OpenAI 官方回应称涵盖 GPT-6 Astra 在内,通用模型已能直接用于部分机器人任务,因而决定将相关研究进一步整合到通用模型研发中。
OpenAI 希望让能理解文字、图像和代码的 AI,也成为机器人的“大脑”,而不再完全依赖专门为机器人开发的模型。报道援引英伟达一项研究称,Astra 驱动的智能体在一项模拟操作测试中取得 74.8% 的成功率,高于该榜单领先专用模型的 51.0%。
通用模型能否全面接手机器人,仍有现实难题。受访学者指出,大模型运行速度和云端联网需要构成限制,机器人企业积累的现场经验与数据也仍有价值。据报道,OpenAI 正研发自有人形机器人,目标于 2027 年首次公开演示。
本文来源:https://www.ithome.com/1/027/004.htm
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。
Thanks to IT Home user "Fly Guy's Left Hand" for submitting this tip!
IT Home reported on November 14 that Fortune published an article yesterday (November 13) revealing that OpenAI has disbanded its standalone robotics team and reassigned related personnel to other departments within the company. OpenAI confirmed this organizational restructuring to the outlet, emphasizing that this adjustment does not represent an exit from the robotics field, as robotics remains at the core of its long-term vision, and that the reorganization is aimed at accelerating research and development.
In terms of the structural adjustment, robotics researchers have been merged into the core research team and will report to Chief Research Officer Mark Chen; the applied robotics and hardware teams have been transferred to the larger-scale applications division, which is led by company President Greg Brockman. IT Home attaches the relevant screenshot below:
As for the reason behind this structural adjustment, OpenAI officially responded that general models, including GPT-6 Astra, can now be directly applied to certain robotics tasks, and therefore the company has decided to further integrate related research into the development of general models.
OpenAI hopes to enable AI capable of understanding text, images, and code to also serve as the "brain" of robots, rather than relying entirely on models developed specifically for robotics. The report cites a study by NVIDIA stating that an Astra-powered agent achieved a 74.8% success rate in a simulated manipulation test, higher than the 51.0% posted by the leading specialized model on the leaderboard.
Whether general models can fully take over robotics operations still presents practical challenges. Scholars interviewed pointed out that the operating speed of large models and the need for cloud connectivity constitute limitations, and the on-site experience and data accumulated by robotics companies still hold value. According to reports, OpenAI is developing its own humanoid robot, targeting its first public demonstration in 2027.
