Tesla 工人抵触为 Optimus 人形机器人录制训练数据,担心被其取代

Ars Technica:AI(RSS)·2026-09-26 05:10·18分钟前· Jeremy Hsu
AI 导读

Tesla 的人形机器人 Optimus 面临制造瑕疵和触觉传感器不可靠等问题,AI 能力尚不足以通用操作,仍需在受控环境中针对特定任务编程。由于依赖模仿学习,Tesla 曾让得州和加州工厂工人穿特殊服装录制动作,但工人因担心机器人最终取代自己而抵触,Tesla 已改为专职团队和训练中心负责数据采集。

Ars Technica:AI(RSS)
63AI 编辑部评分,满分 100

Tesla 工人抵触为 Optimus 人形机器人录制训练数据,担心被其取代

2026-09-26 05:10· 18分钟前· Jeremy Hsu
AI 导读

Tesla 的人形机器人 Optimus 面临制造瑕疵和触觉传感器不可靠等问题,AI 能力尚不足以通用操作,仍需在受控环境中针对特定任务编程。由于依赖模仿学习,Tesla 曾让得州和加州工厂工人穿特殊服装录制动作,但工人因担心机器人最终取代自己而抵触,Tesla 已改为专职团队和训练中心负责数据采集。

The imperfect and manually intensive manufacturing process has led to newly produced robots that require immediate fixes. Touch sensors for the robot’s hands have also proven unreliable enough so that Tesla has now developed a glove-like layer of sensors that can be replaced without replacing the entire robot hand.

The robot’s AI capabilities are also reportedly still insufficient for general-purpose operations, which is unsurprising given the challenges facing every robotics company on that front. Instead, one of The Information’s sources described Tesla’s Optimus robots as currently requiring programming to do specific tasks in carefully controlled environments.

One of the biggest challenges facing all robotics companies is getting enough training data to help robots visually learn a wide variety of manual tasks. The heavy reliance on imitation learning prompted Tesla to have factory workers in Texas and California wear special suits designed to record their physical movements while working.

However, The Information described some workers complaining because they “knew the robots were designed to eventually replace them.” So Tesla has apparently shifted that data collection responsibility to dedicated teams and has also set up “training hubs” for such teams.

Competition to make robots

The Information also makes a point of highlighting Tesla’s continued reliance on Chinese suppliers to make various robot components. The Information previously reported on Silicon Valley startups bringing robotic parts from China to the United States in their luggage.

The US robotics industry’s supply chain reliance on China has persisted despite the Trump administration trying to boost domestic supply chains and robotic production. In July, the Federal Communications Commission banned new foreign-made robots such as humanoid robots and four-legged robot dogs, along with robot vacuum cleaners.

Many of Tesla’s reported challenges in making humanoid robots are not unique. But that may be little comfort when Tesla also faces stiff competition on multiple fronts, as automakers in China, Japan, and South Korea are also developing humanoid robots, not to mention dedicated robotics companies pursuing humanoids. Toyota plans to invest billions of dollars in upgrading factories with robots, including some humanoids, while Hyundai is planning to deploy up to 25,000 Atlas humanoid robots developed by US subsidiary Boston Dynamics over the next several years.

One of the companies furthest along in commercial deployment of humanoid robots is Oregon-based Agility Robotics, which first put them to work at an Atlanta-area warehouse owned by GXO Logistics in 2024. But the overall business case for humanoid robots still has to be proven through more sustained and cost-effective deployments—not to mention showing that such robots can work safely around humans.

来源:Ars Technica:AI(RSS)· arstechnica.com