{"id":23060,"date":"2025-05-23T09:21:27","date_gmt":"2025-05-23T09:21:27","guid":{"rendered":"https:\/\/metaverseplanet.net\/blog\/?p=23060"},"modified":"2026-01-06T13:04:48","modified_gmt":"2026-01-06T13:04:48","slug":"apple-starts-using-vision-pro-on-robots-after-failing-to-sell-it-to-people","status":"publish","type":"post","link":"https:\/\/metaverseplanet.net\/blog\/apple-starts-using-vision-pro-on-robots-after-failing-to-sell-it-to-people\/","title":{"rendered":"Apple Starts Using Vision Pro on Robots After Failing to Sell It to People"},"content":{"rendered":"\n

Apple’s mixed reality headset<\/strong>, the Vision Pro<\/strong>, which failed to generate the expected interest among consumers, has now become the focus of a scientific experiment. The U.S.-based tech giant has begun using the Vision Pro on humanoid robots<\/strong> as part of a research initiative. The aim of the project is to train humanoid robots<\/strong> using Vision Pro. But was the research team successful?<\/p>\n\n\n\n

A few years after launching the Vision Pro<\/a><\/em><\/strong> to lackluster consumer response, Apple<\/a><\/em><\/strong> decided to repurpose the device in a scientific setting. The Vision Pro is now being utilized to aid in the training of humanoid robots<\/strong>, turning the once commercial failure into a valuable research tool.<\/p>\n\n\n\n

Apple collaborated with scientists from MIT<\/strong>, Carnegie Mellon<\/strong>, University of Washington<\/strong>, and UC San Diego<\/strong>. Together, they compiled a massive dataset consisting of over 25,000 human activities<\/strong> and 1,500 humanoid robot activities<\/strong>. This dataset was then used to develop a powerful AI system<\/strong>, intended to make humanoid robots behave more like real humans.<\/p>\n\n\n\n

\n\n\n\n

Vision Pro Speeds Up and Lowers the Cost of Training Humanoid Robots<\/h2>\n\n\n\n
\"\"<\/figure>\n\n\n\n

One of the biggest challenges in humanoid robotics is the slow and expensive training process<\/strong>. However, using the Apple Vision Pro<\/strong>, the team was able to overcome these obstacles. The device enabled the researchers to record high-resolution videos quickly<\/strong>, and with a specially designed algorithm, they could slow down these videos by up to 4 times. This allowed the robots to capture and learn even the smallest details.<\/p>\n\n\n\n

As a result, the training process became faster and more cost-effective<\/strong>. The method is detailed in a research paper titled “Humanoid Policy \u223c Human Policy”<\/strong>, published on GitHub<\/strong>. Although the paper presents Apple\u2019s technique as just a proposal, the implications are significant.<\/p>\n\n\n\n

With the humanoid robotics industry rapidly advancing<\/strong>, companies like Tesla<\/strong>\u2014which are already investing in humanoid robot development\u2014could potentially adopt Apple\u2019s Vision Pro approach to accelerate training.<\/p>\n\n\n\n

\n