IT之家 AISource publication

Tesla is building a "school" for Optimus, and Musk says training robots is like raising children

IT之家 News on October 23: Tesla plans to send its humanoid robot Optimus to "school," allowing the robot to master more skills through practice.

IT之家 News on October 23: Tesla plans to send its humanoid robot Optimus to "school," allowing the robot to master more skills through practice.

According to Business Insider, on the 21st local time, Tesla CEO Elon Musk said on the earnings callComparing training Optimus to raising a young child. He explained that Tesla's autonomous driving system can learn from the data accumulated by its massive fleet, but Optimus does not have such conditions and can onlyBy observing humans work, and then practicing personallyto master skills, in a way similar to how humans learn.

Musk told investors: "Just look at little kids. When they first learn to walk, they wobble and can't do anything, and then they finally learn to wipe their own bottoms—that's a major milestone. Then they learn to read and learn to do all sorts of things, and all these skills are learned through play. We have to let Optimus do this kind of training at scale as well, so that it can truly handle all kinds of real-world tasks."

Tesla is also building a training facility called "Optimus Academy," where robot prototypespractice various tasks and accumulate datato be used for training robots introduced later. "To collect enough data to train Optimus, we first have to build a large number of robots. I think we'll need roughly 10,000 to 20,000 units, and then have them train autonomously."

Tesla's Fremont factory has already begun producing Optimus, and Musk expects customer deliveries to start by the end of 2027.

Musk, whose net worth is close to $1 trillion (IT之家 note: approximately RMB 6.69 trillion at the current exchange rate), also reminded investors that the production process may not be smooth sailing, as there are still many challenges to solve in humanoid robot development. The report noted that making robotic hands both durable and dexterous, and developing AI systems capable of handling open-ended tasks, are among Tesla's main challenges.

Related reading:

Original source

IT之家 AI

Content notes

Original publication and rights belong to the source.

Machine translation · Refer to the original