Watch Tesla Optimus doing house chores, cooking, helping in car manufacturing and more

Tesla has published a new video of the Optimus robot performing productive tasks, including house chores, cooking dinner, and helping on the car manufacturing line.

A 60-second video showed the robot picking up a garbage bag, opening the lid, and placing the bag in the bin. Up next, Optimus is cleaning up the table with a brush and dustpan. The bot tears a paper towel from the roll. It used a vacuum cleaner to clean the floor with swift hand movements.

Besides house chores, the robot was spotted doing some cooking tasks, including picking up the spoon and stirring the pot. That’s not it, the company showcased the next part of the clip, including Optimus picking up the Model X fore link from the right cardboard box and placing it on the right ramp of the dolly.

Some other tasks showcased in the video are:

• Opening the cabinet
• Push the microwave button
• Opening/Closing the curtain
• Ironing cloths
• Placing an object into a place
• Cleaning the whiteboard
• Watering the plants

Tesla has confirmed that these tasks are done by a single neural net and learned directly from human videos. The model training has helped the robot to learn faster, and more developments are still underway.

Milan Kovac, Chief of Optimus Program, explained the latest improvements via his post on X.

“One of our goals is to have Optimus learn straight from internet videos of humans doing tasks. Those are often 3rd person views captured by random cameras etc.

We recently had a significant breakthrough along that journey, and can now transfer a big chunk of the learning directly from human videos to the bots (1st person views for now). This allows us to bootstrap new tasks much faster compared to teleoperated bot data alone (heavier operationally).

Many new skills are emerging through this process, are called for via natural language (voice/text), and are run by a single neural network on the bot (multi-tasking).

Next: expand to 3rd person video transfer (aka random internet), and push reliability via self-play (RL) in the real-, and/or synthetic- (sim / world models) world.”