Meet Chinese humanoid robots learning to steam clothes, make sandwiches

The Manila Times

Meet Chinese humanoid robots learning to steam clothes, make sandwiches  In a warehouse on the outskirts of Shanghai, nearly a hundred humanoid robots are learning to make sandwiches, clean counters and arrange flowers for 17 hours a day. Their repetitive movements are not just for show—they are part of a data-driven effort to advance AI-powered automation. Chinese robotics firm Agibot is developing humanoid robots capable of handling everyday tasks, from household chores to retail work, as part of a broader push to integrate artificial intelligence into daily life. The company has set up simulated environments, such as a fully stocked convenience store and a furnished apartment, where data acquisition robots, called the Agibot Genie-1, practice real-world tasks with human controllers. Each action generates data that is fed into Agibot’s central embodied AI system, refining its large-scale learning models.  REUTERS VIDEO  Subscribe to The Manila Times Channel - https://tmt.ph/YTSubscribe    Visit our website at https://www.manilatimes.net    Follow us:  Facebook - https://tmt.ph/facebook  Instagram - https://tmt.ph/instagram  Twitter - https://tmt.ph/twitter  DailyMotion - https://tmt.ph/dailymotion    Subscribe to our Digital Edition - https://tmt.ph/digital    Check out our Podcasts:  Spotify - https://tmt.ph/spotify  Apple Podcasts - https://tmt.ph/applepodcasts  Amazon Music - https://tmt.ph/amazonmusic  Deezer: https://tmt.ph/deezer  Tune In: https://tmt.ph/tunein   #TheManilaTimes #tmtnews #robots  #humanoid #artificialintelligence

Transcript

00:30The robots will start to be useful first in some structural environments and also the

00:45tasks are quite clear and specific because it's a contained kind of environment.

00:52You can exactly tell the robots to do repetitive but maybe boring jobs and replace people at

00:59or we should say free people from those very boring or hazardous jobs.

01:07That's the first step and the second step I think it will gradually enter some semi-structural

01:13environments like grocery stores.

01:40When we reach a certain amount of data I think the generative AI's capability will give us

01:46a combinatorial generalization capability where the robots can first understand open

01:56and vague natural language instruction and second can imagine and plan combinations of

02:03different skills and objects.

02:06And at that time I think we could have a chance to invite the robots in our home.

02:36Because of China just like Japan or other East Asia countries is facing a very hard

02:56challenge on the aging problem for the whole society so elderly care is a very, very large

03:04market and there's less people, well-trained people can do those jobs.

03:10So people definitely expect the robots in five or ten years can do those works to for

03:19example organize the room for the senior people or to pick up a package or some food delivery

03:28and even to transfer people from bed to washrooms.

03:58The payload of the robot arms these days are still a few kilograms, it cannot hold

04:09a whole people like adult people.

04:12So we are doing some very simple household jobs for the elderly people these days.

04:20As I mentioned bring the water to throw a trash or pick up a package or to monitor the

04:27health condition.

04:28This I think is quite important, monitor the safety and the health condition of those elder

04:34people.

04:57Robots can do routine tasks like people and especially manipulations using the arms and

05:13the hands because those are where the values are created.

05:18So we invest a lot in AI development including our computing power, ML infrastructure.

05:49We want to use general purpose robots to create unlimited productivity.

05:56Just imagine that one day in our own robot factory our robots are assembling themselves.

Category

Transcript

Recommended