00:25What you see here is our third generation humanoid robot.
00:28We call it Mobi3.
00:31We already deliver our robot to our customer.
00:35So we iterate together with our customer, learn from the deployment.
00:53Labor shortage problem.
01:02So we are working with customers across manufacturing, semiconductor, and logistics for the tasks that
01:09are typically labor-intensive or even hazardous for their people.
01:14So tasks including repetitive tasks, what you see here is called material induction.
01:18It's a very common task that happens in almost all industries like construction, logistics, and manufacturing.
01:47Victor, our robotics engineer, is from one of the top robotics labs in the world.
01:51So he is currently controlling the robot.
01:55That serves two purposes.
01:57First, we use that approach to train the robot skills.
02:01There's no programming.
02:03There's no C++, no Python through this process.
02:07Victor is using the demonstration to show the robot how the task is done.
02:12Then that becomes data.
02:14Now we use that to train the model.
02:15And when we are operating autonomously, Victor is standby and ready.
02:21So when the robot makes a mistake, Victor can take over and teach the robot the correct behavior.
02:40Our target for training the robot skills accomplished tasks.
02:43We try to make it as fast as onboarding a person.
02:46We are not there yet.
02:47But we see the development of AI really give us the line of sight to achieve that goal.
02:54So once you have achieved that goal, imagine when we ship a robot to a factory.
02:59You open up the package.
03:00In two hours, the robot starts creating value for a customer.
03:04That's the goal we want to achieve.
03:24In my head.
03:26Yes.
03:28.
03:42.
03:53There's no self-praigency on our website.
03:54Yes, we want to go before it's going to beaster
03:59You
