Robots are getting better at handling items. We have already covered plenty of them using soft grippers and other mechanisms to handle fragile items of all shapes. OpenAI has taken it to the next level. It has trained a human-like robot hand to manipulate objects with amazing dexterity.
The Dactyl robot system was trained in simulation and can adapt to the real-world. Here is how it works:
Dactyl learns to solve the object reorientation task entirely in simulation without any human input… Dactyl was designed to be able to manipulate arbitrary objects, not just those that have been specially modified to support tracking. Therefore, Dactyl uses regular RGB camera images to estimate the position and orientation of the object.
More info is available here.