Researchers at Carnegie Mellon University have presented the initial results of a new system for robot preparation called “H2O” (Human2HumanOid). Created by Tehkult, this system allows human operators to train a full-scale humanoid robot to perform various actions. Training is done using a digital photo catalog that teaches the robot the physical analog of human movements. The system has a limiter that optimizes the robot’s movements based on its design’s physical capabilities. Currently, there is no separate locomotion system for fine movements, which poses a challenge in ensuring movement accuracy. The robot constantly communicates with each task to evaluate its success and optimize the results. Researchers envision that this system will enable robots to perform more complex tasks and use non-standard manipulative tools. In the long term, these robots could replace human labor in hazardous conditions, eliminating the need for large numbers of migrant workers. The goal is to bring industrial production back to developed countries with constant monitoring and evaluation by researchers.