OpenAI, an artificial intelligence research company, has been studying machines for years that can learn and do things like solve a Rubik’s Cube. The company has now disbanded the robotic team responsible for the study. One of the company’s co-founders told Podcast that OpenAI focuses on areas where data is more accessible.
this company announced Which can be further improved by having ready data. He also said that the decision to disband the robotics team was a very difficult decision, but it was necessary for the development of the company and its future prospects.
A company spokesman said in an interview that in the fall of last year, after a Rubik’s Cube machine learning project, a decision was made not to continue robotics studies and to focus on other projects instead. The reason for this decision was to accelerate the development of artificial intelligence capabilities, which was achieved with the help of reinforcement learning with human feedback.
OpenAI launched its first massive robotics project in the fall of 2019. At the time, he published his first study of a five-finger robot that, at its best, could successfully solve a Rubik’s Cube between 20 and 60 percent of the time required. The robot was also able to find a way to solve the challenge when its fingers were tied. He could even work with leather gloves.
This robot was the result of two years of efforts of this company. In the spring of 2017, OpenAI introduced Roboschool, an open source software designed to control robots in a simulation. In the same year, the company unveiled a robot that was fully trained in a simulation and could learn and do something at a glance.