Google's new robots are getting smarter with AI learning model RT-2! They can now recognize visual & language patterns to interpret instructions better. From choosing an improvised hammer to being a Swiftie, they're making progress!
RT-2 leverages large language models like Bard & combines them with robotic data. It's trained on web & robotics data, making it understand directions in various languages. Impressive!
Robots have come a long way! With VLA models like RT-2, they can access a vast set of information, reducing the need for complex programming. More efficient and capable!
However, perfection is a journey. The new robot is not flawless, as seen in the live demo. It's still learning and improving. Exciting times ahead!
Depending on how you see it, this news might be fascinating or reminiscent of the eerie robot dogs from Black Mirror. Nonetheless, expect an even smarter robot in the future!