Last summer, Google launched a vision-language-­action mannequin known as RT-2. This model gets its common understanding of the world from the online textual content and pictures it has been skilled on, in addition to its personal interactions. Thanks to the AI growth the focus is now shifting from feats of bodily dexterity achieved by expensive robots to constructing “general-purpose robot brains” within the form of neural networks. Instead of the standard painstaking planning and training, roboticists have started using deep studying and neural networks to create methods that study from their setting on the go and adjust their conduct accordingly. …