Summary

  • Google’s AI subsidiary, DeepMind, has combined its top large language model with robotics to create Gemini Robotics, which it says allows robots to be more dexterous, work from natural-language commands, and generalise across tasks, all of which have been difficult for robots to achieve previously.
  • Examples of uses for the new model include helping robots to play soccer, enhancing their capabilities when visibility is limited, and using AI-generated images to teach robots how to act.
  • Gemini Robotics is the latest in a series of projects that combine LLMs with robotics and is part of a trend that is seeking to make robots more useful.
  • Other stories in today’s edition of The Download include European Union pushing back against Donald Trump’s tariffs by slapping the US with its own levies, big tech calling for new nuclear power stations with the notable exception of Microsoft, and Tesla’s competitor BYD quickly gaining on it in the electric car market.

By Rhiannon Williams

Original Article