This article first appeared on GuruFocus.
Google DeepMind (NASDAQ:GOOG) is giving robots a major brain upgrade. On Thursday, the company rolled out two new AI modelsGemini Robotics 1.5 and Gemini Robotics-ER 1.5that are designed to help machines plan, reason, and even use the internet to solve problems.
The first model, Gemini Robotics 1.5, acts like a vision-language-action system. In plain terms, it thinks before it moves, showing its reasoning process before making a decision. That makes it better at handling tricky, multi-step jobs and learning skills across different types of robots.
The Robotics-ER 1.5 model goes a step further. It can reason about the physical world, pull in outside tools like Google Search, and create detailed plans to complete a mission. Alphabet CEO Sundar Pichai said the upgrades bring robots closer to being general-purpose helpers that can transfer knowledge from one machine to another.
Developers can already start experimenting with the Robotics-ER 1.5 model via Google’s AI Studio, while the Robotics 1.5 system is being rolled out to select partners.