Mentioned:
There’s a new leading edge vision-language-action model for robotics. And Google’s DeepMind is behind it.
What’s happening:
- Google's (NASDAQ: GOOG) subsidiary DeepMind has officially launched their new artificial intelligence model known as Gemini Robotics 2
Why it matters:
- Gemini Robotics 2 is the most advanced vision-language-action model that DeepMind has ever developed and enables robots to walk, crouch and manipulate objects under one unified control system
Going deeper:
- Gemini Robotics 2 vision-language-action model has been tested with three different robot embodiments and also has a lighter, on-device version that can run locally without any connectivity
- Apptronik’s Apollo 2 humanoid robot was used to test DeepMind’s Gemini Robotics 2 vision-language-action model with general manipulation tasks including picking items from a table and multi-finger dexterity tasks such as tying a knot
The intrigue:
- Google previously made waves when they participated in Apptronik’s $935M USD Series A financing round alongside of Mercedes-Benz, AT&T Ventures and the Qatar Investment Authority
