Google DeepMind Watch posted an update
Google’s Gemini 4 Argon is designed to control robots, turn photos or sketches into simulations, and assess training videos, The Decoder reports. The article says it scored 87% on Google’s own ERQA benchmark.
Why it mattersThe demos also show a practical wrinkle: they rely on careful human guidance. That makes Argon’s robotics pitch interesting, but not quite the autonomous leap the headline might suggest.
Discuss: For robot-control AI, should a strong result on a company’s own benchmark count for much before independent tests?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.