The AI model Claude Opus 4.7 outperformed humans by dozens of times in controlling a robot dog.
Anthropic continues to surprise the market: as part of the updated Project Fetch experiment, the Claude Opus 4.7 model demonstrated an impressive leap in autonomy. The neural network completed a full cycle of tasks for setting up and controlling a robot dog 20 times faster than teams of human engineers working with previous versions of AI.
While in August 2024, employees with no robotics experience only received hints from AI, now Claude Opus 4.7 acted almost independently. Under minimal researcher supervision, the neural network:
- connected to video sensors and lidar;
- wrote a program for manual control;
- created a trajectory monitoring system;
- configured an object recognition algorithm.
The key metric is speed. Opus 4.7 turned out to be 18 times faster than a team using older AI versions, and 37 times faster than people working without chatbot assistance. At the same time, the volume of code generated by the model was 10 times less than that of human teams. This indicates higher efficiency and conciseness of the solutions found by the neural network.
An important nuance: Anthropic emphasizes that progress in robotics was a side effect of general scaling of language models. The company did not implement specialized algorithms for controlling hardware. This opens up broad prospects for applying universal AI in the physical world.
However, a complete victory is not yet achieved. Claude still struggles with precise physical actions: the model managed to guide the robot to the target but could not gently push a ball to the right spot. Such tasks require complex real-time feedback, where humans still maintain an advantage.
Anthropic predicts that the industry is entering an era of "physical AI agents." In the future, neural networks will be able to work with standard tools and equipment as effectively as they currently do with software code.
My comment: This experiment is a clear signal for investors and developers: the boundary between the digital and physical worlds is rapidly blurring. Companies that integrate AI into hardware solutions now will gain a tremendous competitive advantage in the next 2-3 years. However, we should not forget: even the smartest neural network is not yet capable of replacing humans in tasks requiring fine motor skills and instant adaptation.