Anthropic has proven: AI controls a robot dog 20 times faster than a human — a new frontier for physical agents
The artificial intelligence market continues to surprise: this time, Anthropic's Claude Opus 4.7 model, as part of the updated Project Fetch experiment, demonstrated the ability to configure and control a four-legged robot at speeds 20 times faster than human engineering teams. This is not just a test—it is a signal that AI is beginning to transition from a purely digital environment into the physical world.
Autonomy Without Intermediaries
While in the first phase of the experiment (August 2024), AI acted only as an assistant for people without robotics experience, the picture has now radically changed. Claude Opus 4.7 operated almost entirely autonomously, under minimal researcher supervision. The neural network independently completed four critical stages: connecting to video sensors and LiDAR, writing a manual control program, creating a trajectory monitoring system, and configuring an object recognition algorithm.
Numbers That Speak for Themselves
Comparative analysis shows a colossal leap in performance. The Opus 4.7 model proved to be 18 times faster than teams using previous AI versions and 37 times faster than people working without chatbot assistance. At the same time, code quality also improved: the volume of code written by the neural network was 10 times less than that of human teams. This means not only speed but also efficiency—fewer lines, fewer errors.
Side Effect of Scaling
Notably, Anthropic did not introduce specialized algorithms for controlling the hardware. The progress in robotics became a side effect of the general scaling of language models. This discovery confirms the hypothesis that universal AI systems can adapt to tasks for which they were not originally trained.
However, there were limitations. Despite successfully guiding the robot to its goal, the model failed at the task of gently pushing a ball to a specific point. This requires complex real-time feedback—an area where humans still maintain an advantage.
Looking to the Future
Anthropic is confident that the industry is entering an era of "physical AI agents." In the near future, neural networks will be able to use standard tools and equipment as effectively as they work with software code today. This raises new questions for the market—from ethics to safety.
Expert opinion: The results of Project Fetch are not just another performance record. They demonstrate that AI is ceasing to be a tool for text generation and is beginning to explore the real world. However, the incident with the ball reminds us that full-fledged physical intelligence is still far off. Investors should closely monitor the development of this niche—this is where the next major breakthrough may lie.