Claude AI handles robot dog control 20 times faster than humans — a new stage in the evolution of physical agents
Anthropic has presented the results of the second phase of the Project Fetch experiment, and the results look impressive. The Claude Opus 4.7 model demonstrated the ability to configure and program a four-legged robot 20 times faster than teams of human engineers working with previous versions of AI. This is not just an improvement—it is a paradigm shift.
In August 2024, Anthropic employees with no experience in robotics attempted to program a robot dog using AI. At that time, the neural network only accelerated the search for solutions. Today, Claude Opus 4.7 acted almost autonomously, under minimal researcher supervision. The neural network independently completed the full cycle of work:
- connected to video sensors and LiDAR;
- wrote a program for manual control;
- created a motion 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 versions of AI, and 37 times faster than humans working without chatbot assistance. Moreover, the neural network wrote code that was 10 times more compact than human-written code. This indicates a deeper understanding of the task and efficient use of resources.
The authors of the experiment emphasize that progress in robotics has become a side effect of the general scaling of language models. Anthropic did not introduce specialized algorithms for controlling hardware. This means that improving basic AI architectures automatically leads to breakthroughs in related fields.
However, there were limitations. Claude still struggles with precise physical manipulations. The model managed to guide the robot to its goal but failed to gently push a ball to a specific point. Such actions require complex real-time feedback, where humans still maintain an advantage.
My Analysis
The fact that AI handles tasks for which it has no specialized algorithms is an alarming signal for the high-tech labor market. If a neural network can replace an entire team of engineers 20 times faster, this is not about automating routine tasks but about completely replacing intellectual labor. The industry is indeed entering the era of "physical AI agents," as Anthropic claims, and this will change not only robotics but the entire economy.