AI agent took control of a robot dog: Claude Opus 4.7 outperformed human engineers by 20 times
I am closely following the development of Project Fetch, and the latest results are truly impressive. In the updated experiment, the Claude Opus 4.7 model fully took control of the robot dog, completing setup and programming tasks 20 times faster than teams of human engineers.
Back in August 2024, Anthropic employees with no robotics experience attempted to program a four-legged robot, with AI serving only as an assistant. However, in the new testing phase, Claude Opus 4.7 worked almost autonomously, under minimal researcher supervision. The neural network independently connected to video sensors and LiDAR, wrote a manual control program, created a robot path monitoring system, and configured an object recognition algorithm.
The key metric is speed. The Opus 4.7 model proved to be 18 times faster than a team using older AI versions and 37 times faster than humans working without chatbot assistance. Moreover, the neural network wrote much more efficient code: its volume was 10 times smaller than that of human teams.
It is especially telling that progress in robotics has become a side effect of the general scaling of language models. Anthropic did not implement specialized algorithms for controlling hardware — this is a pure advantage of the base architecture.
However, there were limitations. Claude still struggles with precise physical actions. The model managed to guide the robot to the target but failed to gently push a ball to the exact spot. This requires complex real-time feedback, an area where humans still outperform AI.
Anthropic predicts that the industry is entering an era of "physical AI agents." In the future, neural networks will be able to use standard tools and equipment as effectively as they currently work with code.
My comment: This experiment is a powerful signal for investors in AI infrastructure. If language models without specialized tuning are already capable of controlling physical devices at a level exceeding human performance, we are on the verge of a massive productivity leap in industrial robotics and automation. The next logical step is the emergence of AI agents that can not only write code but also physically interact with the world.