Claude Opus 4.7 destroyed the engineering team: controlling a robot dog 20 times faster
The artificial intelligence market continues to surprise with its pace of development, and Anthropic's latest experiment is a clear testament to that. As part of the second phase of the Project Fetch initiative, the Claude Opus 4.7 model demonstrated the ability to fully autonomously configure and control a robot dog, completing tasks 20 times faster than the best team of human engineers working with the previous version of the AI.
While in August 2024 the neural network only helped employees without robotics experience find solutions more quickly, the situation has now changed dramatically. In the new testing phase, Claude Opus 4.7 operated virtually without researcher oversight, taking on the entire cycle of tasks:
- connecting to video sensors and LiDAR;
- writing a program for manual control;
- creating a system for monitoring the robot's trajectory;
- configuring an object recognition algorithm.
The numbers are impressive: compared to a team using older AI versions, Opus 4.7 was 18 times faster, and compared to humans working without chatbot assistance, it was 37 times faster. Moreover, the neural network wrote code 10 times more compactly than human teams, indicating higher algorithm efficiency.
It is important to emphasize: Anthropic did not introduce specialized algorithms for controlling the hardware. The progress in robotics is a side effect of the general scaling of language models. This means that with each new generation of AI, we will observe exponential growth in related fields.
However, there were limitations. Claude still struggles with precise physical actions in real time. The model successfully guided the robot to its target but could not gently push a ball to the exact spot. This requires complex feedback, where humans still maintain an advantage.
Anthropic is confident: 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 currently work with software code.
Analyst's opinion: This experiment clearly demonstrates that we are on the verge of a fundamental shift. The speed at which AI is mastering the physical world exceeds the most optimistic forecasts. For the market, this means not only new investment opportunities in the robotics sector but also serious risks for traditional professions related to equipment operation.