Crypto news

Anthropic has unveiled new results from the Project Fetch experiment, and they are impressive. The Claude Opus 4.5 model took full control of a four-legged robot, completing setup and programming tasks 20 times faster than a team of human engineers working with previous AI versions. This is not just an improvement — it's a paradigm shift.

In August 2024, when Project Fetch first launched, the AI only acted as an assistant: it helped employees without robotics experience find solutions more quickly. Today, the situation has changed dramatically. In the new testing phase, Claude Opus 4.7 worked almost autonomously. With minimal researcher oversight, the neural network independently:

• connected to video sensors and LiDAR;
• wrote a program for manual robot control;
• created a trajectory monitoring system;
• configured an object recognition algorithm.

The numbers speak for themselves. Opus 4.7 proved to be 18 times faster than a team using older AI versions, and 37 times faster than people working without chatbot assistance. But the key factor is code quality. The volume of code written by the neural network turned out to be 10 times smaller than that of human teams. This means not only speed but also efficiency: fewer lines mean fewer errors and easier maintenance.

Particularly noteworthy is the fact that Anthropic did not implement specialized algorithms for controlling the hardware. As the authors note, progress in robotics has become a side effect of the general scaling of language models. This confirms the hypothesis that fundamental AI models can adapt to physical tasks without additional tuning.

However, it is too early to idealize the situation. Claude still struggles with precise physical actions. The model successfully guided the robot to its target but failed at the task of gently pushing a ball to the right spot. This requires complex real-time feedback — an area where humans still maintain superiority.

At Anthropic, they believe the industry is entering an era of "physical AI agents." And I agree. In the coming years, we will see neural networks begin to use standard tools and equipment as effectively as they currently work with software code. This opens up enormous opportunities for automating manufacturing, logistics, and even household tasks.

My expert conclusion: the pace of progress in this field has exceeded my expectations. If a year ago we were talking about auxiliary tools, today we are discussing fully autonomous agents. The job market for engineers and programmers is facing a serious transformation — and it has already begun.