The State of Current Robotics

in StemSocialyesterday

The State of Current Robotics



Image source:


According to the vision of Boston Dynamics


When the famous Boston Dynamics, perhaps the best known among those who enjoy the world of robots, decides to speak openly, it is because something important is happening. Amid an industry accustomed to fast-paced announcements and promises of full autonomy, the company presented a rare analysis of the real state of humanoid theft, and the message is direct.


According to them, the entire industry is still only in phase one, behind the new all-electric Atlas emerges a much more cautious development philosophy than that of competitors who boast months of uninterrupted operation; Alberto Rodríguez, former MIT professor and director of Boston Dynamics, divided the life cycle of a robot into three parts.


First, build the machine and win the reliability battle, second, survive the real test with the customer, dealing with the realization that almost all the assumptions were wrong, third and last, scale to thousands of units and deal with long-term wear and tear issues.




His assessment was dry.


All humanoids, without exception, are still stuck in the first stage, automotive manufacturing then appears as the ideal testing ground, not only because it is a large market, but because it represents the maximum point of physical manipulation, complex tasks, extreme precision, infinite variation.


Rodríguez described the scene as an emblematic example, fishing a loose screw from a container, orienting it in the air and inserting it into the body of a car using the other hand; if a robot masters that, it masters almost everything.


The automotive industry demands exactly this kind of skill on lines that can produce five different models with thousands of part variations, but perhaps the most profound change is happening in the brains of these machines.


The old processing pyramid is giving way to end-to-end neural networks, the term physical AI adopted by Hyundai summarizes this transition, it is no longer about perfect code, it is about systems that fail, learn and continue, while some competitors turn every milestone into marketing, Boston Dynamics seems to be betting on a slow maturation strategy, I just can't say how much of this is the company's reality or its capacity limit.


Other companies, both from the United States and China, have shown to be quite advanced, I would say at least in phase two of the scale mentioned by Boston Dynamics.


References 1


Follow my publications with the latest in artificial intelligence, robotics and technology.
If you like to read about science, health and how to improve your life with science, I invite you to go to the previous publications.