Agility Robotics’ Post

I remember having a boss who would have been happy had us as worker's could simply replace the hands we were given with multi use adaptable hands, the thing is if you're not doing one specific thing and are required to perform a wide variety of tasks, its still easier to pick up a tool designed for the specific purpose. In single use applications I can see the benefit of the hands being the tools, but again it will be for specific unchanging tasks, especially if we're trying to pitch robots to work in a multi faceted team environment that requires cooperation in carrying out the parts of a given operation. Peace

Hands are about humans with external universality. EOATs are about machines with internal universality. We have to stick to the modular morphology of machines, if we want to accelerate things. There’s an analogy to bipedal locomotion.

Basic specific terminal device for basic repeated task (AKA today) Everage complex 3 fingers robotic hand for Everage repetitive task on 3 to 5 different grip patterns Wide complex 5 fingers robotic hand for versatil grasping and social communication... Still 5 fingers needs robustness and easy control for the 22 dof, there are mechanical solutions that have never been developed for these past 60 years....

Not having 5-finger hands but dedicated end-of-arm tools is as good as certain in industrial applications, even in the 'humanoid' form factor. The only place where I see 5 fingers justified is in the home environment, where everything is made to work with the human hand in mind. Figure is building a Home robot, not a Factory robot.

Very insightful and valuable point. An alternative perspective is that developing solutions for intelligently manipulating a large number of degrees of freedom within a limited space can benefit other robotic tasks as well. Achieving a human-like, high-performing robotic hand represents a significant milestone in robotics.

Melonee Wise make an excellent point! Let’s not reinvent decades of robotic end-effectors.

Well said Melonee Wise. Robot end effector designs that will scale will be largely dependent on and optimized for the range of tasks for which they are built.

For sure a thoughtful and interesting take.

Insightful! This is really good questions but it seems like there is not just one answer to that. Robots might need hands for some operations and for some not, it depends on the process they are involved in. Good discussion. Thanks for this. 🤖