Agility Robotics’ Post

Lack of a clearly defined problem they are going solve beyond demos and pilots. Why do we need humanoids and not other type of mobile manipulators? One could argue that legs provide versatility, but for what specific use case is this kind of versatility desired and how valuable is it from the customers perspective given the price, performance, and reliability challenges?

After having worked with AMR's in warehouse and factory settings for around 5 years, I would say that making the systems robust enough to work mostly unattended, or at least correcting themselves without needing constant robot wrangling is a must. The entire point of the bots is to alleviate humans of the repetitive tasks, yet if the machines break constantly, the job changes from the original work to constant robot maintenance as the repetitive task. Otherwise, having the robot do a task that needs increases in efficiency past human limits is another key such as working non stop shifts with small opportunity recharging breaks.

Cost, speed and proof! Proof is a little chicken/egg but widespread trails will only come when the cost/ROI is worth it and the speed can match or exceed existing systems. I believe in a future with widespread use of humanoids, but we still have a long way to go!

"Cost" is not the main hurdle for scaling humanoid robots; the key challenge is achieving efficiency. Unlike humans, robots don't need (breaks, salaries, or specific working conditions) and can operate continuously with just power and maintenance. They can also work in harsh environments, making them ideal for Disaster response, Space exploration, and Hazardous material handling. However, this requires overcoming significant technological challenges in (AI & ML), power sources, and durability.

Scale seems to be the biggest hurdle. For all the deficiencies, humans are an incredible instrument for the hourly rate. The human eye is a 576 megapixel camera, with extremely dexterous fingers, arms, and bodies, and a controller (brain) that is HIGHLY adaptable to slight variations that is powered with VERY efficient fuel sources (chemical reactions at a mitochondrial level) that are not tethered to anything. That's a LOT to try to replicate all in a small footprint.

Reliability of the autonomy. Robust manipulation and perception. Safe perception and safe behaviors. It‘s not costs, those will fall quickly enough. This applies to all mobile autonomous robots that interact with their environments. Humanoids are just one form of these systems. The challenges are equally hard still for quadrupeds, wheeled systems etc.

Ethics. Creating enough buy-in from society that people can embrace the change as adding to overall social welfare. I saw a recent Parkour demo from Boston Dynamics, and as amazing as it was, my first reaction is that the military must drool when they see this tech. Create regulations that keep the tech from being militarized against human beings. Develop use cases that benefit the working class rather than replace their jobs. Build robots that tutors a child who is falling behind in school. Otherwise, people will fight scaling robots globally tooth and nail.

Cost, speed, battery life, safety, enough correct applications that cannot be done cheaper or faster using traditional industrial robotics. Lots of hurdles there.

Short term: Getting the cost low enough and the capability high enough, to compete with low cost labour markets. Long term: Figuring out how all the people who used to do those jobs can still have a fulfilling life and take care of their families.

There are at least 12. All of these issues, I believe, can be addressed with the general public by an executive in season 3 of Going Public, the first-of-kind interactive investing show on Marketwatch TV with Dow Jones Distribution.