The robotics head of Amazon, Tye Brady, has expressed a skeptical and realistic view of the trend of humanoid robots, which are booming, in particular those that appear smooth but do not necessarily translate into usefulness in the real world.
In an interview with the media at the Web Summit in Lisbon, Brady claimed that many of these human-like machines are technology for technology’s sake
On Making Form Subservient to Function
The main argument being put forward by Brady involves the transformation of robot companies to focus more on the appearance, on how futuristic and humanoid it appears, rather than getting into deep thought on what kind of role the robot is supposed to play.
When we consider robotics, we consider, A, what problem are we trying to solve? And, B, the then function. “By functional, we possess form,” he said.
He thinks this is retrogression: “It does go the wrong way in case you begin with form. He does not want to come up with robots that resemble humans; instead, he wants to see machines that are functional and viewpoint specific.
As Brady noted, Amazon already has more than one million robots that work in its warehouses as picker-sorters and wheeled load pullers and other automated machines that collaborate with people
Such robots do not need a humanoid form; however, they offer actual worth in industrial environments.
Humanoid design weaknesses
Brady is not completely opposed to the humanoid idea; though bipedal locomotion can enable robots to move about difficult areas or even up the stairs, he is of the opinion that current humanoid designs fail miserably where it is required the most.
He claimed that after a humanoid robot comes somewhere, there is the actual test: “A humanoid robot can get a humanoid robot somewhere, and that is when the difficult part starts.”
There is, probably, some task that you need to perform, and it will now include some kind of touch, some kind of handling, Brady explained.
Stated otherwise, a robot must not be merely movable and dextrous but must also have fine motor control, dexterity, and be able to engage objects in a meaningful way.
To another point made by Brady, lots of humanoid prototypes will be of interest to audiences, yet they will not be resilient enough to perform any work in the industry or the warehouse.
He said that work was required until the robots could handle the entire battery of objects that they may experience in the environment.
The Alternative to Amazon: Competent and Purpose Driven Robotics
Instead of pursuing the humanoid direction, the Amazon policy is inclined to functional automation bots processing routine, monotonous operations in an economically efficient manner.
In his book, Brady noted that robots at Amazon are created to help augment and bring utility to individuals by not fully eliminating people but rather by getting rid of repetitive tasks.
He cries foul over the idea of 100 percent automation as being unrealistic. To him, it is not to create glorified androids but rather to create intelligent machines to supplement human workers.
The existing fleet of robots deployed in the warehouse operations of Amazon consists of the machines able to pick the objects, move them, and even understand the fragile ones, however, these types of solutions are mounted, mobile, and extremely task oriented.
An example of this was the Amazon Vulcan robot, which employs a powerful sensing system to prevent potential damage to objects, but it is a floor mounted machine, not a humanoid.
Why Brady’s View Matters
- Humanoid hype reality check: With the bipedal robots constantly being developed by many different companies that have apparently made it easier to look cool, Brady views the industry and reminds them that they need real life use.
- Robot economics: The price of humanoids, when it comes to both finance and technology, is very high, and the strategy adopted by Amazon dictates that returns cannot justify this investment yet.
- Future direction: According to Brady, one day there will be robots that combine both mobility and manipulation not only to walk like humans but also to accomplish some tasks upon reaching their target.
- Human robot collaboration: His words support an idea of robots as the partners and facilitators, and not substitutes as complements to human work and as liberators of the personhood of monotonous or dangerous operations.
Amazon robotics director Tye Brady is giving a more down to earth refutation of the present-day robot buzz about humanoids.
Although it is true that fully humanoid machines could still be a source of amazement at the trade show, Brady states that it is not yet prepared to meet the harsh reality of industrial applications in the real world.
To him, the future of robotics should not be in how to create things that mimic us, but rather, in creating a machine that does what we require it to do in an efficient, safe, and cost-efficient manner.
Regardless of whether his vision prevails or the humanoid dream is still playing a vital role in the headlines, one thing seems obvious: Amazon is gambling on the functional rather than the formularies of robotic automation and that may be what happens next in the decade to come.
