05-10-2022 | By Robin Mitchell
At the end of this month, Tesla will be hosting an AI event where their latest robot, Optimus, is expected to be unveiled, but it is likely to be a disappointment for those who understand robotics and the engineering behind it. What benefits does a general robot present, what should we expect at the event, and why is it unlikely that anything groundbreaking will be announced?
What benefits does a general robot present?
The field of robotics dates back many decades, while the concept of robots goes even further (the Victorians had ideas of steam-powered entities handling mundane tasks). And yet, in all this time, general-purpose robots still don’t exist. Instead, current robotic platforms are specialised in a specific task, and while some hardware platforms may support multiple functions, no robot on earth can be taught a new task in the same way humans can.
While many would think that the lack of general-purpose robots results from hardware challenges, the truth is that the difficulties lie in software. Companies such as Boston Dynamics have already developed bipedal robots that can navigate around rooms, run, jump, and even do backflips with precision and accuracy, but all of these actions come from predefined functions and carefully written programs. The robot can’t be asked to do generic tasks that most humans would know how to do, and exploration is something that AI simply hasn’t developed yet (e.g. go find the box of spanners in the workshop).
This general-purpose capability comes from intelligence, and this intelligence decides how best to use the robotic platform that it finds itself in. For example, an intelligent AI housed in a wheeled robot with a single arm would understand its capabilities and then try to execute an order to the best of its ability.
If a general-purpose robot could be built, then it would mark a new era for humanity. A general-purpose robot in a bipedal form would no longer be limited by the platform’s intelligence but by the robot’s physical abilities, which are easily solved with new actuators and sensor technologies. Such a platform would be able to understand and perform commands given to humans, operate in harsh environments, and even remove the vast majority of humans from laborious jobs.
Tesla to unveil their Optimus robot by October 2022
At the end of this month (September 2022), Tesla is expected to unveil the developments in their Optimus robotic platform, which has been touted as the robot of the future. The bipedal robot has been designed to be physically slow, use electric motors for controlling limbs, and built into a sleek body to make the robotic platform more appealing to humans. Furthermore, Tesla believes that if the platform can be successfully developed, it would generate far more revenue than their EVs.
When Elon Musk announced the development of the Optimus robot last year, the world was in shock and awe, thinking that it wouldn’t be long before we all have a personal robot assistant. Furthermore, some worried about robot uprisings, so Elon Musk put the public mind at ease by insisting that the robots would be slow (so that people could outrun them). Finally, a person in a spandex suit danced around pretending to be a robot which had the audience laughing.
For those who are fans of Elon Musk, it’s time to swallow an uncomfortable red pill on the situation; they are unlikely to demonstrate anything of importance at the event. It is highly likely that they will establish a bipedal robot with fully operational legs, arms, and hands, and this robot may even walk about in the room. But while this may seem like progress, it has already been done before by hundreds of companies all over the world.
In fact, the Optimus robot may even be programmed to look like it is doing something worthwhile during the show, but it is most likely that these actions will be preprogrammed routines, just like how Boston Dynamics does the same with their robots during demonstrations. That isn’t to say that these robots are useless; it’s that they don’t have the level of intelligence that they mimic.
Why is it that Optimus will unlikely do anything special?
The reason why we believe that Optimus will do nothing of interest comes down to a simple fact; general AIs simply don’t exist yet, and there are so many companies working on the problem. While Tesla self-driving vehicles (which themselves are extremely controversial) are themselves impressive feats of AI, navigating along a road is nothing compared to performing general tasks requested by humans in environments exposed to constant change.
Companies such as Boston Dynamics are arguably THE industry experts in robotic design and fully understand the challenges of movement, balancing, and actions. The chances of Tesla magically solving the problem in a year while Boston Dynamics have slaved over the problem for decades is simply unlikely. The same applies to AI development; since other companies have also been working on the AI problem, it is doubtful that Tesla has magically come across the solution. As previously said, navigating on the road is different from being able to dynamically navigate a robot while performing generic tasks fluidly.
Overall, I am not holding my breath on the Optimus reveal, and those who are fans of Elon Musk should keep in mind all the false promises that have been made; Cybertruck, electric lorries, solar roofs, hyperloop, underground tunnels, and robot taxis.