Two weeks ago, I watched something that blew my mind. Figuratively speaking of course.
No, it wasn’t a Michael Bay movie or a particularly captivating TikTok video, but a company’s annual presentation. Ordinarily an ordinary affair, but then this wasn’t just an ordinary company. It was the electric carmaker Tesla’s second-ever annual AI Day where they unveil all the cutting-edge AI (artificial intelligence)-related technologies that they’re developing.
Amidst a sea of impressive technologies that they demonstrated, one stuck out the most – for obvious reasons. They unveiled Optimus, a general-purpose humanoid robot.
For context, a humanoid robot is an anthropomorphic machine that mimics the form and movement of a human body. It is often the first approximation of what a robot looks like when the layperson imagines it, even though the term “robot” itself doesn’t necessarily mean a humanoid robot.
For instance, when people ask me what I do for a living and I say I design robots, they immediately assume that I build humanoid robots; and I often have to explain that I build robots that don’t resemble humans but have embedded intelligence that qualifies them to be called robots.
Coming back to Tesla’s AI Day 2022, the still under-development Optimus humanoid robot, albeit minimally functional, can already perform tasks such as walking around autonomously and watering plants, and has machine vision capabilities that allow it to sense what is in its immediate environment and navigate around it.
Sure, it’s far from complete and it’ll probably be at least five years before it even becomes commercially viable. Still, the fact that the uber-talented Tesla engineering team has gotten this far in less than a year is laudable.
Those who point to Honda’s Asimo humanoid robot, which was developed decades ago, and say Tesla is merely replicating similar technology or that Optimus is far inferior to the humanoid robots Boston Dynamics develops are missing the point.
Both Honda’s Asimo and Boston Dynamics’ robots are largely expensive research projects rather than affordable, mass production-worthy robots. Also, they lack the wealth of data and artificial intelligence that powers Tesla’s partially self-driving cars – something that is integral to developing autonomous robots.
Its potential is immense and far-reaching and I would go so far as to say that in the true fullness of time, we’ll realise that its impact is of the magnitude of the personal computer and the smartphone.
Hyperbole you say? Hardly.
Imagine a humanoid robot that can accomplish any physical task that you can perform, and more. Now imagine a humanoid robot that can learn how to do things you have no idea how to do, and do them at an expert level within days.
For instance, you could purchase a humanoid robot instead of getting a house helper. The robot could do your laundry, fold your clothes, cook your meals, pick up groceries, take care of your kids and elders, and much, much more.
As valuable as that is, Optimus will be able to address an even more acute problem plaguing the world today – labour shortage. Many countries’ economies, Malaysia included, are hurting due to the severe labour shortage afflicting them, which was exacerbated by the ill-advised Covid lockdowns.
Factories don’t have enough workers and construction sites lack the manpower to complete projects on time. Humanoid robots could fix this problem in a fell swoop. They can replace human factory and construction workers.
No, it won’t be fixed overnight as these humanoid robots have to be trained to be able to operate effectively in their respective environments – just as a human needs to be trained. But they can be trained, and once a single humanoid robot is trained, this learning can be passed on to every humanoid robot in the factory/construction site.
This is because, unlike humans, who have to learn individually, only one humanoid robot needs to be taught and this learning will be passed on to its non-biological compatriots instantaneously and wirelessly via code.
As for its price point, Elon Musk, the CEO of Tesla, contends it’ll cost less than $US20,000 per robot (RM 94,500) once it goes into initial mass production. That’s an extremely attractive price point as many industries pay much more to a single worker over one or two years, especially if the cost of bringing the foreign worker in from another country and providing room and board is included.
And of course, over time, and as economies of scale compound, this cost will only go down, making it ever more economically viable.
Will it happen tomorrow? No. In two years? No. In five years? Probably not. But in 10 years? Yes, most likely.
The writer can be contacted at [email protected].
The views expressed are those of the writer and do not necessarily reflect those of FMT.