Roboticists were awed by how quickly Tesla’s engineers put Optimus together but claimed the device wasn’t even close to a breakthrough.
The much-anticipated Optimus robot prototype has finally been unveiled by Elon Musk. The bipedal robot, which the Tesla CEO envisions being sold as a “general purpose” bot that is less expensive than a car and equally capable of working in factories and performing household tasks, is a machine that can walk on two legs.
The Optimus presentation at Tesla’s AI Day was loaded with the customary hype and deception from Elon Musk. It did, however, give roboticists a commendably thorough view of some of the machine’s genuine capabilities. Their assessment? Although the experts at Tesla have made amazing progress in such a short period of time, Optimus is still a research robot that will take several years to develop into anything useful.
Let’s review what we saw on the day, though, before we get into the experts’ perspectives.
- We saw two robots. The first, known as Bumble C, was a “rough development robot” that could be seen moving boxes and picking up watering cans as it crossed the stage during prerecorded demonstrations. Musk referred to the second robot as having an external covering and being “pretty close to what will go into production.” But it was only ever shown waving its hands and was only ever used as a prop, wheeled around by Musk.
- Musk reiterated his goal to develop “a usable humanoid robot as rapidly as possible,” one that can be “produced in extremely high number, eventually millions of units,” and that will “cost much less than a vehicle — far less than $20,000.” Later on, he asserted that the creation of such a device would enable “a future without poverty” and “a major revolution of civilization as we know it.”
- For Optimus, we now have some basic specifications, however it’s not obvious if they’ll hold true in the future updates like they did with the concept bot. It is 73 kg in weight, with 28 degrees of articulation in its limbs, and is powered by a 2.3 kWh battery, which according to Tesla, can last all day at work. The mechanical aspect that stands out is hand-modeled after a human’s, and Tesla demonstrated how it has adapted the navigation and control software used to power self-driving cars.
- Musk reiterated the rumor that the bot could one day “be like a friend” in a Q&A session following the presentation. He also stated that customers will be able to order one “within three years, probably not more than five years” (though it’s unclear if he meant businesses or individuals), and that it will be available for purchase.
So, that’s the bare facts about what was said and shown on stage. But what did the experts have to say? Here’s some of their reactions, from interviews with The Verge and judgments shared on social media.
“Am I blown away? No. Am I laughing? No.”
Florida State University robotics professor Christian Hubicki shared his opinions in a Twitter conversation, saying that while the Tesla team had “come a great way in approximately a year,” the actual skills of Optimus were “normal (but not mind-blowing) for humanoids.”
Hubicki observes in his thread that the Optimus robot appears to walk using a robotics technique known as zero moment point, or ZMP (though this is not confirmed by Tesla). This form of locomotion has been used for years by well-known robots like Honda’s Asimo. (When you watch Optimus, you can definitely see the similarity in the way it carefully shifts its weight from one foot to another and has its knees bent in a semi-crouch.) According to Hubicki, this walking style is “very safe, but not mind-blowing in 2022.”
He adds that the presentation’s dependability is still another important topic that is left unanswered. “How frequently does it fall? You can’t tell from a cool video, or even from a live demo,” the speaker asserts. Reliability is obviously a major consideration when it comes to using robots in manufacturing facilities, since any downtime on an assembly line may have a big ripple impact.
“An extraordinarily brave live demonstration of a herculean effort that sadly lacks novelty and imagination”
The first two Optimus bot prototypes, according to Will Jackson, CEO of Engineered Arts, were “decidedly deficient” and had “any innovation” in terms of design.
They share a lot of similarities with Honda’s Asimo robots, whose research has since been discontinued, according to Jackson. “The hands are fairly rudimentary, and the overall design is heavily built, clunky, and power wasteful. The only saving grace is a clutch mechanism in the finger activation. Compare the gear revealed this year to the reveal from last year, which featured a man in a robot suit, to see how far they have come from human level movements and skills.
According to Will Jackson, CEO of Engineered Arts, the first two Optimus bot prototypes were “decidedly poor” and lacked “any novelty” in terms of design.
According to Jackson, they have a lot in common with Honda’s Asimo robots, whose research has since been put on hold. “The design is bulky, clumsy, and power inefficient; the hands are very primitive. Only a clutch mechanism in the finger activation offers hope. To demonstrate how far they have advanced from human level movements and talents, contrast the gear unveiled this year with the revelation from the previous year, which featured a man in a robot outfit.