Why would a car company such as Tesla start to develop general-purpose humanoid robots? That is the question on many people’s minds.
A car company may develop humanoid robots for a variety of reasons. They may want to use them as assistants in their factories or showrooms, or they may want to develop robots that can drive and park cars. Probably not all that strange as Tesla is into the idea of automated self-driving cars.
However, Musk’s plan for developing human robots is much broader than this. To begin, Tesla intends bots to take over jobs that are dangerous, monotonous, or both and replace humans in those roles. However, the long-term goal is for them to provide services to millions of households, such as caring for the elderly, mowing lawns, and preparing meals. Elon Musk even considers that people may begin gifting his company’s bot to their elderly parents at some point in the future.
Have you heard of Optimus?
Unless you have been living under a rock, you have probably heard of Optimus. On August 19, 2021, during Tesla’s Artificial Intelligence (AI) Day event, Elon Musk, the company’s CEO ao, the intention to build a general-purpose robot called Optimus was first disclosed to the public. Over a year later, Optimus, no longer a theoretical concept, walks onto the stage at Tesla’s 2022 Artificial Intelligence (AI) Day event. Now the critics’ opinions are being circulated amongst the press, most notably ABC business writers Tom Krisher and Matt O’Brien, who wrote a relatively unflattering review of the event.
Their article states that the Tesla early prototype robot walks awkwardly and waves but does not perform complex tasks.
The demonstration failed to impress AI researcher Filip Piekniewski, who called it “next-level cringeworthy” and a “complete and utter scam.” He stated that it would be “good to test falling, as this thing is going to fall a lot.”
In their article, Tom Krisher and Matt O’Brien point out that Tesla was not the first car company to dabble in creating a general-purpose robot. In fact, according to their report, Honda seems to take credit for that.
ASIMO, have you heard of him?
Honda ceased production of the Asimo robot in July 2018. Honda decided to discontinue all Asimo robot development and production to focus on more practical applications of the technology developed during Asimo’s lifespan. The name was inspired by Isaac Asimov.
Isaac Asimov was an American author primarily remembered for his work on the Foundation series and his stories involving robots. Between the years 1942 and 1949, the authors of the Foundation stories produced the works that would later be compiled into the Foundation trilogy: Foundation (1951), Foundation and Empire (1952), and Second Foundation (1953).
It was also convenient that HONDA could make ASIMO’s acronym (Advanced Step in Innovative Mobility), using a name associated with popular stories about robots. The development of ASIMO represented the pinnacle of Honda engineers’ twenty years of work in the field of humanoid robotics. ASIMO was capable of running, walking on slopes and surfaces that are uneven, turning without jerking, climbing stairs, and reaching for and grabbing objects. ASIMO was also capable of understanding and responding to straightforward voice commands. ASIMO could recognize moving objects, postures, gestures, the surrounding environment, sounds, and faces, which enabled it to interact with humans. This gave it the ability to communicate with them.
Which other car companies have dabbled in robotics?
It turns out that Honda and Tesla are not the only car companies to have entered the robotics market. As a result of its acquisition of the robotics company Boston Dynamics in 2021, Hyundai owns a collection of humanoid and animal-like robots. Ford has partnered with the Oregon-based startup Agility Robotics, which creates walking and lifting robots with two legs and two arms.
According to Ryden, the research that automobile manufacturers are conducting into humanoid robotics has the potential to result in machines that can walk, climb, and overcome obstacles; however, impressive demonstrations in the past have not resulted in an “actual use scenario” that lives up to the hype.
Can Tesla succeed where Honda failed and create robots that live up to the hype?
It has been challenging for companies to create general-purpose robots in the past successfully. Ayonga Hereid robotics engineer, thinks that bespoke single-purpose robots should be how the robotics industry develops in the future.
To succeed, businesses that develop robots for general purposes need to solve a problem known in the trade as Moravec’s paradox. At the tail end of the 1980s, a brilliant computer scientist and roboticist by the name of Hans Moravec came up with an intriguing paradox. According to Moravec’s theory, computers can perform tasks with ease that are challenging for humans (like intelligence). However, they have difficulty with perception and mobility, two things that humans find relatively easy.
Critics of Tesla’s new bot, such as robotics expert Cynthia Yeung even query the need for a robot to look human. Why does a general-purpose robot need to have five fingers? “None of this is cutting edge,” Yeung tweeted. “Tesla, hire some PhDs and attend some robotics conferences.”
Yeung also questioned Tesla’s decision to equip its robot with a human-like hand with five fingers, noting that there’s a reason why warehouse robots use pinchers with two or three fingers or vacuum-based grippers. Employees working for Tesla stated that Optimus robots would have four fingers and a thumb with a tendon-like system to mimic human dexterity.
“When it comes to developing a robot that is both affordable and useful, a humanoid shape and size isn’t always the best way,” said Tom Ryden, executive director of the nonprofit startup incubator Mass Robotics.
What did the critics think of Optimus’s appearance on AI Day?
Krisher and O’Brien described the tasks performed by the robot as basic. It had exposed wires and electronics and a later, next-generation version that had to be carried onstage by three men.
It appeared that even Musk was embarrassed. Musk suggested that the issue with flashy robot demonstrations is that the robots do not have the intelligence to navigate themselves and are “missing a brain.” However, on Friday, he did not provide much evidence to support his claim that Optimus was more intelligent than the robots developed by other businesses and researchers.
According to a statement by Musk, the early robot walked onstage for the first time untethered. According to him, Tesla’s objective is to manufacture an “extremely capable” robot in large quantities (possibly millions) at a price that is lower than the price of a car, which he estimated would be less than $20,000 per unit.
Tesla was prepared in case Optimus’s performance went wrong on the day. Tesla showed a video of a robot carrying boxes and inserting a metal bar into what appeared to be a factory machine. However, Krisher and O’Brien wrote that no live demonstration of Optimus completing these tasks was provided.
When will Optimus be available for sale?
Employees of Tesla told the Palo Alto, California crowd, as well as those watching via Livestream, that Optimus has been in the works for six to eight months. People will be able to purchase an Optimus “within three to five years,”
By then, it is hoped that the robot will be powered by massive artificial intelligence computers that analyze millions of video frames from “Full Self-Driving” vehicles. These computers will be used to teach the Optimus robots tasks.
Krisher and O’Brien conclude that experts in the robotics field are skeptical that Tesla is any closer to mass-producing legions of human-like home robots that can do the “useful things” Musk wants them to do, such as make dinner, mow the lawn, or keep an elderly grandmother company.