Last week Elon Musk introduced the Tesla Bot, a 1.7m, 56kg robot straight out of a sci-fi movie (yes, they actually had an awkward dancer in a spandex suit on stage).
I want to tell you why it’s BS propaganda, and what we should pay attention to instead.
Let’s start with the BS part.
Elon claimed that the Tesla bot will take care of “dangerous, repetitive, boring tasks”. An example he gave is going to buy groceries (not really dangerous, maybe boring). The timeline to have the Tesla bot roaming around the street is “sometime in 2022”.
Elon envisions Tesla-branded humanoids walking in supermarkets within a year. He also reassures us that the Tesla bot is designed to be friendly, and can be outrun or overpowered by most people.
Let me argue why you shouldn’t be worried about running from a Tesla bot, and it’s already a miracle if one of these walks out of someone’s home.
Technically, the most difficult task that the Tesla bot has to deal with is navigation. This includes crossing the street, walking on the sidewalk, and shopping in a supermarket. Obviously, the Tesla bot has to do everything without bumping into humans or things.
Can Tesla pull of such a technology? We can take a guess by looking at their self-driving cars as a proxy. Tesla’s cars have the tendency of crashing into parked emergency vehicles, and sometimes mistake the moon for a yellow traffic light. The same faulty technology won’t be as dangerous on a humanoid (it’s much slower and doesn’t carry people), but how far from home do you think it can go?
I don’t believe that Tesla will make any exponential progress with their AI within the next year. But let’s assume that they did. Try imagining for a second what people would do if they met a Tesla bot on the street. They’d group around it to take selfies at best, beat it with a baseball bat for fun at worst.
A humanoid-shaped robot is a dumb idea on so many levels. If Elon cared about building a product to actually go pick up groceries, he’d copy Domino’s pizza, not Will Smith’s “I, Robot”. Even the most advanced robotics company in the world (not Tesla) doesn’t put just two legs on its robot and doesn’t let them walk around the street. I’m talking about Boston Dynamics obviously, whose robot “Spot” has four legs and is sold for industrial applications.
I think that Elon is smart enough to know all these things. The point is that he doesn’t care about building a robot like this in 2022. Why did he announce that then, other than “for the lolz”? I’m not sure about it, but there’s a sentence that can give us some hints, and is the only thing we should pay attention to.
Elon justified the Tesla bot by saying that “Tesla is arguably the world’s biggest robotics company because our cars are like semi-sentient robots on wheels”.
Think about this. Tesla is not positioning itself as a car company anymore. It’s a robotics company.
It’s pretty easy to see why Tesla wants to take some distance from the car industry. My father has been driving BMWs for 20 years, so I grew up used to sitting in a German car. A Tesla Model S is twice as expensive as most BMWs my father has owned, but you wouldn’t tell sitting in one.
Let’s just say it: Tesla’s quality is shit. Poor leather, poor plastics, misalignment issues. The thing is, you don’t care about any of these issues if it’s the only legit electric car on the market and accelerates like a roller coaster. For the past 10 years if you wanted an electric car you had a single name to look at: Tesla.
Today, that first-mover advantage is starting to fade. Mercedes will go full-electric by 2030, Volkswagen has an electric car for €25k, and Hyundai even made one that can outperform Ferraris and Lamborghinis.
Elon has two options: fight the “car pros”, or position Tesla as “something else” and elude the fight altogether. It looks like he went for the second option.
It took more than 10 years for “old” traditional car companies to catch up with Tesla’s electric car features, and that’s the easy part. If you move the playing field to robotics, the Mercedes of the world will have to get good at software. Writing software for someone who’s used to wielding metal is like teaching to play chess to a boxer. They’re just two different games.
Tesla is competing more with Google’s Waymo and the Apple car (yes, Apple is likely building a car). Who is going to win this fight? I can’t say yet, but be ready for war.