Do Tesla robots dream of infinite planetary resources?

Elon first revealed Tesla’s plans to create robot helpers last August, in a theatrical display using, err, humans. Given the electric car maker’s notoriety for announcing products, only to see them delayed by up to half a decade, FT Alphaville thought it would be a while until the project got off the ground.

But on Tesla’s latest earnings call, following better than expected fourth quarter results, Elon said the most important development work this year will be on the robot — christened “Optimus”. Cybertruck be damned.

JP Morgan’s analysts note Tesla’s pivot from carmaker to robot maker could have a significant impact on its wider automotive plans.

From a note out overnight:

This implies that the Roadster and Semi (originally slated for launch in 2020) and the Cybertruck (originally slated for launch in 2021) will be delayed until at least 2023, including as the firm prioritises development instead on a humanoid robot concept that will serve as an “incredible buddy like C3P0 or R2D2” with a personality that evolves to match its owner while also solving the economy’s labour shortage problem. The robot, dubbed “Optimus”, was said on the call to be the most important product development work taking place at Tesla this year, given its “potential to be more significant than the vehicle business over time” even as Rivian, Ford and General Motors deliver battery electric pickups to customers.

We don’t know if Elon Musk, Technoking of Tesla, is a fan of the show Caprica, a spin off prequel series of the re-imagined (and highly superior) Battlestar Galactica, where humans create robots which then decide to obliterate humanity.

But we presume he must at least be familiar with the genre. His numerous public declarations about how superhuman AI could be very bad for humans, certainly suggests as much.

So why then is Musk accelerating the mechanical takeover of the planet by ploughing head first into personal robot manufacturing?

We can only hope it’s because he feels it’s essential to get ahead of “bad guy” competition, i.e. those who might wish to programme robots for destructive rather than buddy purposes. Perhaps he thinks that by being first to market he can be sure robots are endowed with fundamentally benign protocols a la the Three Laws of Robotics in the Asimov novels. (Though, if you read the Asimov novels you’ll know even these laws can potentially be undermined or hacked.) Elon, presumably, doesn’t think his own motivations can be hacked or corrupted either.

Whatever the case, there’s a bigger conceptual problem with developing robotic slaves to fill a labour shortage, or avoid unionisation. And it’s not the obvious economic one everyone is worried about. In a new robot-powered economy, displaced menial labour would eventually adjust to human-touch services, such as care or therapy, or simply making TikTok videos. There would always be jobs. Just different, and potentially more frivolous ones. (Who could have conceived of a social media manager in the 1980s?)

Who will be the cheaper serf?

The bigger issue is how robots will compete over resources with their meat-bag equivalents. Who will be the cheaper serf to hire will be the ultimate consideration. And it’s not at all clear it will be the robot.

Elon’s mechanised robots will, we presume, be made of scarce rare metals and commodities. The fact they will be rolled off a production line fully primed with intelligent software may compensate for some of that cost. Even so, it will be tough to compete with the cheapness and abundance of humans.

Robots will still need to be powered by electricity, lubricated with oils, cared for and maintained. They will still need reboot, downtime or maintenance time like their human equivalents.

Humans, on the other hand, are comprised of the planet’s most abundant resources (oxygen, carbon, hydrogen and nitrogen). They are also capable of self-replication without a manufacturing plant. That means they will always be fundamentally cheaper and easier to make. Yes it may take 16 years to grow a functioning human worker, but the pipeline is solid, with stock continuously being replaced if not added to. (Though the fact Elon feels the human replacement rate is in trouble, could be another influence over his thinking.)

Decommissioned humans, meanwhile, can also be recycled far more efficiently than robots. From ashes to ashes.

Elon’s fans might counter that the whole of point of Optimus is to liberate the world’s poorest from their daily grinds.

But while that is in theory a fine goal, it’s very possible that without a significant leap in material science to overcome resource constraints, the effort might actually end up making things worse. Any shortage of multitasking humanoid robots at all, and it’s clear they’ll continue to compete with humans for service positions in, say, elite human households.

The only rational reason to go head first into a robot slave economy is if the full cost of creating and maintaining a robot slave will be cheaper and friendlier for the planet (in terms of resource consumption) than that of a human.

For a true dystopia, consider what might happen if a competitive robot industry were to develop. To keep costs down and to optimise the use of planetary resources, manufacturers would eventually turn to biology. This could see robots emulate our own human bioengineering to the point they become indistinguishable from humans. Competitive forces might then encourage the growing of robot slaves in labs. Or in controlled self-replication hubs. There these robots would be pre-loaded with software that ensures they love their owners and would never hurt, revolt or try to displace them.

The only difference between them and us would be their allegiances, their aspirations and their willing subservience to their masters. At least, for as long as their programming didn’t get corrupted or hacked.