Tesla's Robot has a Secret Sauce

Tesla’s Optimus robot has collected lots of bouquets and rotten tomatoes since the first two prototypes hit the stage during Tesla’s second AI day. Someone’s dead wrong about the shuffling, slumping metal lumps, but whom?

"This Tesla AI day is next level cringeworthy. Complete and utter scam." Reacted researcher Filip Piekniewski. However in contrast, Lex Fredman, of MIT and podcast fame, said: "Tesla Bot is truly amazing. It's the early steps of a revolution…”

As said, they can’t both be right. In defense of Piekniewski, Optimus doesn’t represent a marked advance in what robots can actually do. Head on over to the Boston Dynamics website and watch Atlas in motion if you have any doubts. Optimus hasn’t touched the leading research edge of robotic capability. To think so is laughable.

But.

Babbage’s wooden and metal cog computer was highly capable; just not practical. Metal cog iphones were never going to be a thing.

The wonderful Atlas is all about hydraulics. That’s what moves it’s limbs even though it’s powered by electricity. Messy, finicky, expensive hydraulics… but precision hydraulics aren’t very likely to be a thing either, for robots. Not if there’s any other way to do things.

And Tesla just showed there is. Let me translate Lex Fredman’s enthusiasm: “Actuators? You can do everything you need to with just actuators?? Holy… hold on to your pants, kids, we may just have passed a hardware inflection point, here.” The clincher to this view of the argument is the price Tesla is hoping to hit, just one quarter the price of a real robot now; though they believe their robot will be highly capable, with superior AI.

“Actuators” are small powerful electric motors, and the term goes back a long way. The first radio-controlled aircraft used tiny electric motors to control the ailerons and rudders and called ‘em actuators not motors so nobody thought they were running the propeller. That was powered by gas. Powering robot limbs has been the catch, the critical path, the largest single bottleneck, the necessary invention that still wasn’t in place. Tesla is going to do it all with small, powerful electric motors. No more bottleneck, clear sailing ahead. Bloody fantastic; and it just makes sense.

I’m sure researchers are going to keep researching artificial muscles of one kind and another. I hope they do; I hope they make breakthroughs. But those breakthroughs just became, in the strictest sense, unnecessary.

Elon likes to leverage knowledge gained in one area to make progress in another. Metallurgy is a rich source of trade secrets, and both Tesla and SpaceX have troves of metallurgical trade secrets, so Elon employed that knowledge to advance the art of tunneling into the earth, where damned hard (but not too expensive) rock cutters are a necessity.

With Optimus, Tesla is leveraging its knowledge of how to build compact, powerful, high-torque, instantly accelerating (or decelerating) electric motors (including metallurgical trade secrets regarding metals incorporating rare earths) to discard clumsier ways of powering limbs. Tesla will just get the job done with some motors. Then toss in AI that’s not so different from so-called “full self-driving,” already well underway, and you’re golden. This is great work if you can get it; if you can accomplish the feat, that is - and it certainly looks like Tesla can and will. Using small electric motors you can mass produce really is a terrific advance – for practical production. For price. It clears the way and makes making a worker dirt cheap. Maybe $10,000 for the body, and $10,000 for the computational brain, and you’ve got yourself a helper that can be rolled off assembly lines. Eat your heart out Babbage.

(No doubt the price will creep up along with some mission creep that adds considerable capability before you get to buy one; that’s technology for you. But the price will come down from there.)

I know, you were hoping for a bot that could outdance anything from Boston Dynamics and soothe a crying child; and that might happen one day. But the Roadster wasn’t the first electric car either, it was nearly a century too late for that. The Roadster was the first step towards practical, profitable electric cars (although not profitable itself.) It and Tesla’s subsequent models made the path to such cars clear as day, and many other car companies are now following, as quickly as they can.

The Optimus robot’s “awkward” steps and corny gesticulations were also a clear demonstration that a barrier has fallen, and that the androids ARE coming. Cheaply. It can be done. Without hydraulics. Expect other companies to adopt the same actuator strategy, themselves. Get ready for you robot underlords.