Don’t Forget to Tip Your Robot Bartender: Rodney Brooks on Humanoids

A few minutes into our interview, Rodney Brooks offers a cautionary tale. What, precisely, the legendary roboticist is cautioning against, however, is up for interpretation. It could be humanoid robots, generally, or specifically bipedalism. Maybe it’s about not mixing different kinds of hard liquor — or, perhaps, he’s preemptively warning against inviting the Robust.AI CTO to your next black-tie event.

“I was 25 feet away from a well-known humanoid two weeks ago,” he chuckles. “It was at a cocktail party. No one was interested in it. I said to someone, ‘hey, why don’t we go try robot tipping?’ She said, ‘yeah.’ We weren’t really going to do it, but we turned to look at the robot, and at that moment, it fell flat on its face.”

(Robot tipping, for those not from the American Midwest, is not a reference to slipping the mechatronic bartender a couple of bucks for making your double a triple.)

It wouldn’t be proper to classify Brooks as an outright humanoid hater. He was, after all, the driving force behind Rethink, the company that gave the world the humanlike Baxter robot back in 2011. For reasons Brooks is quite candid about during our conversation, Baxter would ultimately have a much greater impact in the research world than the industrial settings for which it was originally designed, leading the company to cease operations in 2018.

Rethink has been rethought — and re-ceased — multiple times since then, but that’s something we’ll have to retread in some future newsletter.

Brooks is, however, more than happy to take the proverbial pokers to an overinflated hype cycle. He has, after all, lived through more than most.

His annual January 1 Predictions Scorecard is a recommended New Year’s hangover cure, when coupled with a lot of water and a greasy breakfast. Humanoids were, naturally, a key topic in this year’s essay, and Brooks kicks off that conversation by recounting another story involving a robot bartender. The iRobot cofounder does seem to have a lot of these, but as someone who has been operating in this world for a bit, I can confirm that robo bartenders at industry mixers is a very real phenomenon.

This story gets to another deeply felt concern about automation: job loss.

Quoting Brooks,

In March 2023, I was at a cocktail party and there was a humanoid robot behind the bar making jokes with people and shakily (in a bad way) mixing drinks. A waiter was standing about 20 feet away silently staring at the robot with mouth hanging open. I went over and told her it was tele-operated. “Thank God,” she said. (And I didn’t need to explain what “tele-operated” meant). Humanoids are not going to be taking away jobs anytime soon (and by that, I mean not for decades).

There are myriad reasons why the timeline may well be significantly longer than some in the industry have led us to believe — which is to say we’re essentially being sold self-driving cars all over again. (Is it still Achilles paradox if the finish line perpetually moving?) Safety is a really big, obvious one. Training, too. The problem of contact manipulation keeps surfacing in conversations I’ve been having lately around embodied AI (see the Dyna conversation in this very newsletter). Also, all of these things need to be accomplished at massive scales.

That last bit is where Rodney’s grizzled veteran status really shines. Robust’s Carter robot is, notably, even less humanoid than Baxter. It is, as its name implies, a robotic cart. That’s selling it short, of course. It’s got autonomous navigation, along with the ability to be passively moved around by its human coworkers. Point being, it looks less like someone who works at Home Depot than something that someone would push around. That someone very much remains in the loop in a world where Carter is cruising around.

Much of Brooks’ humanoid skepticism is born of his experience successfully — and unsuccessfully — launching product. As a cofounder of iRobot, Brooks was directly involved in the most successful home robot launch of all time. But the path to the Roomba is littered with robot baby dolls, lunar landers, and other projects that never panned out. More than 99% of these things never do.

Hardware is, in a word, hard.

“Very few [executives at humanoid companies] have ever deployed robots,” Brooks tells me. “I know how hard it is to deploy robots or how hard it is to make something that a customer's going to pay for. It’s gotta be very, very reliable, and it's gotta work, with a bunch of nines — 99.999% of the time. If you get higher failure rates than that, it's really frustrating and perhaps dangerous.”

Showing that your company is quick to embrace new technologies can win favor with shareholders, but that will only take you so far. At a certain point, ROI is the only spreadsheet column anyone really cares about.

“In general, customers don't care about the particular technology they’re using,” Brooks adds. “They want cost effective results. They want it to somehow make them more efficient than they were without the technology, and that's what matters to them.”