Robots Made of Meat

This story was first posted on LinkedIn. You can find the original here.

A few months ago, I was trying to persuade my then-roommate that self-driving cars were going to be a big deal. I thought I made a pretty strong case: on-demand vehicles, more efficient interstate shipping, the disruption of two million jobs, all that. I had that argument locked down, when he said something I wasn't expecting.

"Well, sure," he said, "that all makes sense, but it's not a fundamental change, is it? From your point of view, Uber cars are already fully automated, because you call them entirely with an app and don't need to speak a word to make it work. The on-board navigation system is just powered by bread instead of electricity." With apologies to Jai for my probably significant paraphrasing.

We went on to talk about unit economics and Uber's customer experience, and finished the original discussion, but his idea stuck with me. Are Uber drivers and other on-demand providers just laborers of a different type, not fundamentally different than contract workers of ages past? Or, has Uber successfully distilled their drivers down to rentable single function robots, and in doing so fundamentally changed the role they serve?

The need for further experimentation was clearly indicated. And from this was born the Meat Robot Challenge.

"The on-demand economy is attracting more than 22.4 million consumers annually and $57.6 billion in spending." -- Harvard Business Review

The rules of the challenge were simple:

  1. Go one week without doing any chores that can be done by an on-demand app. No driving, no grocery shopping, no cleaning, no carrying bags, nothing. Can't so much as rinse a dish.
  2. Get all of these chores done using on-demand labor apps, and only the app. No face-to-face communication with the person doing the work whatsoever. Pretend they are a machine.
  3. All of these chores must be done as they are normally done, to just as high or better a level of quality. No letting things pile up.

When I launched into this experiment, I fully anticipated failure. I was sure there would be some task that the apps couldn't do, or that required face-to-face clarification about what I wanted. Things didn't go as expected.

For the first few days, the challenge had some practical difficulties. Most Uber drivers insist on you confirming your name and destination when you get in, while services like Instacart frequently require you to sign for your deliveries. Once I understood the problem though, workarounds were reasonably easy to develop.

For Uber, that workaround took the form of a bright text display on my phone: "TRISTAN MORRIS drive to JFK." Holding that up, saying nothing, and making a slashing motion across my throat almost always got the point across, and there are at least a few Uber drivers in NY who now believe I'm mute. For Instacart, it took the form of a slot under the door that "automatically" signed delivery forms placed into it. I don't think I fooled anyone, and there's a few Instacart delivery guys who think I'm agoraphobic, but they played along with the fiction and delivered groceries without ever seeing me. And after that, things worked fine.

That's it. It worked fine. Groceries showed up, dishes got cleaned, rooms got neatened. I got a new IKEA desk delivered and someone assembled it. A personal assistant application (Fancy Hands) booked my weekend trip for me, and the itinerary was good. I also got them to manage some of the other apps for me, which was convenient for having one button that does everything.

I did spend a good deal of time and effort setting up a webcam to monitor people when they were in the apartment without me, but I might as well not have bothered. Nobody ever took anything or left the door unlocked. They did a really good job.

After the initial hiccups were over, I managed to get through eight straight days of the challenge before I stumbled. Security in my apartment block didn't like all the contractors who were going up with no resident to escort them, and changed their policy to say I had to call down before they could go up. I could perhaps have argued, but the point of the experiment was made.

For over a week, I had all my chores done in such a manner that if the person working had been abruptly replaced by a robot, I probably wouldn't have noticed. As a person, I know that the human beings doing that work are complex individuals with their own needs and work environments (and yes, I tipped), but as a consumer, they might as well have been Rosie the Robot from the Jetsons future.

"Uber has put its first self-driving cars on the road." -- Financial Times, May 2016

To be clear, I'm not coming out against the on-demand economy. It's more efficient, and from a consumer perspective, there are times it's actually better. Sure, sometimes I want to chat with the guy who actually does the work, but there are also times I just want to come home to vacuumed carpets and actual groceries and not have it be more complicated than that. But I feel like, as a society, our notion of what being a worker means is at-odds with the realities the on-demand economy creates.

We have notions of individual accomplishment, of laborers who stand out by the superior quality of their work and eventually better themselves. But robots don't get promoted—even the best made machines are just used until they break or get replaced. And while we may argue about the details of when that will be technologically feasible, the financial incentive to move that way is clear.

The on-demand economy doesn't create jobs like we imagine them. They aren't vehicles for advancement, or a way to learn new skills. They are functions. They're functions that generate money, the same way a robot's functions generate welds and shape plastic, but we don't think the robot is going places. And we don't think that because we know that no matter how good the robot is, there's always going to come a Mk II model, and then the old version is going into the trash.

It may seem like a little thing, but the vocabulary matters. When an engineering firm or a tech company creates a thousand jobs with some expectation of longevity and support, that does good for society. But when a delivery company creates a need for a thousand on-demand laborers? They haven't made any jobs at all.

They've just assigned those functions to some robots made of meat, until the ones made of steel get a little bit cheaper.

Tristan Morris