"Bots" that aren't bots

Tags: #<Tag:0x00007f736bf96600> #<Tag:0x00007f736bf96470> #<Tag:0x00007f736bf962b8>

Not sure if this is the right category for this, but…

For those of you who are in the Bay/Berkeley, I wanted to share this bit of news about your local delivery robot startup. Maybe you guys know this already…

Why I’m sharing this: the bots aren’t bots at all (by which I mean not “autonomous” or independent in any way) but are actually basically drones, being piloted remotely by users in Colombia.

The Kiwibots do not figure out their own routes. Instead, people in Colombia, the home country of Chavez and his two co-founders, plot “waypoints” for the bots to follow, sending them instructions every five to 10 seconds on where to go.

As with other offshoring arrangements, the labor savings are huge. The Colombia workers, who can each handle up to three robots, make less than $2 an hour, which is above the local minimum wage.

This is a hilarious and evil thing to do: outsource delivery work, already a rough job for low pay, to drones piloted by even lower-paid foreign workers in places possibly with dreadful labor laws (I don’t know anything about labor laws at the current time in Colombia).

This reminds me of this story I read recently about a 1959 cultural exchange between the US and the Soviet Union.

The US used the trip as a chance for spying (about half of the “student ambassadors” were spies working for either the CIA or RAND Corporation) and as an opportunity for capitalist propaganda demos, putting on a display of “the kitchen of the future.”

It was then recycled as futurism for Look magazine in 1959:

She presses a button, and the dishwasher “walks” an electronic track to the dining table. Kitchen also has a robot floor cleaner and automatically adjustable sinks. By pressing buttons, Anne can prepare a complete meal without leaving the kitchen’s control panel.

But the robot floor cleaner, the proto-Roomba, was just an elaborate fake to impress/scare the Communists:

“They had a two-way mirror with a person sitting behind it that could see the room,” former designer Joe Maxwell told me over the phone. “And they radio-controlled the vacuum cleaner and the dishwasher.”

eta: the source and context for the Kiwibot drones was this Twitter thread, arguing that self-driving cars are probably also a grift.


Omg, fuck those things! One of them almost ran over @susan on Shattuck.

Frankly I’m surprised no one has made banners to attach to the machines explaining how they are operated…


I’m surprised no one has made banners to attach to the machines explaining how they are operated…

OMG. someone really needs to do this. Perfect guerilla jamming campaign. Tell all the Berkeley-ites you know, maybe someone will implement it…

Even stickers would work.


In this vein:

1 Like

So a question for myself as a non-expert who has written about this stuff before (I did some stories about machine learning in medical tech, and wrote about self-driving cars in relation to med tech)…

How do we assess at a broad level the utility and capability of machine learning systems to take over tasks if there’s this scammy stuff going on constantly (real people behind the curtain, or sitting under the Mechanical Turk if you will)?

If anyone has any suggestions for smart and informed critics who write regularly about this stuff from a skeptical perspective, please share! I know a couple but I’d like suggestions, especially from people who actually work on/around/near tech like this…


There was an online journal I used to follow that was pretty AI criticial for a while, but I stopped following it. They started getting lots of submissions from libertarian singularity promoting types and the signal to noise ratio got so bad I stopped following it. It’s name escapes me at the momment.

I would be curious of other resources like that though.


I KNEW IT! i first saw one with my sister and i kept yelling that it must be remote-controlled by an intern somewhere in berkeley watching us. lol.

i’ve learned to call this wizarding, as in the wizard of oz.

  • i remember years ago one of the most talked-about examples of this was a food startup that claimed to use ML and computer vision to tell you the caloric count of your meal if you just sent them a picture. actually it was humans with some “training” or something looking at your picture with their human eyes and looking up the ingredients on the internet and basically guessing the calories. hahahahaaha
  • when working for a consulting company, i totally was part of building a different startup’s wizard program. the wizards were college interns. bright-eyed.

i want it to be a bridal receipt veil. like this.

1 Like

oh my god. you remember yarnbombing? why aren’t the robots all dressed in cute sweaters by now? oh my god lol

1 Like

I wonder in some ways what the long term impact (goal ?) of this will be if left unchecked. Considering that along many facets (delivery, call centers, etc) these technologies are being sold replacements for human beings. Is the whole thing just going to abstract away workers even further, making them invisible to the public? And if so thats going to make transparency about their workign conditions and environs a bit difficult right?

Im reminded by all of those studies that demonstrate how it is easier for soldiers to dehumanize an enemy the further they are physically from them, and the greater the level of abstraction is. How it’s hard for them to commit violence up close and personal, but psychologically easier if its at a distance through a drone camera.

1 Like

yeah, and the fucked up thing is if you get bright-eyed college interns you sell them on the idea that they’re generating data and methodology to replace themselves because they know they’ll ascend in society.

i imagine the people being paid less than $2/hr are not fed a heavy diet of brainwash about how wonderful it is that they are generating the extremely valuable training data for all the machine learning models that will make it possible.

i just watched the fyre festival documentary last night. wizarding is the same as billy saying “let’s put up a cabana for fifty thousand dollars and see if someone buys it” … wizarding is “before we prove it is possible, we need to show that someone in the world will pay for this. if we can do that, then we can get vc money and then research the thing”

1 Like

yeah, and especially if it’s someone you already don’t think of as a person. =((((((

My opinion is that this is 100 percent the longterm goal/result…In an immediate sense it functions through the incentive to replace relatively higher labor costs (not much for delivery bikers!) with lower automated costs (or in this case just an internet connection to foreign workers).

I mean, you can see this happen in McDonald’s branches with automated tablet ordering, right? You don’t have to interact with the humans, they can just stay in the back and make your fast food and you can ignore them.

In the longer term, getting tech good enough to replace “unskilled labor” is the goal. I think (not a theory expert) this fits really well with a Marxist prediction about capitalists, that they would automate/mechanize labor to an extent that they don’t have to pay labor, which causes a deadly contradiction since not paying labor means you’re not paying the people who you expect to buy your products. Like, imagine if several industries were erased like truck driving would be if automated cars (long haul trucks) get good enough. That’s a lot of unemployed people.

I tell everybody about this I think because it’s one of my favorite works but the 1920 Czech play, R.U.R. (Rossum’s Universal Robots) that invented the word “robot” is all about this. I don’t want to lay out exactly why because if you haven’t read it I really want you to, it’s still funny and interesting 100 years later. There’s a PDF here.. I need to write about this somewhere!


Another brief thought: the inability thus far to build robots to totally replace “unskilled labor” really gives the lie to the idea of “unskilled labor”…


Re: the Marxist prediction, it’s called the tendency of the rate of profit to fall. There is a Wikipedia article about it, but I just went and looked at it and it was rewritten in 2018 by a power-crazed reactionary to be gigantic and kind of horrible. If you are interested and have the time you can go and look at the talk page and enjoy the argument (warning for a transphobic comment).

Anyway here is a basic explanation that is better than the Wikipedia page I think.

I think my interpretation is actually inaccurate to Marx/talking about a secondary effect, since Marx said that the tendency of the rate of profit to fall was the “ratio of surplus value to investment falling across the whole system,” i.e. automation/mechanization removes human labor, less human labor produces less value which means less surplus value to appropriate which leads to falling profits. (I think I have this right.)

It’s weird to me that for a long time I thought that focusing on human labor as the source of value made this prediction inaccurate, but that now in the present day I can see this mechanism actually happening in places like Amazon warehouses…


please do!!!

1 Like