On this episode of Uncanny Valley, we introduce you to DoorDash’s new delivery robot and discuss what the growing robot population means for humans.
Coco delivery robots navigate the streets of Santa Monica, CA.Photo-Illustration: WIRED Staff; Jay L. Clendenin; Getty Images
All products featured on WIRED are independently selected by our editors. However, we may receive compensation from retailers and/or from purchases of products through these links. Learn more.
Earlier this week, DoorDash unveiled its own new autonomous robot called Dot. The company says it’s part of its goal to have a “hybrid” model for deliveries. It’s the latest sign of a renewed interest in the industry of delivery robots after years of challenges. WIRED’s Aarian Marshall joins us to discuss why this matters for all of us, whether we’re ordering in or not.
Mentioned in this episode:
DoorDash’s New Delivery Robot Rolls Out Into the Big, Cruel World by Aarian Marshall and Boone Ashworth
This Food-Delivery Robot Wants to Share the Bike Lane by Alex Davies
You can follow Michael Calore on Bluesky at @snackfight, Louise Matsakis on Bluesky at @lmatsakis, and Aarian Marshall on Bluesky at @aarianmarshall. Write to us at uncannyvalley@wired.com.
How to Listen
You can always listen to this week’s podcast through the audio player on this page, but if you want to subscribe for free to get every episode, here’s how:
If you’re on an iPhone or iPad, open the app called Podcasts, or just tap this link. You can also download an app like Overcast or Pocket Casts and search for “uncanny valley.” We’re on Spotify too.
Transcript
Note: This is an automated transcript, which may contain errors.
Michael Calore: Hey, Louise, how are you doing?
Louise Matsakis: Hey, Michael, I’m good. How are you?
Michael Calore: I’m doing well. Thanks for being here this week. Filling in for Lauren.
Louise Matsakis: Yeah, Lauren is on a really exciting trip to Arizona that I’m sure we’ll hear more about soon. So, as her editor, I am happy to fill in when she’s off on an adventure.
Michael Calore: All right, well, today we’re going to be talking about something that, a little bit about robots and delivery apps. It’s also an intersection that WIRED’s Aarian Marshall has been following closely for years. So please welcome back to the show, Aarian. Hi.
Aarian Marshall: Hello.
Michael Calore: Given today’s topic, I’m curious to know what is the most unexpected thing you’ve both ordered through a delivery app?
Louise Matsakis: So I was on a press trip to China recently, and I was taking notes during an interview, and I got pen all over my white pants. So I ordered the Chinese equivalent of a Tide pen. I think that’s probably the weirdest thing I’ve ever gotten on demand.
Aarian Marshall: I am so constitutionally against just paying for delivery even before there were apps. I hated paying delivery people. Much respect what they do, but I’d rather spend my money in other ways. So I think recently I got really desperate and ordered pad Thai, and that was wild for me. So that’s where I am.
Michael Calore: That’s your most unexpected thing?
Aarian Marshall: Yeah, the fact that I used it at all is unexpected.
Michael Calore: Yeah, I mean I live in one of the most population-dense parts of the west coast of California, and I rarely use delivery apps, but I would say probably the most unexpected thing I’ve ever ordered is a bottle of Jameson Irish whiskey.
Aarian Marshall: Respect.
Michael Calore: This is WIRED’s Uncanny Valley, a show about the people, power, and influence of Silicon Valley. Today we’re talking about how after years of struggles delivery app companies are still trying to deploy their robots right to your doorstep. At an event earlier this week, DoorDash unveiled its own new autonomous robot called Dot. The company says it’s part of a goal to have a hybrid, quote, “Work model for deliveries going forward, working with humans, but also drones and autonomous vehicles in the mix.” But DoorDash is not alone in its efforts to have your dinner delivered by a robot. The industry of autonomous deliveries has been steadily evolving and has faced considerable challenges along the way. We’ll dive into why some of these companies are still betting big on delivery robots, the race to create the right technology to use in them and what having these robot fleets in our cities could mean for all of us. I’m Michael Calore, director of consumer tech and culture.
Louise Matsakis: I’m Louise Matsakis, senior business editor.
Aarian Marshall: And I’m Aarian Marshall, a staff writer covering cities and transportation.
Michael Calore: So Aarian, you attended this event this week hosted by DoorDash where the company unveiled their new robot. Its name is Dot. Tell us all about it.
Aarian Marshall: Yeah. So I attended this event virtually. My colleague Boone Ashworth actually went in person to the event in San Francisco. And we knew there was going to be some kind of robot, but I was really struck by how cute Dot is, but also how big Dot is. So most delivery robots you see out in the world today, there aren’t a ton, are pretty small. They usually come up to maybe an adult human’s knees maybe a little bit taller. This one’s almost five feet tall. It’s covered in cameras. It is nine cameras and seven other sensors on it. It is bright red, it has these big pixelated eyes and to get into the delicious DoorDash stuff inside, its mouth opens kind of like a Pac-Man. It also in the promotional videos had this kind of WALL-E like voice that talked to consumers. So altogether a very cute thing. We did come away with some questions about how this thing was going to operate in the real world. DoorDash said it’s going to be on sidewalks, it’s going to be on roads, and it’s also going to be on bike lanes. And that’s a really complicated challenge for an autonomous vehicle to have to handle all of those things at once, all of the different kinds of stuff you see in all of those different places. So it’s going to be interesting to watch going forward.
Louise Matsakis: So I’m particularly obsessed with one detail, which is that Boone gave Dot a little kick to see how the robot would react. Can you talk about that? You mentioned bike lanes, you mentioned all of these obstacles, but what about the threat of people? Is that part of the reason that DoorDash decided to give it these cute details so that maybe people would be less likely to kick it the way the Boone did?
Aarian Marshall: Definitely. There is a whole bunch of researchers who focus on how to make robots more palatable to humans. Something I ended up talking to experts about for the story and didn’t make it in were some of the technologies that it’s not necessarily that Dot does, but other robots do to kind of protect themselves. Someone told me that Starship Technologies, which is another leader in this space, has actually trained its robots to stay away from children because children like to mess with these robots and they’ve trained them to, if there’s an adult present, to go stand by the adult because adults are safer than children. So it’s something that these companies are definitely thinking about, how to make sure that their robots stay safe in this big complicated world we live in. And it turns out one of those ways is to stay away from kids.
Michael Calore: Also, they should just not deploy them in Philadelphia.
Aarian Marshall: Yes. Another good tip.
Michael Calore: So Dot is not DoorDash’s first robot, right? Earlier this year the company announced a partnership with Coco Robotics. Is that right?
Aarian Marshall: Yeah, that’s right. So Coco is a LA-based company right now they’re deployed in a handful of cities including the west side of Los Angeles and they have this kind of smaller, more traditional sidewalk robots and they are operating, you really do see them all over certain neighborhoods in LA. They just hang out on corners. They are out delivering food. So they are out and about and they are partnered with DoorDash and they have been for a few months now.
Michael Calore: But Dot is the creation of DoorDash Labs, which is a wing of the company that works on automation and robotics. So why is DoorDash sort of trying to approach this from two angles?
Aarian Marshall: Yeah, it’s a good question. The first thing that DoorDash told me is that they have data from millions and millions and millions of deliveries and they think that that data is going to help them run the best robotics service that can be run. But the other thing is they’re really in control of the whole thing and they like that idea. And I also think that part of it is that DoorDash, like so many other companies in tech and in Silicon Valley are really trying to promote how thoughtful they’re being about artificial intelligence and robotics and by running their own robotics company they can show investors and then probably also potential employees that they are really serious about the busiest things in tech right now.
Louise Matsakis: I have a question for you about these robots. When the Coco Robots first were deployed in LA, I live on the west side of LA, so I’ve seen them everywhere, actually, I looked it up and it seemed like they were still mostly being remote controlled by contractors. So instead of a delivery person, it was just somebody in an office building somewhere who was basically playing a video game. And has the technology progressed a lot since then, or how advanced is the tech really, and is that part of the reason that DoorDash wants to go in-house now, that we’ve gotten to the point where they’re much closer to being fully autonomous?
Aarian Marshall: Yeah, it’s a great question and one that is not entirely transparent. I will say that Coco in particular, I’m hopefully going to make my way to LA in the next few months to actually hang out with their teleoperations team, which is the team that theoretically helps the robots behind the scenes get around. And what Coco has told me specifically is that they do have a lot of their robots driving more autonomously than they have been in the past. There’s still people there sitting behind screens, seeing what the robot sees and intervening when the robot gets in trouble. We don’t know how many times that happens, but it does seem like the technology has gotten better than it was a few years ago. Now I think there’s a big difference between that and what you see a company like Waymo doing, those autonomous vehicles are out on the roads and they are the vast majority of the time really driving autonomously. So there are a bunch of strides happening in autonomy right now and that’s really exciting. The other thing about Waymo is that they have spent billions and billions and billions of dollars in this over the years. Is DoorDash willing to spend that amount of money to get a perfectly operating robot that is truly autonomous? I think that’s an open question. They wouldn’t tell me how much they’re actually investing in this right now, which I thought was interesting. So I do think it’s an interesting time to jump in, but if they’re going to jump in, they’re going to have to spend a lot of money to get there.
Michael Calore: It’s an interesting time for them to jump in just because the excitement around delivery robots really feels like it was at its peak right before the pandemic and in the first couple months of the COVID pandemic, so 2018, 2019, in the first part of 2020, and then it really cooled off. And I’m just wondering if there are other companies that we can sort of point to that have kept the torch burning or companies that have faced speed bumps then decided to pull over.
Aarian Marshall: So the pandemic was really a high point for enthusiasm about these robots because people were scared of contact with other people. So it was like, well, let’s let robots do the delivery and we won’t put delivery people at risk. So you saw Amazon, you saw FedEx investing a lot of money in these things. A few other companies. Postmates started a company called Serve Robotics that’s still around. It’s now a spinoff on its own. And then over the years you saw them back off a bit. Amazon abandoned that project, so did FedEx and it sort of fell out of favor. And I think that’s partly because this is such a complicated technical problem to create a robot that can operate in so many different contexts. And roads are complicated, but they have very specific rules. Sidewalks, not so much. People don’t have to put on their turn signals. People don’t have to follow specific lines or walk in a straight line. And people also, some of them are in wheelchairs, some use mobility aids. There’s just a huge variety of things you see on the sidewalks. It’s just a difficult problem. So we’ve seen less investment and less enthusiasm for sidewalk robots over the past few years, but they’re still around, and we saw Sam Altman invested a bunch of money in Coco Robotics earlier this year, so it’s happening. It’s just smaller than it was.
Louise Matsakis: Yeah, I’ve certainly seen a lot of videos of these delivery robots getting into all sorts of unexpected jams. I noticed here in LA that often the tree roots totally disturb the sidewalk, so there’s almost many mountains that the Coco Robots have to get over and I’m always afraid that they’re just going to totally face plant. So how realistic is it that we’re actually going to be able to overcome all of these obstacles?
Aarian Marshall: Yeah, I’d say it comes down to, and this is the thing I always say about autonomous vehicles, it’s not when, it’s where. So I think it’s possible to build a robot that can deal with specific contexts. For example, DoorDash says they want to operate in the suburbs, and part of the reason is because the sidewalks there can be wider, though not always. They sometimes have bike lanes, and they think this is a less complicated place. So if you keep limiting the amount of space where these have to operate, they can actually get pretty good at it. Coco, for example, has told me that in LA they know, to your example, Louise, they know this corner has that crazy tree root and they know that they have to pay attention to that both in terms of the data collecting and also when the humans have to intervene. So they’re like, OK, the robot is coming to this corner with the crazy tree root. Maybe that’s a solution for the future. How much does that cost? Does that actually save you money in terms of labor? Which is theoretically the point of these robots. I’m not sure. I’d be interested to know.
Michael Calore: I mean, it makes sense then that a lot of the action that we’re seeing in delivery robots right now is on college campuses, because if you think about the layout of a college campus, you basically, you can roll a robot anywhere on wide paths that are well paved, well lit, areas that are blocked off for pedestrians only at all times. And it just seems like it would be a good testing ground for this technology, because it doesn’t have a lot of the same complex infrastructure that a city sidewalk does, especially in an area that is dense enough for you to be able to drive food around to people and still make money doing it.
Aarian Marshall: Yeah, that’s a great point. And that’s actually the strategy we see from a company called Starship Technologies, which is Estonian. I’d say they’re probably the biggest company in this space right now that stuck around, and they’re really going after university campuses. As you say, Mike, they’re better maintained in terms of sidewalks. People are often friendlier and from the kind of business perspective, the college students at 2 A.M, maybe you’re coming back from whatever you’ve been doing and you’re a little bit less price sensitive about how much that pizza costs or how much that burrito will cost and delivery is looking like a really good option for you. So it might be kind of the ideal market for this sort of thing.
Michael Calore: All right, well we need to take a break, but when we come back we’re going to dive further into what robots like Dot might mean for us living in cities, whether we are ordering in or not. So stay with us. Welcome back to Uncanny Valley. Today we are talking about all of the delivery robots and the renewed interest in them across the industry. Now Aarian, I have to say something that caught my attention from your reporting was this one detail that immediately made me internally grown a bit. The fact that Dot is trained to also use bike lanes. And this might seem like a minor thing, but for me and for many, many other people, it actually opens up a whole can of worms. We have somewhat good bike infrastructure in San Francisco where I live. I know it’s a little less nice in other cities where these robots are rolling out, particularly places like Phoenix, which has almost no good bike infrastructure outside of the city center. But the bike lanes have already become quite crowded with not only but e-bikes and cargo bikes and delivery bikes and two-wheeled scooters, which is great because it gets people out of cars when you have a car out of bike lane. But it’s also bad because now people feel like they have the right or are allowed to drive a five-foot-tall robot in the bike lane. As somebody who covers transportation closely and as somebody who lives in a city, I’m curious what you make of all this.
Aarian Marshall: Yeah, I agree, Mike. It’s something that definitely gives me pause, and as you pointed out, this is not the first time we’ve seen private companies try to infringe on that very limited bike lane space. We’ve seen companies like Amazon and DHL experiment with using smaller but still kind of big bikes to deliver packages, and it’s like, OK, that’s cool. As you say, it takes people out of cars, but that also takes up space that commuters might be using, might make it harder for more nervous cyclists. So I think it’s a great point. DoorDash is not the first company to experiment with bike lanes and robots. There’s a company out of Ann Arbor that’s a few years old that’s still, it’s smaller than it used to be, but it’s still sticking around, so this has been happening for a little bit. I’ll also point out that it’s also kind of complicated when it comes to regulations. During the height of the sidewalk robot craze, we saw Amazon actually going from state to state, and we broke the news on this, Amazon going from state to state, lobbying state legislatures to change their definitions of vehicles to allow these robots to operate on sidewalks. Now, that doesn’t happen everywhere, and there are also regulations in a lot of places that determine who’s allowed to operate in bike lanes as well. We’ve seen, for example, in a bunch of cities that e-bikes are technically banned from bike lanes, so it’s going to get complicated if these companies really want to operate these delivery robots in bike lanes all over the world, that they’re going to have to talk to a lot of politicians, which I know is their favorite thing.
Louise Matsakis: Man, I find this lame. I feel like we barely have the bike lanes for bikes in many parts of the country and the idea of now having to compete with robots kind of sucks. But if it means more bike lanes, I am sort of for it. I’d rather have the tech companies on the side of wanting to build more bike lanes than perhaps less of them for their autonomous cars. So I don’t know, the enemy of my enemy is my friend.
Aarian Marshall: Yeah, definitely.
Michael Calore: I think it would be hilarious if the law of induced demand came to bike lanes and we just started adding a third and a fourth bike lane in every place where there’s already one.
Louise Matsakis: That would be great. Honestly, more bike lanes is sort of my bottom line here.
Michael Calore: Fully. Well, OK, so that’s a transportation front, but we should also talk about the labor front, because the United States is home to the largest food delivery market behind China. We use food delivery apps a lot, and many tens and hundreds of thousands of people make their living from food delivery as contracted delivery people. So for all of the talk of AI coming to take our jobs, it seems like we are nearing an inflection point with more robots and drones becoming more capable. So my question to both of you is, do you think that this hybrid universe that companies like DoorDash are trying to build, that robots will outnumber delivery people? Will there be a balance? What will that look like?
Aarian Marshall: I find it hard right this moment to envision a world in which there are more robots than human delivery people. Just because when people pay the extra money to have someone bring their food to their house, and it is extra money, I think what a lot of people want is that food to show up at their door. And that might mean going up a few steps. It might mean buzzing through the gated community, entering a code, and these are all things that are really hard for robots but really, really easy for humans. It doesn’t take much for humans to do that. So really getting robots to that level of proficiency is going to take a ton of investment and I’m just not sure that’s worth it. But Louise, I’m interested to hear about the experience in China, which I know is different from here.
Louise Matsakis: So the interesting thing about China is that sometimes there is a robot portion of the journey and then a human part of the journey. And where you often see this is in hotels. So if you order food to your hotel room, usually the delivery worker will drop it off with the robot and then all the robot does is get into the elevator and get off at your floor and go to your room. That’s a really simple, predictable journey. And then you open your hotel room and there’s the robot with the boba you ordered or whatever. But that hasn’t changed the fundamental labor dynamic, which is that what’s happening is that that delivery worker can now do more deliveries, right? Because they don’t have to go all the way up to the room. It’s not that demand for their labor overall has actually been reduced. Delivery workers are still an incredibly large share of the overall working population in China, and I don’t see that changing for exactly the reasons that you described, Aarian. People want their food delivered to them right away. They want it to be really fast and the chaotic streets of a city like Shanghai are just really hard for an autonomous robot to navigate at this point.
Michael Calore: So that really is a true hybrid model and I like that, but I don’t live in a hotel.
Louise Matsakis: For sure. Yeah, it happens in some residential complexes as well, but it’s just the same really quick route in the elevator.
Aarian Marshall: That reminds me, Louise, of something someone told me while I was reporting out this article, which is these two researchers at Carnegie Mellon followed around this sidewalk robot test in Pittsburgh for a few weeks and they found that a lot of the human labor of delivering was actually just displaced onto other workers. So for example, someone had to clean out the robot when things spilled and someone had to make sure it was all in the right slots, and then as we said, someone had to come out to the curb and get the food. So it might not mean there’s less work, it’s just work in different places or more work for other people.
Louise Matsakis: Yeah, I think that that’s an incredibly good point. It’s often about modifying the labor rather than eliminating it.
Michael Calore: All right, we’re going to take a quick break and then we’ll come right back with recommendations. Thank you, both, for a great conversation about robots. We might not be DoorDash’s cute new Dot, but we are here to deliver some recommendations for our listeners. Do you like that? Is that good? Don’t answer. Aarian, you get to go first. What do you recommend this week?
Aarian Marshall: That was really good Mike. I’m going to recommend the JBL Clip 5, which is a Bluetooth speaker. I will caveat this by saying I’m not an audiophile. I’m not one of those really picky people, but I think it has really great sound. I got one recently for my birthday and I like to carry it around. It has this very handy clip that you can clip onto your shower, you can clip onto a backpack. If you’re traveling, you can clip onto, I don’t know, your dog and have traveling pretty good sound, and it’s a relatively inexpensive for I think a pretty good speaker and I’ve really enjoyed using it.
Michael Calore: Nice. How much is it?
Aarian Marshall: It is currently 69.95.
Michael Calore: That’s so cheap for a good Bluetooth speaker.
Aarian Marshall: I think so too.
Michael Calore: I mean, I’m sure there are cheaper ones, but that one is actually pretty good. It’s like my second favorite, I would say, of the ones that when people ask me which Bluetooth speaker should I get, I tell them the UE Boom or the JBL Xlip. I make it sound like people ask me that all the time, but really it’s only a couple of times a year.
Louise Matsakis: I was like, wow. I was like, wow, Mike, are you a gear editor or something?
Michael Calore: Yeah, yeah. They’re like, “Hey, you’re the WIRED guy. What should I get my kid for his birthday?” That kind of thing.
Louise Matsakis: That makes sense. They’re not like, “Best Bluetooth speaker of 2025. Go.” That makes more sense. Get ready to roll your eyes. But my recommendation is journaling. I am sorry, but I thought that this was terrible advice or just would not actually help me, but I’ve been doing it over the past month and it’s actually made a big difference in my life. No pen or notebook recommendations. However, I think whatever works for you though, I have really been enjoying an unlined notebook. So the blank page.
Michael Calore: Do you prefer the unlined notebook so that you can draw or write it a slant?
Louise Matsakis: Sometimes I put some stickers that I was gifted in one page the other day. I think just having the ability to make a chart or to write however space you want it to be for whatever reason, works better for me. I find the lines kind of constraining.
Aarian Marshall: Louise, do you go back and look at it afterwards or do you just feel like the process of doing it in the moment is helpful?
Louise Matsakis: Both, I think. I was very intrigued by a friend who has been doing it daily for I think now over two decades, and he mentioned being able to go back that far and see how he felt at the time, and that was incredibly appealing to me. So we’ll see if I end up going back. But so far, I do remember a few weeks ago, so I haven’t totally gone back. But I do read it when I’m done and just look back at what I just spent a half an hour writing or whatever.
Aarian Marshall: That’s lovely.
Louise Matsakis: What about you, Mike?
Michael Calore: OK, I’m going to recommend a feature that is probably on your phone if you have an Android phone that’s relatively new. It’s called Ask Photos. And it’s inside Google Photos, and it’s a conversational photo editing feature. So basically you go into Google photos, you click on the edit button, and then you have a new option to ask Google to do something to the photo. And you can say, “Can you make it better?” You can say, “Can you fix the skin tones here? Can you fix the lighting? Can you make this look more colorful?” You can give it English language instructions of basic things that you want onto the photo. This is all powered by Gemini, which is Google’s large language model conversational chatbot thing, and it debuted on Pixel phones a while ago, and Google has now rolled it out into Google Photos on every Android phone that is powerful enough to support the chatbot features. So I don’t really think that it’s important to use it to make your photos better. I think if you take a picture, you’ve taken a picture, and the picture that you share and the picture that you saved forever should be the one that you actually took and it should not be something that is like a simulacrum of something you took or something that you wish you had taken. Right? The truth of the photo is the photo that you snap in the moment and editing it afterwards is convenient. So I’m not advocating that you edit your photos to make them better, but you should try this feature because it’s kind of incredible. It shows you how LLMs and chatbots can actually fit into situations that you did not imagine LLMs and chatbots fitting into. Am I making sense?
Louise Matsakis: Yeah, totally. This sounds awesome. I think as someone who is kind of a camera noob, I don’t actually know what saturation technically means or a lot of these other features that come standard on smartphones, I think be able to just say the colors seem kind of washed out. Can you make them a little brighter? Or whatever. I think being able to interact that way with natural language sounds really cool and useful.
Michael Calore: And I gave a sort of similar caveat to this the last time I talked about using chatbots more, which is that I think that in order to properly engage with these technologies on intellectual level, you have to actually use them. You have to understand what they’re about. It’s like why I take a Waymo every once in a while, even though I don’t really like choosing the option that takes money out of somebody’s pocket, because you’re not paying a driver to drive you somewhere. But I think in order to really see how these things fit into our lives and understand why people are excited about them, you actually have to use them. So I’d recommend that you start talking to your phone to edit your photos, even though I don’t want you to be editing your photos.
Aarian Marshall: And don’t we live in a simulacrum anyway?
Michael Calore: You’re one of those. All right, well, thank you both for being here this week.
Aarian Marshall: Thanks for having us.
Louise Matsakis: Yeah, thank you. This is fun.
Michael Calore: And thank you for listening to Uncanny Valley. If you’d like what you heard today, make sure to follow our show and rate it on your podcast app of choice. If you’d like to get in touch with us with any questions, comments, or show suggestions, you can write to us at uncannyvalley@WIRED.com. Today’s show is produced by Adriana Tapia and Mark Lyda. Amar Lal at Macrosound mixed this episode. Mark Lyda is our San Francisco studio engineer. Matt Giles fact checked this episode. Kate Osborn is our executive producer. Katie Drummond is WIRED’s global editorial director. And Chris Bannon is Condé Nast’s head of global audio.