I was late for dinner when a neon-pink Coco bot rolled up my block and paused as a man slept on the curb. A teenager shoved it, laughed, and the machine whirred away like nothing had happened. I felt a small, private dread: those bots know more about the city than we admit they do.
I write about technology that changes the city because I want you to see the trade-offs before the PR departments paper them over. You’ve probably spotted the bots—googly eyes, names, a polite hum as they weave past bike racks and puddles. Companies hope charm will override suspicion. That may be wishful thinking if those bots keep bringing footage to police departments while trying not to get stomped on.
A Coco bot hesitated outside my corner bodega last week.
That moment—machine, human, and cracked sidewalk sharing the same three feet of space—tells you everything about the rollout problem. Delivery robots are solving one obvious friction (your fries delivered without a tip jar) while colliding with another: the city itself. Narrow sidewalks, glass towers that scramble GPS, and the unpredictable human element make route planning a nightmare.
Enter Niantic, the company you remember from your old Pokémon Go afternoons. Their idea is simple and raw: use the millions of geotagged AR photos players uploaded to rebuild cities in much higher detail than GPS allows. Imagine a quilt stitched from tourist snapshots; that’s the map the bots could one day use. For Coco Robotics, which favors bright pink bodies that are hard to miss, those stitched maps promise fewer wrong turns and fewer angry encounters with passersby.
How does Pokémon Go data help delivery robots find addresses?
Niantic’s AR snapshots fill in where GPS blurs. In dense downtowns your phone’s blue dot can wander 50 meters and put you on the wrong block; Brian McClendon, Niantic’s CTO, told MIT Technology Review that urban canyons are the worst place for GPS. By turning player photos into 3D models of sidewalks, store fronts, and lamp posts, robots get more reliable visual cues. That’s the technical pitch: less guesswork, fewer dead-ends, and shorter delivery windows.
Downtown, someone kicked a delivery robot and walked away without looking back.
Vandalism is not hypothetical. Newspapers like The Economist have tracked a rise in attacks, and stories have shown bots being used as catchalls for public anger. But there’s another, less-visible worry: surveillance. Emails and reporting have revealed instances where robot camera footage made its way to police departments such as the LAPD. Now imagine those cameras paired with Niantic’s dense visual maps.
I’m not asking you to panic; I’m asking you to weigh what you want a delivery bot to be. Cute mascot? Mobile camera? Both? The partnership between Coco and Niantic ties a beloved game’s data to machines that physically traverse neighborhoods—a mix that will comfort some and alarm others. The robots are starting to look less like pizza carriers and more like neighborhood canaries.
Are delivery robots sharing camera footage with the police?
Yes—there’s documented evidence. Internal emails and local reporting have shown companies providing footage to law enforcement. The public reaction was predictable: anger, distrust, and more attempts to disable the hardware. That backlash is one reason operators lean into PR moves—product placement on Netflix, cartoonish faces, celebrity cameos—to soften the optics.
On my phone, the blue dot still wanders between blocks when I walk beneath glass towers.
You probably remember Pokémon Go’s explosion in 2016: streets full of players, digital critters pinned to real-world places. Niantic says the game’s footprint remains large—hundreds of millions of installs and tens of millions still playing years later—which translates to a unique data set. Those AR photos aren’t just snapshots; they’re a crowd-sourced atlas of granularity most mapping firms can’t match.
Coco and Niantic argue that bolting that atlas to autonomous delivery will cut delivery time and errors. Competitors like Serve Robotics and apps such as Uber Eats are still wrestling with reliability in cities where buildings jam GPS signals. If this partnership reduces wrong turns and idle bots, companies avoid repeated re-routes that cost money and cost patience.
Will Niantic’s data make bots less likely to get lost or beaten up?
It could lower the “lost” part by giving robots better visual navigation. Whether it reduces attacks is a social question more than a technical one. Better routing means fewer awkward sidewalk standoffs, but it won’t fix the resentment that comes from people who see these machines as symbols—of gig-economy pressure, of surveillance, of changing public space.
Here’s what I advise you to watch: how companies handle camera data, the transparency of police requests, and what local governments decide about robot behavior in public. You should also watch what happens when PR stunts—like Serve’s Saymo cameo on a Netflix talk show—meet real-world friction. Niantic and Coco can make deliveries faster; they cannot alone make people like the thing doing the delivering.
So you tell me: will grafting our childhood game onto these machines make city life better, or will it just give the robots better maps to be watched by the rest of us?