Tag: GPS

Posted on
by

Back in December 2024, Niantic posted about a large geospacial model they were building using data from Pokémon Go.

That led to a lot of online talk, with takes ranging from:

  • Well yeah, that’s what we figured they were doing with the “Scan a Pokéstop to build better AR models” feature.
    to
  • Watch out! Niantic is building a global AR model using every image that passes in front of your phone’s camera!

I was in the “of course that’s what it’s for” camp. When the game rolled out the Pokéstop scanning feature a few years earlier, it seemed obvious that it was training 3D machine vision, like how all the “pick the squares with bicycles” CAPTCHAs are obviously training for self-driving cars. I figured there was a good chance someone would use it for some harmful purpose or another, probably surveillance, so I skipped those tasks.

Anyway, after a week or so, Niantic updated the article to clarify* that it was using the deliberate Pokéstop scans in public places for Pokémon Playground, not any of the other AR features like taking a photo of your buddy in the kitchen.

This made sense, because if they were using that data, it would have eventually gotten better at placing a Pokémon in my kitchen. (The floor’s a grid. You’d think that would help, but noooo….)

Drones (And not just Beedrils or Combees)

Those scans are back in the news, because as DroneXL reports, that geospatial model is being used for camera-based drone navigation.

Including military drones.

Because of course everything has to be weaponized. Allegedly even Pixar’s RenderMan.

Admittedly, GPS itself started as a military technology long before it became civilian infrastructure. Military and civilian tech really do just have a revolving door between them, don’t they?

Training Data

Among other sources, DroneXL cites a Dutch-language article at Trouw, who asked the defense contractor (Vantor) directly whether it uses Pokémon Go data: Vantor initially said no, but later walked back any guarantee. Niantic Spatial, however, has stated that the Pokéstop scans were used to train an “early version” of their model. That means the data (or weights produced from it) is still in there, just blended so much by training process that it can’t be identified anymore.

Kind of like you probably couldn’t confirm my old blog posts are in the training data for an LLM by looking at the LLM weights, but you can find pages from hyperborea.org in Common Crawl data, and assume any model trained on Common Crawl still has it in there somewhere.

Maybe scans made since Scopely (US-based, Saudi owned) bought Niantic’s gaming division last year haven’t gone into the map built by Niantic Spatial (still independent), so Vantor technically isn’t using current player data. Or maybe Niantic Games continued passing scans along to Niantic Spatial for a while, under the separate TOS, and Vantor’s spokesperson just hadn’t made the connection.

Quietly Dropped

Curiously, the Pokéstop scanning task I’d left in my list for years just disappeared a few days ago.

At first I deleted the tasks as I got them, but every time I scanned an eligible stop it would add a new one if I didn’t have one in my list. So after a while I just left one there and ignored it like an ad banner.

It turns out Pokémon Go discontinued the features on June 2, just three days before the Trouw article was published. (New tasks stopped appearing that day, and it took a few days for old tasks to disappear.)

Coincidence? Maybe. But the timing’s certainly suspicious.

Notes

* Before Niantic published their update, I e-mailed them asking for clarification. It took them over a month, but they did eventually reply:

Hi Trainer, we appreciate your patience. Thanks for your questions about AR Mode and our Privacy Policy. I’ve shared some additional information below:

For Pokémon GO, only AR scans from the PokéStop Scanning feature will contribute to the development of the Large Geospatial Model. As noted in the PokéStop Scanning Help Article (https://niantic.helpshift.com/hc/en/6-pokemon-go/faq/2519-scanning-a-pokestop/): information gathered during PokéStop Scanning allows Niantic to generate accurate, dynamic 3-D maps of real-world objects and their relative locations, and help devices understand the surroundings in AR real-time. As noted in the Editor’s note to the blog post, merely playing the game does not train an AI model.

When using AR or AR+ mode, we do not store your photos on our servers. For PokéStop Scanning, once a PokéStop scan is voluntarily uploaded, the video recording and associated camera data is retained on our servers in accordance with our data retention policies. For more information please see our Privacy Policy (https://nianticlabs.com/privacy).

Posted on
by

GPS navigation options we need:

  • I know how to get to the freeway from home.
  • I know how to get home from the freeway.
  • Don’t send me down someone else’s narrow residential streets just to save two minutes.

If I’m trying to get somewhere other than home after work, I’ll use GPS to get an idea of the time remaining and the fastest route. Since I’d rather avoid the freeway during rush hour, it keeps trying to send me on these zigzag paths through residential neighborhoods to avoid backed-up arteries or just avoid busy intersections. I used to follow those routes, but after a while I started noticing other cars ahead of me that were clearly doing the same thing. It’s not just one car being added to that lumpy narrow road with lots of driveways, stop signs, kids on bikes and people taking out the trash. It’s a lot of cars. And of course we’re following the same apps drawing from the same data, so we’re all taking the same side streets, not spread out among all of them.

If there’s a big difference, that’s one thing, but for two or three minutes? What’s the point?

Of course the navigation app seems so testy when I decline to be part of the problem, and it has to keep recalculating…

Posted on
by

I use navigation on my Android phone to pick out the best route to work each morning. The problem is, it bases time estimates on traffic conditions now — not traffic conditions as they’ll be when I get to each point along the route. I’ve gotten used to the morning drive taking at least 15 minutes* longer and the evening drive taking around 10 minutes less than predicted, but a little more precision would be helpful.

Obviously, Google isn’t psychic. They can’t predict where and when car crashes will happen. But they do have historical traffic data. If you go to Google Maps on the web and display traffic, you can switch between live data and an average for a given time and day of the week.

It would be fantastic if Google used that data to predict how much slower (or faster) traffic will be moving at each point along each projected route, and use that for the time estimates. It would be nice for the “Are we there yet?” factor, but it would be incredibly useful for route planning!

*Sometimes more. This morning, it predicted a 55-minute trip. It took me an hour and 35 minutes.

Posted on
by

I’ve never been a fan of actually using GPS navigation. Sure, I’ve always thought it was insanely cool that it was possible, I just didn’t want to use it myself. For unfamiliar destinations I generally prefer researching a route first, and for familiar ones I generally prefer just relying on my local knowledge. But I’ve found something that I do like using it for: Traffic.

I recently started a new job, exchanging a fairly short commute for a ~40-mile trek across the Los Angeles freeway system. Under ideal conditions, it’s about 45 minutes. When the freeways are bogged down (i.e. when I’m actually going to be driving), it can take an hour and a half or more.

When I landed the job, I replaced my phone with a G2. It’s a heck of a lot faster than my old phone, plus it can handle newer software…like Google’s turn-by-turn navigation app for Android. After trying a couple of different routes the first few days, I tried it out…and discovered that it factors in live traffic data when calculating the remaining time.

The upshot: I can walk out the door, start up the app, and figure out which of three main routes will get me there fastest. (Well, least slowly, anyway.)

Of course, it’s not perfect. It’s based on traffic now, and over the course of a predicted hour-plus, the route could easily get more congested. That’s not even counting potential accidents. It does seem to update frequently, though, and knowing I’ve avoided a 100-minute drive in favor of 70 minutes really outweighs the annoyance of a mechanical voice telling me how to get to the freeway from home.

I do have to remember not to rely on it too heavily at the end of the trip, though. I left it on by mistake after selecting my route to the LA Convention Center for Adobe MAX this morning, and instead of turning it off, I let it direct me straight past the parking garage.

Oops.

»All pages site-wide with this tag