Interestingly, I have some weirdly relevant direct experience to share here.

Pokémon Go was released at the beginning of July 2016. A week later, the Air Force kicked off its Red Flag exercise in Nellis AFB outside of Vegas. For the several thousand active duty folks participating, this is a month-long TDY from their normal base to Vegas. The premise is a large-scale war simulation, and it encompasses essentially all major wartime functions. I was directly involved in supporting drone operations (including live strikes) during this exercise.

The thing that’s funny about your comment is that because Pokémon go was just launched a week prior, a huge percentage of the participants were playing it in their downtime between exercises. You have to understand: these are thousands of 20/30-somethings (and occasionally even teenagers), meeting back up with friends from all over the world in Vegas. On and off base, people are socializing, having fun, and playing games. While phones were limited to outside of SCIFs, most of the base had no such restrictions. I recall wandering around base at 2am with friends playing it.

What’s funny is the same was happening with our deployed friends as well at the same time. This was a game that all their friends back home were playing, and when deployed, you need all the morale you can get. There were technically OPSEC policies this all probably violated but this was before any blowback from Strava accidentally revealing military bases or other similar incidents, so there was no specific guidance or moratorium on it.

All this to say, I understand what you’re trying to arrive at via deduction, but I think your understanding of the world in this case may be a bit too limited to meaningfully speak to this. That said, is the headline sensationalist? Probably.

Even the Pokémon Go world model headlines were stretching the reality of what the model captured.

If you’ve played the game, the scanning function is only for what they call Pokestops: These are points of interest that you can walk to and get items in the game. The game gives you points if you walk in a circle around one and take a short video.

They’re relatively sparse. At most, they captured some 3D models of some things like signs, small landmarks (up close) and the fronts of some buildings.

The images captured by something like Google Maps are a million times more useful for someone trying to construct a world model with a lot of coverage. The Pokémon Go captures would be useful if you wanted something like a detailed 3D scan of the sign in front the student building or something.

> The overlap in the locations of Pokemon Go Player data and any active Drone heavy theaters of war is a tiny sliver.

Currently active theaters. And now there are detailed locations of our cities. We might not get killbots today but we will get pacificationbots.

> The overlap in the locations of Pokemon Go Player data and any active Drone heavy theaters of war is a tiny sliver (or zero?).

But presumably the images/models at ground level can be used to train/improve the general performance of Vantor's aerial (satelite based) navigation system so it works better elsewhere?

> overlap in the locations of Pokemon Go Player data and any active Drone-driven Theaters of War would be a tiny sliver

Is Pokémon Go not played in the Middle East, India, Taiwan, Korea or Japan?

Pretty sure Pokemon Go and Ingress was played in Kiev long before the war

As a Pokemon go player, I would say it isn't.

There's even a Pokemon exclusive to the middle east region: sandstorm pattern Vivillion. Lots of players there.

> The overlap in the locations of Pokemon Go Player data and any active Drone heavy theaters of war is a tiny sliver (or zero?).

For now.

Will the drones at least look like Pikachu?

i thought Maxar was mostly 3d images based on satellite inference. 1/2 of a pixel difference in a morning vs. noon vs. night sat photo can determine shadow and therefore height, etc. etc.

rapid 3d modeling of topography and cityscapes + supplementation with other data, e.g. pokemon. But ultimately that's supplementation, not the main effort.

The current drone-heavy theatres.

That could change in an instant.

Vantar: "None of these places in our training data are in active theaters of war!"

Also Vantar: "The superpower of generative AI is that data in one task generalizes to other tasks!"

This is a massively weak argument. It's like its own strawman, one does not see this often, lol.

If you train a soldier in the US, is he unable to do those things outside the US?

So by this conclusion we can assume, that these drones will be somewhere else. Somewhere in heavily populated areas right?

""As someone who works in this space, the headline is a bit of a stretch. The overlap in the locations of Pokemon Go Player data and any active Drone heavy theaters of war is a tiny sliver (or zero?).""

Can you elaborate?

GPS can be faulty in cities.

Pokémon Go scans, are primarily in cities.

The mapping in the article, is specifically saying to use visual cues when GPS is faulty.

How is this not directly 1-1 overlapping, the gap and the solution.

>As someone who works in this space, the headline is a bit of a stretch. The overlap in the locations of Pokemon Go Player data and any active Drone heavy theaters of war is a tiny sliver (or zero?)

are you saying that drone training in quiet residential neighborhoods is not training? are you saying self driving cars can only drive in theaters where they've been trained, because autonomous training is always specific by neighborhood? are you saying that if a particular region has some novel terrain that all previous training must be discarded?

It's not like there's a moral high ground about not collaborating with the military. Unless you want to advantage America's adversaries, namely China, Putin's Russia and Iran's current regime. There's always this implicit, sometimes explicit, "war bad" childish political philosophy in posts like this. In reality war is a given and you have to be prepared to have the upper hand.