Hacker News Reader

Comment by Shank | Hacker News Reader

I know someone did a detailed analysis on this, but you can trivially see that something is wrong by looking closely at the 2nd to last photo [0]. The photo clearly shows multiple phones with the exact same image on the screen, i.e., they are pixel perfect copies. While I don't think the article is implausible, the 2nd to last image is certainly photoshopped.

[0]: https://simonweckert.com/img/googlemapshacks/2.JPG

View on HN

Not conclusive, but it was edited in photoshop [1], whereas the other images were not [2]

[1] https://fotoforensics.com/analysis.php?id=b98e447834ffbc36d3...

https://fotoforensics.com/analysis.php?id=6f4b50d6f01c001874...

Damn, all this website does is show an image file's metadata. I thought it was going to be some cool AI thing that detects whether an image is photoshopped via some kind of fancy ML model. Does such a thing exist? Is it possible?

It was also edited in GIMP on Mac before PS on Windows.

While also not conclusive, but apparently the editings seem to be re-save only. But again, you can choose not to save (or only save some) edit histories in XMP.

If they are all running Google Maps at the same location why would the image be different on them?

Because gps is not that precise. Put two phones on top of each other and each will show a slightly difference in location data

Maybe, but how can you tell that there is a few pixels of difference by looking at that photo?

They're all open in Maps, of course they have the same location and thus screen image.

The EXACT same location and compass orientation? I can't even get 2 phone magnetometers to agree most of the time, let alone a wheelbarrowfull

Why not ? They get the quite exact same « wrong » data from GPS, so should be getting quite exact same « wrong » location.

In my experience, my Iphone 11 is quite often wrong up to 20-30 meters, depending on nearby buildings or other environmental factors. Other phones at the same location at the same time can also differ the same distance, but not necessarily in the same direction.

Depending on the use case, within 50 meters of each other may or may not be considered ”the same location”.

Magnetometers don't use GPS and IME can even be thrown off based on their proximity to metal objects. Things like multipath also won't impact all devices equally if they're adjacent

And if they're on top of each other, wouldn't they create issues with the touch screen? They would certainly move the screen a bit.

Good catch, despite that, there is also almost no glare from sun, which is increases the chance to be photoshoped. Given that most phones adapt to env lightning. The ones with maps on it look like paper glass almost.

Yeah, it's been a few years but I recall people settling on it most likely being fake. Unfortunately the original publication used the word "fake" so now all Google results just leads to different rewrites of the same original article. I guess there's a lesson there if you're planning on making up a story.

Reply on news.ycombinator.com