It's astounding that a vision model can do this. When it comes to the fact that this is possible at all following various geolocation and OSINT accounts has ha

It's astounding that a vision model can do this. When it comes to the fact that this is possible at all following various geolocation and OSINT accounts has had a quite sobering effect on me quite some times me ago. I would like to point out josemonkey particularly, because he really explains the process very well. By and large I think it is first and foremost a fairly tedious task, but doesn't require any access to special services or special skills. Of course the vision model could change the tedious part. Overall I think what the OSINT people do and what the vision model from the article simplifies is entertaining, but ultimately not super relevant, because reality is more like xkcd 538[1]. Most people have a device on them that in some way reveals their location more or less accurately. Location data in the grand scheme of things is a commodity for quite some time already. While I think it is worth fighting that, I guess pragmatically we should all live under the assumption that our location is public information. [1] https://xkcd.com/538/ https://xkcd.com/538/