For months members of the public have been using GeoSpy, a tool trained on millions of images that can find the location a photo was taken based on soil, architecture, and more. It's GeoGuesser at scale.
Honestly it’s a bit trickier. I listened to the podcast episode and my 1st thought was like “Damn… obviously bad” and during the FOSS part I couldn’t but think “Ah… for my photo with missing EXIF GPS data… that’d be neat” specifically because I’m using Immich to host my photo on my server. Now Immich has face recognition, cf https://github.com/immich-app/immich/tree/main/machine-learning and more. Would it be wrong to extend it?
Honestly it’s a bit trickier. I listened to the podcast episode and my 1st thought was like “Damn… obviously bad” and during the FOSS part I couldn’t but think “Ah… for my photo with missing EXIF GPS data… that’d be neat” specifically because I’m using Immich to host my photo on my server. Now Immich has face recognition, cf https://github.com/immich-app/immich/tree/main/machine-learning and more. Would it be wrong to extend it?