O3 beats a master-level GeoGuessr player, even with fake EXIF data

Re:O3 beats a master-level GeoGuessr player, even with fake ... by ceph_ on Tuesday April 29, 2025 @02:00PM (#43835936)

The AI should be forced to use the same rules as the human. Not the other way around. The AI shouldn't be using outside resources.

[ Parent | on HN ]

Re:O3 beats a master-level GeoGuessr player, even with fake ... by voxic11 on Tuesday April 29, 2025 @03:08PM (#43836743)

Another rule bans "using third-party software or scripts in order to gain an unfair advantage over other players."

So is it even possible for O3 to beat another player while complying with the rules?

[ Parent | on HN ]

Re:O3 beats a master-level GeoGuessr player, even with fake ... by ben_w on Tuesday April 29, 2025 @04:01PM (#43837364)

If a player uses such a model, the model is third-party and the player is cheating.

But: when a specific model is itself under test, I would say that during the test it becomes "first" (or second?) party rather than "third".

[ Parent | on HN ]

Re:O3 beats a master-level GeoGuessr player, even with fake ... by voxic11 on Thursday May 01, 2025 @12:39PM (#43860081)

I think third-party here means not produced by GeoGuesser rather than not produced by the player.

[ Parent | on HN ]

Re:O3 beats a master-level GeoGuessr player, even with fake ... by bscphil on Tuesday April 29, 2025 @02:38PM (#43836384)

I think that's part of the point they're making, hence "They should've simply redone the test."

[ Parent | on HN ]