world's best ai vs geoguessr pro

Channel:RAINBOLT · Length:25:22 · Published:2023-05-11 · Views:4399484

Watch on YouTube

Summary

GeoGuessr pro Rainbolt plays head-to-head against an AI built by Stanford students that guesses locations from street-view images. The video shows the AI repeatedly beating him.

Why HN recommends it

Cited as an early, concrete example of AI outperforming top GeoGuessr players, in threads on AI photo geolocation.

The best player in the world, Rainbolt, played against an AI out of Stanford — HN, 2024

Machine learning & deep learning · Computer vision, speech & NLP · Gaming culture & analysis

Subtitles (excerpt)

AI. I've 1v1ed AI twice before and I've won both times quite easily. But today, like I reached out to you by these students from Stanford University that said they built a Geoguessr AI for their class that I will not be able to beat. Let's see if it's finally time if an AI can beat a pro Geoguessr at Geoguessr or if I'm losing my job. Lion, we're back. It's time to face the AI again. Quickly take me through your AI that some of my audience might have seen, some of them might have not seen. Where we Where we stand with your AI right now? So, my AI since the last match with you, I would say it also improved a bit. I switched from learning it completely from scratch to something that's called ResNet, pre-learned model. Probably not the best AI today. So, you're saying it's not the best AI. I guess this is a good transition to a Stanford. Okay, so I kind of forgot to just introduce these guys, but just for reference here on what I'm going up against. This is Michael. He's currently getting his master's in computer science at Stanford. This is Lucas or Shawn Mendes. And he's getting his master studies in computer science at Stanford as well. He also is in his last year in AI research focusing on large language models and the

On Hacker News

Some people are really good at GeoGuessr, but also their best performance are more likely to get views. If you want a bot that is extremely strong at geoguessr there is this: [LINK] One forward pass is probably faster than 0.1 second. You can see its performance here: [VIDEO] (rainbolt is a really strong player) — GaggiX, 2025-04
> Technology can identify locations from photographs now. Tbf, it could do that before, and probably still better than the LLM: [VIDEO] But seeing it as what appears to be an emergent capability in such a general model is something else. — c7b, 2025-04
If you watch the video, (one of) the reasons why the AI was winning was because it was using “meta” information from the Street View camera images, and not necessarily because it’s successfully identifying locations purely based on the landmarks in the image. > I realized that the AI was using the smudges on the camera to help make an educated guess here. [0] [VIDEO] — karlding, 2025-04
Rainbolt is one of the world's best GeoGuessr players: [LINK] AI repeatedly winning against him in a head-to-head: [VIDEO] — echelon, 2024-05
The question is, how does this do at GeoGuessr, where users are given a picture from Google street view, and are asked where it is on the world by clicking on a map of the world. Users get points based on how close it is, user with the most points after N rounds, wins. The best player in the world, Rainbolt, played against an AI out of Stanford, so I wonder how this one would do. [LINK] [VIDEO] — fragmede, 2024-05

Related videos