Hacker News Reader

Comment by colinbartlett | Hacker News Reader

I would love to know how much Google spends on StreetView.

When they first started sending those cars around, I think everyone collectively mocked, "Haha. Well you're certainly not going to photograph every street in the world."

View on HN

What about "Haha, Well you're certainly not going to create a 3D point cloud of every street in the world"...

Google's self driving cars are just doing that.

EDIT: I can't find the post I once read, but this post also makes some clear: http://www.linkedin.com/today/post/article/20130502024505-99...

EDIT 2: Here a video of the presentation: https://www.youtube.com/watch?v=YXylqtEQ0tk @ 3:30 you can see the generated (and stored) point cloud.

That bit blew my mind.

I'd love to see some of those point clouds. Maybe Google has an artist-in-residence who could 3D print varies landmark locations?

That was actually a little bit sobering for me. I thought the AI techniques they use were much more general, autonomous and independent of map data. I didn’t know that location information of traffic signs, pavement etc. are taken from map data. But it makes sense to prefer data from most reliable sources.

Google's pretty good at thumbing their noses at the naysayers that say "Haha...you're not going to xxx all yyy".

xxx = crawl and cache, yyy = reachable pages on the web xxx = scan and preview, yyy = published books etc

When I first saw the article for the google patent on free ad-powered taxis[1], I immediately thought that it was another way to get get more cars for street view.

[1] https://news.ycombinator.com/item?id=7114079

I think they buy a fair bit of their data from third party vendors, but I'm not positive. Still, aggregating and organizing all that information is impressive.

Reply on news.ycombinator.com