Hacker News new | comments | ask | show | jobs | submit login

I'm surprised by the comments that imply we'll solve traffic congestion and be snoozing safely in the passenger seat as our robot cars zip us to work in 10-20 years. I live in a technical world where a team of geniuses can barely keep a system up that exchanges 140 character messages.

"Robot cars" don't need to deal with the sort of scale or coupling that causes systems like the one you described to fail (the system you reference itself doesn't really need to be that way, technically... it is that way for the purposes of centralized control).

The "robot car" I may be sleeping in doesn't have to worry about what the other 300 million cars on the road are up to, only the 10 or so that happen to be immediately around me.

Yes, there are real problems to be solved in this area but they are very different sorts of problems than scaling and at the current rate of progress I wouldn't be too surprised if automated driving were available as a high end option on some cars within 15 years.

We already have robots designed for games that drive cars, even besting most of the people skill-wise, and we have a nice traffic congestion example out of the network and medium access protocols too in the telecommunications science. so I guess that those problems are already almost sorted out.

Video-game bots are a nice research field for artificial intelligence which can provide solutions for the real world.

Unrelated: Check this out if you're interested in the topic, it's a nice read http://www.cs.rochester.edu/research/quagents/QuakeIII.pdf

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact