Comment by iamleppert | Hacker News Reader

Comment by iamleppert | Hacker News Reader

It seems to me, that Waymo/Google has no real interest in their self-driving car program.

Anthony Levandowski went to Uber because he is interested in furthering self-driving car technology into practical applications and Uber offered to actually, really prioritize his project and make his dreams a reality. That's the only logical explanation of his actions; he thought Uber would be a better technology and business partner.

Setting aside the legal issues of whether what he did was right or not, I don't think Google would be the right custodian of this project anyway. For the greater good of the world, we need self-driving cars to become a reality.

And Google just isn't going to do it. They are invested in self-driving cars like they are invested in contact lenses, or drones, or robots, or go-playing AI, and half-dozen other things. Uber was/is a far more logical choice for a company that is going to be both motivated and pragmatic enough to bring this technology to market.

Google has been working on it for ages and we have yet to see a serious attempt at a non-academic implementation.

(I work at google)

>Google has been working on it for ages and we have yet to see a serious attempt at a non-academic implementation.

This seems like a claim made without knowledge. The 2016 autonomous vehicle disengagement reports [0][1] show that Waymo is an order of magnitude ahead of any competitor in miles driven and vehicle reliability. And as another user mentioned, Waymo is currently testing with real consumers and a large fleet of vehicles in Phoenix [2].

[0]: https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/disen...

[1]: https://www.wired.com/2017/02/california-dmv-autonomous-car-...

[2]: https://waymo.com/apply/

Greed is another logical explanation. I don't buy the narrative of Levandowski as some tortured idealist who just wanted his dream to shit rainbows, and Google as the neglectful parent stifling his creative growth. If that was the case he would have stuck with Otto for more autonomy.

I think Google just understands the complexity of the issue presented by true self-driving cars. I'm still highly skeptical that they're going to become a thing outside of cities and major highways anytime soon, there are simply too many edge cases to cover. That's not to say it's impossible, but Google is clearly shooting for pure self-driving cars that don't need a steering wheel, or something close to that, prior to any commercial application. Tesla and Uber are taking more iterative approaches, but they can directly benefit from mediocre/incomplete self-driving in the meantime whereas there's no incentive for Google. We'll see who gets to the holy grail first, but I wouldn't discount Google. Large-scale, public, iterative approaches, while the dominant religion in modern tech, are not a panacea.

Umm, Waymo has an early rider program... https://waymo.com/apply/

I don't see how they have no interest, none of their other projects were upgraded to separate company status. Also, they've been working on it for 10+ years so it's outlasted most of their other "projects"

I forget where I read this (somewhere on HN most likely), but there is a major difference between what Google/Waymo is doing and their competion. It all goes back to the DARPA days; The waymo approach is a hybrid one, its terrain detection using LiDAR as well as visual based systems. To see and "understand" the environment using a mix of these technologies is very different than what, say Tesla is doing, which is using cameras to do what is basically advanced lane following (although it is still fantastic technology). I believe that its this dichotomy in approaches which causes what people interpret as Waymo falling behind, or well, not being as ahead as they once were. Many other companies are catching up using visual based systems, but I believe there is a wall they will hit, where they will need to go through a major iteration of their technology to match what Waymo is currently doing.

> Google has been working on it for ages and we have yet to see a serious attempt at a non-academic implementation.

I trust Google way more than Uber to do something like this right (i.e. not kill people), even if it takes "forever."

As much as I personally dislike Google, I have to agree with you here.

Reply on news.ycombinator.com