Hacker News Reader

Comment by Animats | Hacker News Reader

Now that explains some things. The initial Cruise Automation system was lane keeping and auto speed control, which everybody has. It was over-promoted as full automatic driving on the Cruise web site and Youtube.[1] "Just sit back and enjoy the ride", says the video.

The founder was credible enough to sell GM on buying their 40-person startup for somewhere between $500M and $1BN. GM already had the Cadillac/CMU technology self-driving technology, which was better than what Cruise had. But GM bought anyway.

One of the best deals made since Bill Gates sold IBM on adopting DOS.

Meanwhile, Cruise now has multiple Velodyne laser units on a rack on top of the car, like everybody else.

[1] https://www.youtube.com/watch?v=uEQqQj_zyHI

View on HN

GM had the Cadillac/CMU technology, but due to...I'll call it managerial oversight, they lost their initial IP to Google in 2009 (basically everybody had left and took their knowledge with them) and then again to Delphi in 2015 with the Ottomatika acquisition.

GM knew it needed self-driving technology, and Cruise was in the right place at the right time. Kudos to Cruise for pulling it off. Hopefully GM doesn't repeat themselves and effectively kill the technology for a 3rd time.

Ah. That may go down in history as one of the biggest big-company R&D management failures since Xerox couldn't figure out what to do with PARC. That would explain overpaying for Cruise - top management doesn't have to explain that failure to stockholders.

You should learn about the history of GM. Specially from the early 70s to the late 90s. Lots of horrible management decisions. You will learn how they gave toyota and honda the US domestic market on a silver platter.

I know. I read De Lorean's "On a clear day you can see General Motors" (1979) and Lutz's "Car Guys vs. Bean Counters" (2013).

Lutz makes a key point about automotive. He says that all the functional needs from automobiles have been met. You can buy just about any car from any make and get where you need to go reliably. Now it's all about the things which drive non-rational decisions about what car to buy.

Self-driving changes that. It's the first big change in basic car function in decades. That's going to have a big impact on the automotive sector, but nobody knows yet how it will play out.

You are watching a two year old youtube video and then confidently extrapolating to conclude that Cruise has no tech and that GM got duped out of a billion dollars.

And this DMV crash report from February of this year.[1] Cruise hit a parked car in San Francisco at 20mph. The driver tried to intervene and prevent it, but too late.

When you read Google's crash reports [2], it's almost always "autonomous vehicle was rear-ended by human driver". That's happened twice at Phyllis Av. and Grant Rd. in Mountain View. It's worth a look at that location.[3] Notice the tree in the median strip on Phyllis. That's in just the right place to block Google's car-top LIDAR scanners from seeing cross traffic on Grant Rd. So the autonomous vehicle advanced at 6mph to get a better view, detected cross traffic, stopped, and was rear-ended. The same thing happened a month later, last September. The autonomous vehicle did the right thing (not seeing "hit by cross traffic" accident reports), but it wasn't what the driver behind expected.

That's what the problems look like in the later stages of development. Not reports of hitting parked cars.

[1] https://www.dmv.ca.gov/portal/wcm/connect/bc21ef62-6e7c-4049... [2] https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/auton... [3] https://goo.gl/maps/T1Q33rtB67F2

I saw the tech they had in their office right before acquisition. It wasn't anything drastically different from the tech used in 2007 during the DARPA Urban challenge.

Well, all they are showing is that they have no tech.

Even the photos in this article are a bunch of Chevrolet dumps charging.

there is 15+ self driving startups at the moment, GM has prolly seen all of them and went with Cruise

That isn't sound logic. GM has a track record of making terrible managerial decisions, so assuming they know best is ignoring history. The people making the decisions haven't worked on self-driving cars and are not the best people to judge the technology, in fact these are the very people who twice (technically thrice, but that's another story) abandoned their own self-driving R&D which ultimately went to Google and Delphi (and another which was just completely thrown away).

You're also assuming that all of the other companies wanted to sell, which isn't true.

Otto sold for 680 million to Uber in August 2016, the company started in 2015, that's atleast one company that was willing to sell and I figure most of them would, they all have no end game other then being acquired for the most part

http://www.recode.net/2016/8/18/12540068/uber-paid-680-milli...

time will tell if the Cruise acquisition pays off or not

Otto was designed to be sold to Uber before it was created, there's a reason it was self-funded despite having nearly 100 employees by the time it was acquired. Anthony Levandowski met Travis Kalanick years ago and had been planning something like this for a while. It's basically an open secret.

we will see a few more acquisitions play out in this space over the next year or so but I do believe Google will have a hard time acquiring the companies in the space as tons of people are jumping ship and some are even starting their own self driving startups with prolly no intention of going back to google like Otto

they abandoned that retrofitting hardware a while ago too, they only did that for a few months before working on a more full stack solution that is level 4 that would require a partnership with a car company to install their product which they now have

That YouTube video is hilarious, that steering is twitchy as fuck. It's like someone used PID control without a clue.

Tesla is really good at precise lane-keeping.[1]

[1] https://www.youtube.com/watch?v=WZ3DV671CT8

Reply on news.ycombinator.com