Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Google Is Developing Its Own Uber Competitor

bloomberg.com

Post reply on HN
Print view View original

141–150 of 294 posts

Re: Google Is Developing Its Own Uber Competitor

#141

Post by jfoutz »

threeseed wrote: post #128
jfoutz wrote: post #110

Earlier quoted context omitted.

I think the self driving car is a tougher problem to solve than the maps. If uber can buy self driving cars, they can make maps on a city by city basis. If you have a pile of money, you can get maps. If you have a pile of money, it's not really clear you can get an autonomous car. I'm the first to shout they're coming - but they aren't here yet.

Autonomous cars are here: https://www.media.volvocars.com/global/en-gb/media/pressrele... And you can find examples from almost every car company. They all have projects going back decades looking at autonomous technologies. Some you see today e.g. parallel park assist, crash avoidance, lane/stop sign detection etc.

I understand what you're saying, and i agree there are a ton in testing. But it's not like i can go to the volvo website and check the box for the "no steering wheel" option at any price. Furthermore if i'm say, WalMart, I can't call up any truck manufacturers and get a self driving option for my next fleet purchases. That's in spite of having a huge infrastructure for supporting and maintaining my trucks.

Today, unless your title is "Director of self driving car research" you can't just go take one for a spin. Again, they're coming, they are the future. Uber wouldn't have cleaned out Carnage Mellon robotics department if it didn't still require original research.

Re: Google Is Developing Its Own Uber Competitor

#142

Post by vidarh »

maxerickson wrote: post #133
vidarh wrote: post #130

Earlier quoted context omitted.

The TIGER data is a starting point . But it's outdated the day it is complete, and rapidly diverges. For many purposes that's perfectly fine, but as stated above, you need someone driving the street. Even something as major as postcodes changes multiple times in between TIGER updates, and tons of streets are changed or added or removed. The TIGER dataset is fantastic, and certainly helps levelling the playing field b…

Have you looked at the last couple years of updates? It's way more accurate and complete than it used to be, and for getting names on that last 1%, they have drivers and the ability for the app to say "I don't know that street name, is it new?".

I do some work for a company that does mapping, so yes, I've looked at it regularly. It may be way more accurate and complete, but it is still updated too rarely to ever be up to date. Cover any reasonable area, and road situations change daily, especially when you're doing things like driving which also needs to account for temporary road closures and the like.

As I say, it's a good starting point, but it's not in any way sufficient, and keeping it up to date is a huge amount of work, which has been reflected in the value of companies that have set themselves up to be able to keep up to date data sets of this.

That said, as I've commented elsewhere I do agree that Uber is in a good position, but mostly because they don't need full coverage. They need thorough and up to date information, but they need it in very tiny areas to begin with, and as you say they can rely on drivers to cover areas they don't currently know how to deal with. They can apply GPS traces to analyse exactly what areas are most profitable to improve map information for to make self driving viable first, and roll it out area by area.

Re: Google Is Developing Its Own Uber Competitor

#143

Post by benmanns »

ajju wrote: post #82
rottyguy wrote: post #77

Earlier quoted context omitted.

Actually, insurance in driverless cars should be trivial. Accidents should drop to near 0 (over time) due to driver error... at least in theory.

Eventually it might, but that assumes that all cars will be driverless. Even if driverless cars have a 0 error rate, human driven cars will run into them at at least the same rate as they run into static objects. At any rate, this eventual reduction in insurance is why I started by saying "Initially at least". With time, the cost of manufacturing self driving cars will also go down.

That would be fall the human driver's liability and/or uninsured/underinsured coverage, which is going to be much less than comprehensive insurance.

Re: Google Is Developing Its Own Uber Competitor

#144

Post by rottyguy »

pyrocat wrote: post #114
toomuchtodo wrote: post #89

Earlier quoted context omitted.

> That $0.56 number includes depreciation. Electric vehicles would depreciate at a much slower rate than ICE vehicles, as the only components that wear would be moving parts (electric motor, fans for cabin comfort) and any degradation to the battery pack (Tesla has packs at 500K miles that still have 80% capacity).

Don't forget wear and tear from inside the cabin. I don't envy whoever is in charge with cleaning the inside of a self driving car.

Consider the car has the opportunity to get checked and cleaned every time it refuels/recharges, I'd imagine the vehicles would be in decent shape. Until robots take over , I'd suspect that human labor would be involved at these stations.

Re: Google Is Developing Its Own Uber Competitor

#145

Post by xivzgrev »

I'm really surprised. G ventures puts ton of money into uber, uber makes a huge pre order for google cars, uber gets put into google maps.

And now this. I guess google is like "well we got the hard part, self driving cars. Why not just go after this ourselves?"

What is the thing google really wants? Is it more valuable to try to directly monetize all this vehicle traffic, or be the platform every else uses? Google could set up rev share, get access to all the vehicle data (even if they don't use google maps), and set google as default software in the car web experience.

It seems greedy, arrogant, and unaligned with their core ways of making money. They're trying to be vertically integrated player which can work if you control all supply but once Google introduces the tech how long is it going to take for a rival to copy, parents aside?

Re: Google Is Developing Its Own Uber Competitor

#146

Post by kumarm »

threeseed wrote: post #127
softdev12 wrote: post #85

I'm glad that the article highlighted how valuable the mapping information is that Google collects. While Uber probably could still run its current operations without Google map data, there is a much smaller probability that they would be able to run a self-driving car service without a highly accurate mapping product. And to build a mapping product is really difficult and expensive. I was surprised when Google start…

Uber can simply license their maps from various third parties the same way every car company does today for their GPS. Google has added some decent additions to Google Maps but really 95% of what they had they purchased. For example Whereis/Yellow Pages data here in Australia.

There are too many sources and too many processes in world wide Maps data (Apple learnt hard way). I doubt Uber has the capability to have decent maps solution in short term (1-2 years).

Re: Google Is Developing Its Own Uber Competitor

#148

Post by CPLX »

Every single time there is a big discussion thread about Uber on HN about half the commentary is about self-driving cars and how various decisions are being made now based on the inevitability of self driving cars. Every time this happens I always wonder why nobody seems to recognize the obvious fact that self driving cars have nothing to do with any of this stuff and won't make economic sense for a company like Uber in any time horizon that it makes business sense to take into account now.

Re: Google Is Developing Its Own Uber Competitor

#149

Post by rottyguy »

I'm sure this has been discussed on a different (uber/lyft) thread but being a nyc'er and amazed at the prices of taxi medallions through the years, I'm glad to see this bubble finally popping (prices are still insane mind you)!-- now if we can do the same for the cost of higher education... With google coming into the fray, this can only drop further.

http://www.nytimes.com/2015/01/08/upshot/new-york-city-taxi-...

Re: Google Is Developing Its Own Uber Competitor

#150

Post by Aloisius »

potatolicious wrote: post #22
1ris wrote: post #15

Earlier quoted context omitted.

OSM is a way better anyway. This would improve uber.

Not that I don't love OSM - I use it myself for some side projects - but is it really better? Here in NYC I've found several streets mislabeled as one-way in the wrong direction, and not too long ago at that. If these major errors can be found by some guy just fooling around, in one of the biggest cities in the world, how useful is OSM for navigation?

Google Maps tried to route me down a staircase in San Francisco for several months, told me I could turn left on a major street that you have never been able to turn left on and decided my address was not only in the wrong part of town, but in a completely different zip code at one point because someone had created a "business" at my address and set it to be in a completely different place which made it pretty much impossible to order from any service dependent on them like grubhub or eat24hours.

No one has perfect data. At least when I update OSM, it updates relatively quickly as opposed to Google Maps which took months to get updated after reporting the errors.

Post reply on HN