> Do you want to beat Uber? Pay the drivers better.
In a race towards a self-driving fleet, this does not appear to be the winning strategy.
In a race towards a self-driving fleet, this does not appear to be the winning strategy.
Which is hilarious. Tesla already has a semi-autonomous fleet of tens of thousands of cars, improving daily.
Best of luck Uber.
https://en.wikipedia.org/wiki/Autonomous_car#Definition
"Level 4: The vehicle performs all safety-critical functions for the entire trip, with the driver not expected to control the vehicle at any time. As this vehicle would control all functions from start to stop, including all parking functions, it could include unoccupied cars."
I don't even believe laser scanners Google is using are required anymore:
http://gizmodo.com/mit-figured-out-how-to-make-cheap-3d-scan...
Not yet.
Based on what I know of Google's efforts, I think your three-year horizon for Tesla is, well, optimistic. As many miles as Google has logged, they're not testing at highway speeds yet -- I don't think their driverless prototypes (as opposed to the modified Lexus SUVs) are even capable of highway speeds. Also, AFAIK, they're only testing cars in Mountain View and Austin, which means there's a lot of weather conditions -- and to some degree terrain -- that they haven't had to accommodate. The most recently-given figures they've supplied that I can find is that they expect to be able to commercialize their technology between 2017 and 2020 -- and they've been working on this project since 2009. Tesla'a engineering is good, to be sure, but I would be really surprised if they beat Google to market in this area.
Now, once somebody really does start offering these for sale to the general public (or just to taxi/rideshare companies or etc.), the answer to your question shifts from "not yet" to "yep."
Tesla's autopilot functions at up to 90mph.
> I think your three-year horizon for Tesla is, well, optimistic.
http://www.wired.com/2015/10/obviously-drivers-are-already-a...
"Four days later, Reese, Mastracci, and Roy piled into a Model S and took off. They covered 2,994 miles at an average speed of 51.8 mph, a figure that includes the time spent plugged into Supercharger stations along the way. They had autopilot mode engaged 96 percent of the time, Reese says, using it at speeds around 90 mph. It eased the burden on the team, a big deal when you’re in a car for 57 hours straight."
96 percent of the time, at speeds up to 90mph, across 2,994 miles
To tackle all the other problems to truly have a self-driving cars is going to take much longer. Not at least since there will be decreasing incentive to solve these much harder problems in a cost effective way.
4% of the time there's inclement weather, ambiguously signed construction, heavy traffic in snow that occludes the entirety of the road, cars stopped in their lane, detours into the lane normally dedicated to oncoming traffic, and all sorts of other low-frequency events that are much harder to solve than keeping a car moving forward on an interstate.
I'd argue that 96% autonomous driving time isn't even enough to say we're 50% of the way through the challenges of creating an autonomous vehicle. Think twice before accusing me of not understanding percentages.
https://static.googleusercontent.com/media/www.google.com/en...
(I clearly remember a video demonstrating one of the employees pulling on to the freeway and then giving the car control, I found the above while doing a quick search for that)
The second slide here says 300,000 freeway miles circa 2012:
https://www.google.com/selfdrivingcar/where/
and the next couple imply that they then moved on to the harder problem of streets.
1) When you scale from individual autonomous devices to vast networks of them you run into emergent phenomena that you simply could not predict ahead of time.
2) You have to spend enormous engineering effort protecting incumbents. E.g. there has been a ton of work on figuring out how to protect wireless microphones used by churches and stadiums. They are intermittent and have a "dumb receiver" (that never transmits), making them very difficult to handle using a "listen before talk" approach. And that's for fricking church PAs. Imagine the push-back when your technological edge case runs over a kid.
3) You can't deploy commercially until you change the regulatory regime. That can take years to decades if you're asking for seismic shifts in policy.
I would predict another 5-10 years before the technology works correctly at scale, rather than under controlled conditions. Then maybe another 5-10 years after that for limited regulatory approval, and maybe 10-25 years for widespread regulatory approval. I don't expect to be able to catch a self-driving cab in New York or Chicago before 2040.
[1] The end goal was technologies that could obsolete fixed spectrum allocation by letting networks of cognitive radios sort out for themselves what frequencies to use. There are various narrower applications, such as technologies that can take advantage of TV "whitespaces."
Your comment: https://news.ycombinator.com/item?id=10673357
Wha?
You are lucky to be able to consider such a thing seriously!
Why don't you bet something more reasonable, like $10?
The second part of that isn't really something you add to casual repartee.
It has to be nation wide in at least 45 states, it also needs to be a comparable price to what Uber is now ( so like 30$ or less for a 10 mile trip)
Ride-hailing services are a saturated market with major network effects (the more drivers on a service, the faster they can get to you) and thus, expensive barriers to entry.
It just seems like it makes more business sense for Tesla to sell/license as many self-driving cars as they can to any and all services that want them, and concentrate on their core business, rather than branch out into an entirely new market and write a whole new app from scratch and market it.
Can someone make the opposite case?
Not so much. Drivers frequently run Uber and Lyft simultaneously. There is no "stickiness".
> It just seems like it makes more business sense for Tesla to sell/license as many self-driving cars as they can to any and all services that want them, and concentrate on their core business, rather than branch out into an entirely new market and write a whole new app from scratch and market it.
You'd think so, wouldn't you? But apparently that's the direction Elon is going: http://www.techinsider.io/elon-musk-hints-tesla-may-enter-ri...
> write a whole new app from scratch and market it.
I think its humorous this is thought of as expensive. How much does it cost to build an iOS/Android app and run the infrastructure to match riders to drivers? $3-5/million/year? Tesla has hundreds of millions of dollars of credit available to it. And you know, the cars receiving software updates constantly.
I meant from the perspective of users and providers, not drivers (which are irrelevant in the case of self-driving cars...). You have to get all these users to download another app when their existing one works fine. That's money that Tesla could be further investing in their core specialty rather.
>You'd think so, wouldn't you? But apparently that's the direction Elon is going
Yes, which is why I'm asking; the fact that someone is in fact doing it doesn't resolve this mystery.
>I think its humorous this is thought of as expensive. How much does it cost to build an iOS/Android app and run the infrastructure to match riders to drivers? $3-5/million/year? Tesla has hundreds of millions of dollars of credit available to it. And you know, the cars receiving software updates constantly.
Right, but they haven't trodden the path of scaling up a ride-hailing app and made the numerous optimizations that Uber/Lyft have made based on the unique problems of that domain.
That isn't what network effects are. That same argument applies to anyone competing against an incumbent.
Double-sided networks like Uber are hard to gain critical mass. Uber did it, which is extremely impressive, however, whoever comes up with a self-driving car will eliminate one side of that network, making it substantially easy to compete.
I personally don't see how Uber's infrastructure minus the driver network is that important.
Given how Apple approaches everything, it's logical that they're going to create an end-to-end experience, meaning they make the tech, they make the cars, and they make the equivalent of Uber.
I'm not sure what Google's move is once they have commercially-viable self-driving tech. Would they license it to another automaker or group of automakers? Would their tech plug into whatever request app they make?
I think the strategy for Uber is to connect with a manufacturer like Tesla, who seem to be advancing competitively with Google and Apple on self-driving tech. Tesla has less access to cash and debt than Apple. Still, I see Uber's value being far less to Tesla than Tesla's value to Uber.