Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Locked thread
Condiv
May 7, 2008

Sorry to undo the effort of paying a domestic abuser $10 to own this poster, but I am going to lose my dang mind if I keep seeing multiple posters who appear to be Baloogan.

With love,
a mod


Cicero posted:

The first general purpose self-driving cars will undoubtedly still have regular controls so that humans can take over when necessary for conditions they can't handle like really bad weather or signs that they can't understand (e.g. parking rules). You don't need to hit 100% of all use cases for self-driving cars to be useful, having them drive like 90% of the time would still be a huge benefit.

i doubt it will actually be a huge benefit. the 90% ok rate will probably make drivers less vigilant during the times they expect the car to operate normally, and therefore more likely to not be prepared for their responsibilities when the car kicks control back over to them.

and the current technology doesn't seem to have safety benefits over normal cars. the current crop of goog cars have been in twice the average number of accidents than your typical human. while they haven't had an "at fault" wreck yet, it's almost certain its high wreck rate is due to the erratic behavior mentioned by asdf32. even if we replaced all the norma cars with self-driving ones traffic accidents won't stop. two self-driving cars nearly wrecked each other a couple of months ago (goog's and another companies, one cut the other off). when you consider the current rarity of self-driving cars on the road, we can assume this is one of the few interactions between two self-driving cars, and it does not paint a pretty picture of the current tech's safety.

Adbot
ADBOT LOVES YOU

The Larch
Jan 14, 2015

by FactsAreUseless
What I've always wondered about self-driving cars is if they brake for small animals. Or children, for that matter.

Condiv
May 7, 2008

Sorry to undo the effort of paying a domestic abuser $10 to own this poster, but I am going to lose my dang mind if I keep seeing multiple posters who appear to be Baloogan.

With love,
a mod


uber lost it's appeal on employee classification at the california unemployment insurance board

http://www.scribd.com/doc/279230988/EDD-Unemployment-Appeal-vs-Uber-Case-No-5371509

previously, the board had ruled that a specific employee (a driver) had a right to unemployment after uber let her go, as they found she met the definition of an employee. uber appealed (iirc, not more than a month ago).

Condiv
May 7, 2008

Sorry to undo the effort of paying a domestic abuser $10 to own this poster, but I am going to lose my dang mind if I keep seeing multiple posters who appear to be Baloogan.

With love,
a mod


The Larch posted:

What I've always wondered about self-driving cars is if they brake for small animals. Or children, for that matter.

probably, considering they're stopping erratically when encountering floating non-harmful debris. tbht though this is one of the weaknesses of the machine learning approach they're trying here. machine learning is a very powerful tool, and can approximate many complex functions, but the issue is it is notoriously difficult to debug their kernels. unlike with typical software development, you cannot really say with any certainty that a system is bug free or correct, which is not a huge problem in many ML applications. with safety-critical systems though, you kind of want to be able to prove the correctness of the software controlling the vehicle.

computer parts
Nov 18, 2010

PLEASE CLAP

Cicero posted:

The first general purpose self-driving cars will undoubtedly still have regular controls so that humans can take over when necessary for conditions they can't handle like really bad weather or signs that they can't understand (e.g. parking rules). You don't need to hit 100% of all use cases for self-driving cars to be useful, having them drive like 90% of the time would still be a huge benefit.

You might be able to get 90% of miles driven automated (which are not to be confused with use cases), but only because those are on highways in the middle of nowhere.

Also 90% is piss poor, not even 2 sigma.

Shifty Pony
Dec 28, 2004

Up ta somethin'


The Larch posted:

What I've always wondered about self-driving cars is if they brake for small animals. Or children, for that matter.

Absolutely yes. I saw a squirrel run in front of one and the car slammed to a halt. I think it had kicked control over to the driver too because it took five seconds or so to restart.

Having them drive around the neighborhood I can definitely see why they have been attracting a lot of not at fault accidents. They drive slowly and brake erratically when there are no apparent hazards around. Had someone been following it when the squirrel happened they could have easily rear ended it.

a_gelatinous_cube
Feb 13, 2005

ANIME AKBAR posted:

Of the three people I know who work/did work at google X, two of them said that a self driving car capable of operating in real life conditions is still decades away at least. They don't even have a clue how driving in snow is going to work. The third one never worked on the car.

The amount of people trumpeting that self-driving cars are just around the corner the last few years has seems like a techculture fueled insane fantasy to me. If there were self-driving cars available for purchase and they started showing up anywhere outside whatever tiny-sized fleet that they have in San Fransisco, yeah maybe you could think that they would be a major thing on the road in 20-30 years, but we aren't even there yet.

Also, just curious, but how do self-driving cars handle situations where you have to drive aggressively, like making a left-hand turn onto a heavily trafficked street with no light?

a_gelatinous_cube fucked around with this message at 13:03 on Sep 9, 2015

Necc0
Jun 30, 2005

by exmarx
Broken Cake

Zyklon B Zombie posted:

The amount of people trumpeting that self-driving cars are just around the corner the last few years has seems like a techculture fueled insane fantasy to me. If there were self-driving cars available for purchase and they started showing up anywhere outside whatever tiny-sized fleet that they have in San Fransisco, yeah maybe you could think that they would be a major thing on the road in 20-30 years, but we aren't even there yet.

Also, just curious, but how do self-driving cars handle situations where you have to drive aggressively, like making a left-hand turn onto a heavily trafficked street with no light?

http://www.bloomberg.com/news/articles/2015-05-14/daimler-s-freightliner-tests-self-driving-truck-in-nevada

only legal on highways in the nevada desert

kitten emergency
Jan 13, 2008

get meow this wack-ass crystal prison

Zyklon B Zombie posted:

Also, just curious, but how do self-driving cars handle situations where you have to drive aggressively, like making a left-hand turn onto a heavily trafficked street with no light?

they don't/poorly

Anubis
Oct 9, 2003

It's hard to keep sand out of ears this big.
Fun Shoe

Condiv posted:

and the current technology doesn't seem to have safety benefits over normal cars. the current crop of goog cars have been in twice the average number of accidents than your typical human. while they haven't had an "at fault" wreck yet, it's almost certain its high wreck rate is due to the erratic behavior mentioned by asdf32.

Sorry, but the hobby statistician in me can't help but feel that there's a flaw in comparing accident rates for a fleet of cars driving around Seattle and LA to national average accident rates and trying to pull useful info from that comparison. I'd expect a much higher accident rate in LA or New York then if they were doing the majority of their testing in Arizona. Not to say that automated cars are commercial ready just around the corner but they are likely further along then you would expect if you're only comparing to those national numbers.

kitten emergency
Jan 13, 2008

get meow this wack-ass crystal prison
it should be mildly telling that the people who ought to be really, really interested in autonomous vehicles - amazon, wal-mart, fedex, ups, anyone who uses long-haul trucking - aren't really making a lot of noise about it, even though freight is the obvious initial entry for assistive driving. i think we'll definitely see a proliferation of freight trucks in the US that are autonomous on interstate highways in good conditions within the decade.

doesn't really help uber though. they're still hosed in pretty much any estimation because they don't make enough money to sustain their business model without VC.

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

uncurable mlady posted:

it should be mildly telling that the people who ought to be really, really interested in autonomous vehicles - amazon, wal-mart, fedex, ups, anyone who uses long-haul trucking - aren't really making a lot of noise about it, even though freight is the obvious initial entry for assistive driving. i think we'll definitely see a proliferation of freight trucks in the US that are autonomous on interstate highways in good conditions within the decade.

doesn't really help uber though. they're still hosed in pretty much any estimation because they don't make enough money to sustain their business model without VC.

It's a good point so I don't know why you conclude that autonomous trucking will happen within the decade (<5 years).

Again, there is a huge gulf between automation in cars (lane drift alarms, adaptive cruise control, automatic breaking) and self driving cars where you would actually remove the driver. One is already here and will be steadily improving and increasing and the other is multiple decades away and/or may never exist in the form people are predicting today.


Uber isn't hosed, they just may not change the world.

Necc0
Jun 30, 2005

by exmarx
Broken Cake

asdf32 posted:

Uber isn't hosed, they just may not change the world.

No that's exactly why they're hosed: their current valuation mandates that they revolutionize the world. Any less and they crash & burn

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

Necc0 posted:

No that's exactly why they're hosed: their current valuation mandates that they revolutionize the world. Any less and they crash & burn

Specifically it means their investors are hosed which is different. Consider groupon - insane hype/bubble/crash, but now they're earning a profit as a cupon/deal site because cupon/deal sites are viable businesses. The same may happen for uber but uber but uber is probably better in every way.

Necc0
Jun 30, 2005

by exmarx
Broken Cake

asdf32 posted:

Specifically it means their investors are hosed which is different. Consider groupon - insane hype/bubble/crash, but now they're earning a profit as a cupon/deal site because cupon/deal sites are viable businesses. The same may happen for uber but uber but uber is probably better in every way.

So when you say 'Uber isn't hosed' you meant their existence as a recognizable brand, not as a corporate entity geared towards creating an ROI

Main Paineframe
Oct 27, 2010

Radbot posted:

Which courts have "shut down" the Uber 1099-dependent business model? I know there was some non-binding decision in CA awhile back.

I can't find anyone besides Uber claiming that a ruling from the California Labor Commission is "non-binding", and it's not quite clear what Uber means by "non-binding", since the considerable reimbursement and penalty that Uber has been ordered to pay are most definitely binding. I think what they actually mean is that it doesn't automatically extend to all of their drivers, and only covers that one individual driver's case; however, it's still not good for Uber, since any other Uber driver that brings a similar case would likely end up with a similar result, and California judges will almost certainly pay attention to the Labor Commissioner's ruling.

Aside from that, the case to watch right now is O'Connor v. Uber. It's been working its way through the system for a very long time, and recently the judge decided to allow it class-action status, meaning that it now represents tens of thousands of Uber drivers, rather than just the three original plaintiffs. Uber's contracts had arbitration and "no class action" clauses, but those were found to be invalid and unenforceable in O'Connor, as well as a few other cases like Gilette v. Uber. Overall, Uber has not been faring well in California courtrooms, although progress has been very slow - O'Connor dates back to 2013 or earlier, won't actually begin the real trial until 2016, and it's absolutely certain that Uber will appeal the likely-unfavorable decision (the judges have not been friendly toward Uber) all the way up the system.

coma
Oct 21, 2010

asdf32 posted:

Specifically it means their investors are hosed which is different. Consider groupon - insane hype/bubble/crash, but now they're earning a profit as a cupon/deal site because cupon/deal sites are viable businesses. The same may happen for uber but uber but uber is probably better in every way.

What about the other scenario where actual taxi services start using a big-tent Uberlike app to drive business, that's one of the reasons Didi Kuaidi is destroying Uber in China: they cover taxis and rideshares and threw the 'disruption' playbook in the trash. When you strip away the Jetsons futurology away from Uber the mobile ease-of-use and scam-prevention is basically what you're left with, and anybody can do that with taxis and some sort of regulated ridesharing system while junking the risky '1099s and driverless cars to the moon' component that is core to Uber.

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

coma posted:

What about the other scenario where actual taxi services start using a big-tent Uberlike app to drive business, that's one of the reasons Didi Kuaidi is destroying Uber in China: they cover taxis and rideshares and threw the 'disruption' playbook in the trash. When you strip away the Jetsons futurology away from Uber the mobile ease-of-use and scam-prevention is basically what you're left with, and anybody can do that with taxis and some sort of regulated ridesharing system while junking the risky '1099s and driverless cars to the moon' component that is core to Uber.

The core of uber is that when you click an app a car reliably shows up. Uber can continue to execute this under a pretty wide variety of laws and regulations. And despite the simplicity, taxi companies as structured are actually poor at competing with this for a variety of reasons.

Also, competition doesn't spell doom for a business.

coma
Oct 21, 2010

asdf32 posted:

The core of uber is that when you click an app a car reliably shows up. Uber can continue to execute this under a pretty wide variety of laws and regulations. And despite the simplicity, taxi companies as structured are actually poor at competing with this for a variety of reasons.

Also, competition doesn't spell doom for a business.

That's what I was saying inarticulately though, when your real-core is so easy to replicate as opposed to your high-valuation fantasy-core, all you really have is the brand-recognition from first-mover benefit. And even though taxi-services are terrible and stuck in the jurassic era, they're having the screws put to them in such a way that it's not inconceivable that people on the taxi side and some enterprising person from the startup side could look at Didi Kuaidi and take advantage of Uber blowback.

FCKGW
May 21, 2006

This what it's like to stay in the Bay Area's cheapest Airbnbs

quote:

The last place I stayed was maybe the most unusual of all. In Berkeley, there's an enormous house — 4 apartments in one building, really, with a total of 11 bedrooms. It has a large backyard, protected by particularly high fences, that is filled with tents — six enormous tents, pitched side-to-side, with Christmas lights strung between them. These tents are always up, and always for rent — the cheapest Airbnb rental in Berkeley. It looks like a tiny, ramshackle village.

The owner of the house — "compound" might be a more accurate term — has turned the entire place into an Airbnb factory, housing about 30 people a night between dorm-style living in the house and the tents in the backyard. He says he's trying to maintain a sustainable business, and provide affordable living situations in an area where housing prices are on the up. The rent is cheap — the tents start at $25 a night — but he seems to work out a deal with those that want to stay longer. I meet several people who are staying for weeks or months at a time, both inside the house and in the tents. Many are staying there while attending the University of California Berkeley for a set period of time, although there are also tourists and people just looking for a cheap place to spend a night.

The owner says he has had run-ins with the Berkeley Rent Board and doesn't want them to know about the scope of the operation, although he insists it is all entirely legal. He's proud of his operation, though, and despite any potential ramifications, wants to get the word out.

When I first arrive, all of the tents are occupied.

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

coma posted:

That's what I was saying inarticulately though, when your real-core is so easy to replicate as opposed to your high-valuation fantasy-core, all you really have is the brand-recognition from first-mover benefit. And even though taxi-services are terrible and stuck in the jurassic era, they're having the screws put to them in such a way that it's not inconceivable that people on the taxi side and some enterprising person from the startup side could look at Didi Kuaidi and take advantage of Uber blowback.

Brand recognition and first mover benefits arn't nothing though and although simple the barriers to entry are'nt insignificant.

If uber's problem is that it will successfully transform the aptly described jurassic taxi industry into effective competition then the disruption hype will have come true. The reality is probably less grand - both uber and taxis will continue to coexist in semi-competition.

Boot and Rally
Apr 21, 2006

8===D
Nap Ghost
I have never had a problem with Curb (formerly Taxi Magic), which does what Uber does but with actual cabs. As Taxi Magic it pre-dates Uber. From the client perspective it did the same thing (except you had to swipe your card, but no longer). It was weird to hear people describe Uber as this brand new idea for things "like cabs" and watch their face go blank when you tell them that already exists, in entirety, for actual cabs.

Main Paineframe
Oct 27, 2010

coma posted:

What about the other scenario where actual taxi services start using a big-tent Uberlike app to drive business, that's one of the reasons Didi Kuaidi is destroying Uber in China: they cover taxis and rideshares and threw the 'disruption' playbook in the trash. When you strip away the Jetsons futurology away from Uber the mobile ease-of-use and scam-prevention is basically what you're left with, and anybody can do that with taxis and some sort of regulated ridesharing system while junking the risky '1099s and driverless cars to the moon' component that is core to Uber.

Taxi services are already using mobile apps, and have been for some time. Uber's advantages are its lower pricing, the superior degree of control it exercises over its drivers, its willingness to offer refunds or free rides for practically any reason, its willingness to screw over the driver for the sake of the customer, and its ability to avoid most of the costs taxi companies have to deal with by making drivers use their own equipment and simply ignoring any regulations that might inconvenience them.

hobbesmaster
Jan 28, 2008

Main Paineframe posted:

Taxi services are already using mobile apps, and have been for some time. Uber's advantages are its lower pricing, the superior degree of control it exercises over its drivers, its willingness to offer refunds or free rides for practically any reason, its willingness to screw over the driver for the sake of the customer, and its ability to avoid most of the costs taxi companies have to deal with by making drivers use their own equipment and simply ignoring any regulations that might inconvenience them.

The problem is the "apps" part. You land in a new town, uber will work but what app do the local taxis use? Who knows, my uber is already here.

Shifty Pony
Dec 28, 2004

Up ta somethin'


asdf32 posted:

The core of uber is that when you click an app a car reliably shows up. Uber can continue to execute this under a pretty wide variety of laws and regulations. And despite the simplicity, taxi companies as structured are actually poor at competing with this for a variety of reasons.

Also, competition doesn't spell doom for a business.

Taxi companies have trouble with it because you cannot get that level of service with true independent contractors at taxi prices. If you are willing to pay more for something like a limo or town car service it will show up exactly where you want exactly when you want.

Uber is essentially in the business of providing three or four-star level quality at construction site food truck prices. It presently manages to do so through a combination of illegal misclassification of employees and massive driver churn. The latter is only possible with their VC money because they can pour money into driver recruiting incentives, ads, and the costs of things like background checks and vehicle inspections without needing to worry about ROI.

The MUMPSorceress
Jan 6, 2012


^SHTPSTS

Gary’s Answer

Ceiling fan posted:

Who the gently caress knows if there is a tech or any kind of bubble? All I know is that I check out individual companies and invest in the ones that can make money doing what they do and are run well.

Also, it's really hard to find porn companies on stock exchanges. Seriously, they're the most profitable companies on the net, and that's while anyone can find as much free porn as they want. I've been making a killing on boring poo poo like medical technology, consumer technology, semiconductor production, ect. And I can't find a nice porn production stock to save my life.

I can't google the link to the article about this for you right now because I'm at work, but these freebie sites are actually kinda brilliant. The guy who founded one of them (I think pornhub or something) made enough money that he bought up all the other 'tube sites too, and runs them off the same backend. He made so much money on that, that he actually went and bought Brazzers and NaughtyAmerica. This means that he now controls what clips get put up for free and can use that to drive people to the paysites he now owns. This loop makes him enough money that if competing tube sites pop up he just buys them. This one man is now in control of most of the internet's porn ecosystem. I'd imagine he's actively trying to buy the other big producers too, or at least sell them services for control of their content on his tube sites. Back in the day BangBros used to upload lower-quality copies of their own new releases to PirateBay because it drove more traffic to their site than trying to fight the pirates, so it's clearly a good business model.

All of this is to say (besides relating that hilarious situation), that the guy who owns all this poo poo basically has no need for outside capital whatsoever. He has no reason to offer stock because he basically controls it all now.

FCKGW
May 21, 2006

http://www.cnbc.com/id/45989405
http://nymag.com/news/features/70985/index4.html

You're thinking of Mindgeek, formerly known as Manwin. They own something like 80% of all the tube sites (basically the top 10) and have gobbled up several production companies. They are the gatekeepers to online smut these days.

If anyone could do an IPO it would be those guys.

FCKGW fucked around with this message at 16:47 on Sep 9, 2015

Cicero
Dec 17, 2003

Jumpjet, melta, jumpjet. Repeat for ten minutes or until victory is assured.
It's amazing how fundamentally some left-wingers misunderstand technology. "This prototype technology has problems, and thus will continue to have problems indefinitely".

Condiv posted:

i doubt it will actually be a huge benefit. the 90% ok rate will probably make drivers less vigilant during the times they expect the car to operate normally, and therefore more likely to not be prepared for their responsibilities when the car kicks control back over to them.
This part I actually am concerned about, but I think it won't be, "we're about to crash, here take over" because obviously that wouldn't work. I'm talking more about, "the weather is turning/about to turn bad, I'll pull over to the side of the road."

quote:

and the current technology doesn't seem to have safety benefits over normal cars. the current crop of goog cars have been in twice the average number of accidents than your typical human.
This isn't necessarily true. Several of the accidents have been minor ones that regular people might not report.

quote:

while they haven't had an "at fault" wreck yet, it's almost certain its high wreck rate is due to the erratic behavior mentioned by asdf32.
Conjecture. If the crash rate is higher than normal, it's more likely the novelty of the cars with their big ol' LIDARs that's distracting at the moment.

quote:

even if we replaced all the norma cars with self-driving ones traffic accidents won't stop.
What, if they cut crashes in half would that be a defeat? They don't have to stop, they just have to go down.

quote:

two self-driving cars nearly wrecked each other a couple of months ago (goog's and another companies, one cut the other off). when you consider the current rarity of self-driving cars on the road, we can assume this is one of the few interactions between two self-driving cars, and it does not paint a pretty picture of the current tech's safety.
The tech is rapidly improving, but there's a reason it hasn't rolled out commercially yet. Also, they didn't 'nearly wreck each other', that's a ridiculous exaggeration:

quote:

A self-driving Audi owned by Delphi Automotive took “appropriate action” to avoid one of Google’s self-driving Lexus cars after it cut it off on a Californian road in a rare meeting of driverless vehicles.

John Absmeier, who was travelling in his company’s car at the time, said the Delphi Audi was forced to abort its lane change in the incident, which happened earlier this week.

The incident was initially reported as a “close call” by Reuters, which cited Absmeier’s comments, but both self-driving car companies involved have since been at pains to state the incident was being used as an example of the cars operating safely.

A Delphi spokeswoman, Kristin Kinley, told Ars Technica: “The story was taken completely out of context when describing a type of complex driving scenario that can occur in the real world. Our expert provided an example of a lane change scenario that our car recently experienced which, coincidentally, was with one of the Google cars also on the road at that time.”

Delphi later released another statement saying: “It was an anecdote of an interaction, not a ‘near miss’. Reuters completely misrepresented the facts.”

A source at Google also confirmed there was no near collision and said no one was at fault.
http://www.theguardian.com/technology/2015/jun/26/google-delphi-two-self-driving-cars-near-miss

Oh no, it had to abort a lane change, something most people do just about every day that they're driving! Seriously, this is a completely common and normal driving interaction that happens to cars constantly, the only unusual thing here is that the cars involved were computer-driven.

computer parts posted:

You might be able to get 90% of miles driven automated (which are not to be confused with use cases), but only because those are on highways in the middle of nowhere.

Also 90% is piss poor, not even 2 sigma.
I was thinking 90% including both highway and city driving, with the major exclusion being bad weather (and the early cars probably wouldn't be practical for very snowy climes). Parking anywhere that isn't a bog standard parking lot/street parking would probably also require manual intervention at least to tell the car where it's allowed to park.

uncurable mlady posted:

it should be mildly telling that the people who ought to be really, really interested in autonomous vehicles - amazon, wal-mart, fedex, ups, anyone who uses long-haul trucking - aren't really making a lot of noise about it, even though freight is the obvious initial entry for assistive driving. i think we'll definitely see a proliferation of freight trucks in the US that are autonomous on interstate highways in good conditions within the decade.
http://money.cnn.com/2015/05/06/autos/self-driving-truck/
http://www.dailymail.co.uk/sciencet...ads-months.html

Granted the German one set for just a couple years away is only semi-autonomous, for highway driving.

asdf32 posted:

In terms of safety, perhaps (if they're more reliable than humans when they are operating). But the majority of the benefits and the largest impact comes from actually removing the driver from the car. The taxi cab/ups business model doesn't change until that happens.
I bike a lot so the safety aspect is actually the most enticing part to me (although it'll be a long time before there's enough market share to really make a dent). Completely removing the driver from the car will definitely take longer than having something that can drive by itself most of the time.

boner confessor
Apr 25, 2013

by R. Guyovich

Cicero posted:

It's amazing how fundamentally some left-wingers misunderstand technology. "This prototype technology has problems, and thus will continue to have problems indefinitely".

it's amazing how fundamentally some right wingers misunderstand arguments. "This prototype technology won't be viable for a long time, but it will surely save the enormous packet of hype and exploitation branded Uber!"

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

Shifty Pony posted:

Taxi companies have trouble with it because you cannot get that level of service with true independent contractors at taxi prices. If you are willing to pay more for something like a limo or town car service it will show up exactly where you want exactly when you want.

Uber is essentially in the business of providing three or four-star level quality at construction site food truck prices. It presently manages to do so through a combination of illegal misclassification of employees and massive driver churn. The latter is only possible with their VC money because they can pour money into driver recruiting incentives, ads, and the costs of things like background checks and vehicle inspections without needing to worry about ROI.

The worst case for uber is that they're a national brand black car service with an app. It might not be a revolution, it is a viable business. But there's room between where they are now in terms of price/service and existing taxis which they compete with, so they'll probably be something more than that.

hobbesmaster
Jan 28, 2008

Cicero posted:

It's amazing how fundamentally some left-wingers misunderstand technology. "This prototype technology has problems, and thus will continue to have problems indefinitely".

You realize that airplanes, which have had autopilots for 80 years, are nowhere near this level of automation right.

uninterrupted
Jun 20, 2011
It is critical that we remember the real victims of the Holocaust: the Soviets.

Claiming that the Holocaust was about Jews is Zionism. Biden is doing it to provide cover to Israel!

Cicero posted:

It's amazing how fundamentally some left-wingers misunderstand technology. "This prototype technology has problems, and thus will continue to have problems indefinitely".

This part I actually am concerned about, but I think it won't be, "we're about to crash, here take over" because obviously that wouldn't work. I'm talking more about, "the weather is turning/about to turn bad, I'll pull over to the side of the road."

This isn't necessarily true. Several of the accidents have been minor ones that regular people might not report.

Conjecture. If the crash rate is higher than normal, it's more likely the novelty of the cars with their big ol' LIDARs that's distracting at the moment.

What, if they cut crashes in half would that be a defeat? They don't have to stop, they just have to go down.

The tech is rapidly improving, but there's a reason it hasn't rolled out commercially yet. Also, they didn't 'nearly wreck each other', that's a ridiculous exaggeration:

http://www.theguardian.com/technology/2015/jun/26/google-delphi-two-self-driving-cars-near-miss

Oh no, it had to abort a lane change, something most people do just about every day that they're driving! Seriously, this is a completely common and normal driving interaction that happens to cars constantly, the only unusual thing here is that the cars involved were computer-driven.

I was thinking 90% including both highway and city driving, with the major exclusion being bad weather (and the early cars probably wouldn't be practical for very snowy climes). Parking anywhere that isn't a bog standard parking lot/street parking would probably also require manual intervention at least to tell the car where it's allowed to park.

http://money.cnn.com/2015/05/06/autos/self-driving-truck/
http://www.dailymail.co.uk/sciencet...ads-months.html

Granted the German one set for just a couple years away is only semi-autonomous, for highway driving.

I bike a lot so the safety aspect is actually the most enticing part to me (although it'll be a long time before there's enough market share to really make a dent). Completely removing the driver from the car will definitely take longer than having something that can drive by itself most of the time.

Staying in one lane and slowing down/stopping on a highway is a significantly simpler problem than city driving, which is Uber's bread and butter (well, aside from VC money).

In the short term vehicle automation will make the trucking industry much safer. It will not impact Uber's operations for another decade, if Uber even survives that long. Not only would autonomous cars need to be 100% perfected, but there would need to be a legal mechanism for driverless cars to be used in a commercial setting. EVEN THEN it's questionable that Uber would want to deploy/maintain/store fleets of autonomous cars across the world and take sole liability for any accidents, as opposed to their current scheme of having the drivers handle mainenance/repair and commit insurance fraud to subsidize Uber's prices.

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

Cicero posted:

It's amazing how fundamentally some left-wingers misunderstand technology. "This prototype technology has problems, and thus will continue to have problems indefinitely".

This part I actually am concerned about, but I think it won't be, "we're about to crash, here take over" because obviously that wouldn't work. I'm talking more about, "the weather is turning/about to turn bad, I'll pull over to the side of the road."

This isn't necessarily true. Several of the accidents have been minor ones that regular people might not report.

Conjecture. If the crash rate is higher than normal, it's more likely the novelty of the cars with their big ol' LIDARs that's distracting at the moment.

What, if they cut crashes in half would that be a defeat? They don't have to stop, they just have to go down.

The tech is rapidly improving, but there's a reason it hasn't rolled out commercially yet. Also, they didn't 'nearly wreck each other', that's a ridiculous exaggeration:

http://www.theguardian.com/technology/2015/jun/26/google-delphi-two-self-driving-cars-near-miss

Oh no, it had to abort a lane change, something most people do just about every day that they're driving! Seriously, this is a completely common and normal driving interaction that happens to cars constantly, the only unusual thing here is that the cars involved were computer-driven.

I was thinking 90% including both highway and city driving, with the major exclusion being bad weather (and the early cars probably wouldn't be practical for very snowy climes). Parking anywhere that isn't a bog standard parking lot/street parking would probably also require manual intervention at least to tell the car where it's allowed to park.

http://money.cnn.com/2015/05/06/autos/self-driving-truck/
http://www.dailymail.co.uk/sciencet...ads-months.html

Granted the German one set for just a couple years away is only semi-autonomous, for highway driving.

I bike a lot so the safety aspect is actually the most enticing part to me (although it'll be a long time before there's enough market share to really make a dent). Completely removing the driver from the car will definitely take longer than having something that can drive by itself most of the time.

Machine learning is a problem people have been tripping over for decades though. As somone who does engineering development the pattern is obvious: 90% of the functionality takes 10% of the time.

Development is moving rapidly because it's solving all the simple problems that cover most cases. But the problems that haven't been solved get exponentially harder. In the case of cars failure is life or death and regulators and consumers arn't actually going to hand over the keys until 100% of the kinks have been worked out. That's multiple decades out at the least.

Cicero
Dec 17, 2003

Jumpjet, melta, jumpjet. Repeat for ten minutes or until victory is assured.

Popular Thug Drink posted:

it's amazing how fundamentally some right wingers misunderstand arguments. "This prototype technology won't be viable for a long time, but it will surely save the enormous packet of hype and exploitation branded Uber!"
I never said or implied this. For one, I think Uber is viable even without self-driving cars. Switching to employees instead of contractors would mean raising prices and losing some market share, but I don't see why they would collapse entirely (well, if that happened after an IPO, the stock price might collapse). Secondly, self-driving cars that can run without a human in them at all 'just in case' are obviously further away than self-driving cars that can generally drive themselves but need a person around for some cases.

I think Uber has a good chance of figuring out self-driving taxis for certain limited use cases (e.g. extremely dense areas where you can have emergency drivers on standby all the time) within the next decade, but ones that can be deployed everywhere is unlikely.

asdf32 posted:

The worst case for uber is that they're a national brand black car service with an app. It might not be a revolution, it is a viable business. But there's room between where they are now in terms of price/service and existing taxis which they compete with, so they'll probably be something more than that.
I agree with this.

Cicero
Dec 17, 2003

Jumpjet, melta, jumpjet. Repeat for ten minutes or until victory is assured.

hobbesmaster posted:

You realize that airplanes, which have had autopilots for 80 years, are nowhere near this level of automation right.
We've already seen huge advancement in the last decade. The first DARPA grand challenge had cars that couldn't even really navigate open areas (albeit ones with lots of rocks and inclines), and a decade later we have cars that can handle highway driving themselves pretty well and city driving decently most of the time.

asdf32 posted:

Machine learning is a problem people have been tripping over for decades though. As somone who does engineering development the pattern is obvious: 90% of the functionality takes 10% of the time.

Development is moving rapidly because it's solving all the simple problems that cover most cases. But the problems that haven't been solved get exponentially harder. In the case of cars failure is life or death and regulators and consumers arn't actually going to hand over the keys until 100% of the kinks have been worked out. That's multiple decades out at the least.
I don't really think this is true. The bar isn't "can handle all use cases perfectly without fail", it's more like "substantially better than a human driver for almost everything". I don't think that is multiple decades away.

Shifty Pony
Dec 28, 2004

Up ta somethin'


Cicero posted:

Conjecture. If the crash rate is higher than normal, it's more likely the novelty of the cars with their big ol' LIDARs that's distracting at the moment.

It is higher because they don't drive like humans. They brake at the slightest things while driving down roads, they come to a full stop at the line at an intersection even when it is clear nobody is coming, and when approaching a stop or turn they seem to have a tendency to slow down by dropping off the throttle instead of braking. You know how most people's workplace would come screeching to a halt if you followed every rule to the letter? It is like that but on the road.

The big lidar and porcupine of sensors is likely reducing the accident rate because it alerts other drivers that the car is different and most will be more cautious around them.

uninterrupted
Jun 20, 2011
It is critical that we remember the real victims of the Holocaust: the Soviets.

Claiming that the Holocaust was about Jews is Zionism. Biden is doing it to provide cover to Israel!

asdf32 posted:

The worst case for uber is that they're a national brand black car service with an app. It might not be a revolution, it is a viable business. But there's room between where they are now in terms of price/service and existing taxis which they compete with, so they'll probably be something more than that.

Actually, how does all the VC investment fit into all this? Don't they mostly get compensated in equity?

If Uber suddenly has to bring all drivers in as employees/cover commercial insurance/etc and rates raise to around what local cabs charge, can they just go "welp, we're not gonna be the next facebook" and just leave the VCs with stock they overpaid for?

Shifty Pony
Dec 28, 2004

Up ta somethin'


If uber loses their fight for calling their employees 1099 contractors back wages and expenses will instantly bankrupt them.

The investor terms are secret but reports are that VCs learned from last time and put real teeth into their ability to claw back their money in the case of things going sideways. They'd likely force a liquidation over a drawn out reorganization.

asdf32
May 15, 2010

I lust for childrens' deaths. Ask me about how I don't care if my kids die.

uninterrupted posted:

Actually, how does all the VC investment fit into all this? Don't they mostly get compensated in equity?

If Uber suddenly has to bring all drivers in as employees/cover commercial insurance/etc and rates raise to around what local cabs charge, can they just go "welp, we're not gonna be the next facebook" and just leave the VCs with stock they overpaid for?

Yep.

See groupon. Change the world -> cupon emailer. Whoops.

Adbot
ADBOT LOVES YOU

ComradeCosmobot
Dec 4, 2004

USPOL July

Popular Thug Drink posted:

it's amazing how fundamentally some right wingers misunderstand arguments. "This prototype technology won't be viable for a long time, but it will surely save the enormous packet of hype and exploitation branded Uber!"

The City of Sunnyvale shot down a bus-only lane on one of the major thoroughfares through town partly on the basis that self-driving cars will make the need for buses obsolete.

  • Locked thread