|
I can't imagine driverless cars working at all unless there's nothing but driverless cars on the roads
|
#
?
Jun 10, 2019 12:44
|
|
|
|
| # ? Jun 7, 2024 23:01 |
|
|
Booourns posted:I can't imagine driverless cars working at all unless there's nothing but driverless cars on the roads you could have driverless (i'm defining this as cars without steering wheels) for niche applications within certain areas that are relatively unchallenging, like well mapped and orderly urban areas with mostly predictable and favorable weather. los angeles is almost ideal. boston or pittsburgh would be hard mode. and then there are some applications which are unlikely to ever switch away from human control, such as emergency vehicles, off-road vehicles, service vehicles like tow trucks or construction vehicles, etc. also due to cost and enthusiast factors (many car guys are unlikely to switch to self-driving for daily riding, though many car guys are also likely to be the earliest of adopters) you'll see a mix of manually driven vehicles in any sort of mass adoption of driverless vehicles scenario which will diminish any potential positive network effects, assuming the huge technical hurdles of fleet self-driving can be worked out and of course, all of this crumbles in the face of the guaranteed congestion which is a natural consequence of our car-dependent built environment and how it will continue to be clogged to poo poo with cars of any kind as the urban population continues to grow
|
|
#
?
Jun 10, 2019 13:19
|
|
|
The problem isn't if autonomous driving can work, it's what will Uber do with all the vehicles they would then own. They are not set up for that, their whole business model is predicated on someone else owning and maintaining the vehicles. Their whole self driving car project is investor story time.
|
|
#
?
Jun 10, 2019 13:55
|
|
|
A car without steering wheels on a fixed predicable path? Sounds like a train.
|
|
#
?
Jun 10, 2019 14:05
|
|
|
Surprise Giraffe posted:I'd be surprised if someone like Uber doesn't release a driverless car in 5-10 years and whoever invests in these companies will have access to a staggeringly valuable monopoly that was easily worth such a long term investment. Since there is shitloads of free money and nothing more valuable to invest in thats what they do. People seem as convinced ITT that there will never be workable driverless cars as emphatically as tech bros are convinced theyll be here by xmas. They who conquer fusion will hold the world in their hand and stand to make trillions. If you are not yet invested in this life changing opportunity I can help you!
|
|
#
?
Jun 10, 2019 15:04
|
|
|
HootTheOwl posted:A car without steering wheels on a fixed predicable path?
|
|
#
?
Jun 10, 2019 15:04
|
|
|
Rent-A-Cop posted:The entire history of not-train mass transit ideas is pretty much finding ways to reinvent trains but not call then trains. What about that Uber clone who wanted to have a large pool vehicle driving predictable routes around San Francisco with request stops and easily marked locations for it to pick up passengers?
|
|
#
?
Jun 10, 2019 15:06
|
|
|
Uber has already satisfied the audience for driverless cars with the "I don't want to make small talk with the driver" button. Only thing that's left are mandatory robot prosthetics to wear
|
|
#
?
Jun 10, 2019 15:10
|
|
|
Total Meatlove posted:What about that Uber clone who wanted to have a large pool vehicle driving predictable routes around San Francisco with request stops and easily marked locations for it to pick up passengers?
|
|
#
?
Jun 10, 2019 15:18
|
|
|
Rent-A-Cop posted:Buses are simply a degenerate form of train. A train without rails. Since even use overhead wires!
|
|
#
?
Jun 10, 2019 15:19
|
|
|
luxury handset posted:there may never be workable driverless cars, but rather cars which are almost self driving except in all the edge cases where they are not, and this scenario is more likely to happen on the scale of decades than "in the next five years'". anyone who tells you anything different is trying to spin a dream they can use to sell you something I dunno. I see where you're coming from, but machine learning stuff seems to necessarily progress in what looks from the outside like random leaps with gaps of stagnation in between. Even if it's not there yet by 2025, I get the sense that progress will accelerate in the next decade like it seemed to over the last one given the amount of interest. Frankly, given they don't want to do the moral and sensible thing and pour money into education, basic research and social services, the rich and powerful throughout the west don't really have anything else to invest in at this point do they? If anything, the fact that there are all these shonky prototypes out there, billions invested in developing them at multiple companies, suggest that the possible solutions that don't actually work ought to be pruned out in the near term and investment redirected towards better ones, right? I'm not trying to argue that the Free Market will miraculously produce a solution, but that the actual fundamental growth factor behind tech i.e. patronage, is going nowhere while the current bubble stays put, which it seems determined to do. Really unless the government can no longer print the banks free money somehow I don't think the tech craze will ever alleviate, because our economies have so few other avenues for groth to siphon off while the average citizens capacity for consumption and personal growth is limited by public underinvestment. That's sort of the dynamic that's been dominating everything since 2008. So far there haven't been so many scams that technological problems that are potentially profitable to solve don't capture capital's attention.
|
|
#
?
Jun 10, 2019 16:32
|
|
|
A friend shared this, didn't see it posted in the thread: http://nymag.com/intelligencer/amp/2019/06/wework-adam-neumann.html?__twitter_impression=true Anyone have strong opinions? Other than being bougie and all the generically bad start-up culture stuff I don't really know that much about WeWork but apparently it's likely gonna be the next big IPO
|
|
#
?
Jun 10, 2019 16:33
|
|
|
Surprise Giraffe posted:I dunno. I see where you're coming from, but machine learning stuff seems to necessarily progress in what looks from the outside like random leaps with gaps of stagnation in between. Even if it's not there yet by 2025, I get the sense that progress will accelerate in the next decade like it seemed to over the last one given the amount of interest. Frankly, given they don't want to do the moral and sensible thing and pour money into education, basic research and social services, the rich and powerful throughout the west don't really have anything else to invest in at this point do they? If anything, the fact that there are all these shonky prototypes out there, billions invested in developing them at multiple companies, suggest that the possible solutions that don't actually work ought to be pruned out in the near term and investment redirected towards better ones, right? I mean it's always possible someone will make a breakthrough in some way but the first step of 'car that can drive itself in very simple situations' isn't really a sign that things are accelerating and we'll totally have self-driving cars any time now. That's the easy step, and also the one that was sexy to work on. Now you've got the really boring step of making that car not murder any pedestrians, deal with the idiot hellfuckers you see on any given road, and handle weather at all. Those are tricky problems to solve if your particular hammer is machine learning because they're all erratic, unpredictable problems.
|
|
#
?
Jun 10, 2019 16:48
|
|
|
Surprise Giraffe posted:I dunno. I see where you're coming from, but machine learning stuff seems to necessarily progress in what looks from the outside like random leaps with gaps of stagnation in between. Even if it's not there yet by 2025, I get the sense that progress will accelerate in the next decade like it seemed to over the last one given the amount of interest. Frankly, given they don't want to do the moral and sensible thing and pour money into education, basic research and social services, the rich and powerful throughout the west don't really have anything else to invest in at this point do they? If anything, the fact that there are all these shonky prototypes out there, billions invested in developing them at multiple companies, suggest that the possible solutions that don't actually work ought to be pruned out in the near term and investment redirected towards better ones, right? there's an unquestioned assumption in your post that is linked to decades of a certain technoutopian american perspective - that self driving cars are somehow to the solution to some kind of nebulous problem. that problem being, perhaps, that driving everywhere is bad and it sucks and people don't like to do it. we talk about this problem in various smaller ways. driving is dangerous! congestion is bad, wastes resources, and is boring! land use is inefficient and exclusionary! self driving cars aren't really going to 'fix' any of these definitions of the larger problem though we know how to fix this problem, as indicated by the sentiment in your post i've italicized. we can make cities less dependent on cars. this is happening, even in suburbs, today. it is not a solution which requires more research, it requires political buy in and public approval to authorize the funds and regulatory changes necessary to make suburbs less car dependent, to make american cities less car dependent. but social problems seem impossible to solve because they are far more complicated than just applying consumer technology as a solution. there is also little incentive for for-profit entities to pitch non-profit 'solutions' to complex social problems. so there's a whole song and dance, performed by car manufacturers, about how self driving cars will dig us out of the car dependency hole we've created despite that being an unsupported and likely ineffective solution in the first place when you think about it so yeah, capital is going to capital. but why even give them the benefit of the doubt that they can make a magic car which makes all the problems of too many cars less of a problem? Feinne posted:I mean it's always possible someone will make a breakthrough in some way but the first step of 'car that can drive itself in very simple situations' isn't really a sign that things are accelerating and we'll totally have self-driving cars any time now. That's the easy step, and also the one that was sexy to work on. Now you've got the really boring step of making that car not murder any pedestrians, deal with the idiot hellfuckers you see on any given road, and handle weather at all. Those are tricky problems to solve if your particular hammer is machine learning because they're all erratic, unpredictable problems. yeah. we could have had self driving cars years ago. people have been drooling about self driving cars for most of a century. but when you mix people and robots in the same space you end up with a lot of dead people, and introducing robots into the most public of public spheres - the streets - is a far more complex problem than it seems if one just focuses on the technical capacity of the vehicles themselves Mr. Fall Down Terror fucked around with this message at 16:53 on Jun 10, 2019 |
|
#
?
Jun 10, 2019 16:51
|
|
|
Self driving will be outsourced to a mechanical Turk in real time. People trying to post anime porn on 4chan will be required to drive your car for 5 seconds without crashing to prove they aren't a bot.
|
|
#
?
Jun 10, 2019 17:01
|
|
|
BougieBitch posted:A friend shared this, didn't see it posted in the thread: WeWork’s business model seems to depend on real estate prices going up forever. They go into a long term lease (or buy, I’m not sure) a building and then do shorter leases to tenants. So if the price of the building goes up they make money. But if the price of the building (and therefore the rent they can charge) goes down they’re hosed because they’re stuck with the same payment. Also they have some fuckery with how they report their earnings but I don’t really understand quite what it is.
|
|
#
?
Jun 10, 2019 17:14
|
|
|
Feinne posted:I mean it's always possible someone will make a breakthrough in some way but the first step of 'car that can drive itself in very simple situations' isn't really a sign that things are accelerating and we'll totally have self-driving cars any time now. That's the easy step, and also the one that was sexy to work on. Now you've got the really boring step of making that car not murder any pedestrians, deal with the idiot hellfuckers you see on any given road, and handle weather at all. Those are tricky problems to solve if your particular hammer is machine learning because they're all erratic, unpredictable problems. Seems like there's at least a few driverless cars basically trundling around with human supervisors all the time now. Apparently weather conditions are hard to deal with and teslas occasionally try to kill people. It's not exactly discouraging for 5-10 years though. luxury handset posted:there's an unquestioned assumption in your post that is linked to decades of a certain technoutopian american perspective - that self driving cars are somehow to the solution to some kind of nebulous problem. that problem being, perhaps, that driving everywhere is bad and it sucks and people don't like to do it. we talk about this problem in various smaller ways. driving is dangerous! congestion is bad, wastes resources, and is boring! land use is inefficient and exclusionary! self driving cars aren't really going to 'fix' any of these definitions of the larger problem though A mass transit solution that could completely replace cars is itself a techno-utopian idea I think. Systems like that will never cover the whole country, and never be as convenient as hopping in your car to go wherever at whim. I do prefer public transport but then I live in Europe, and I'd still love a driverless car for getting to out-of-the-way locations. The idea of trying mass transportize America sounds completely bonkers outside the cities, and even then a vast logistical challenge you'd have to develop tech so it would function efficiently. A mix of autonomous cars and public tranport is obviously the most sensible solution. You get a system to handle the majority of journeys with autonomous vehicles to fill in the gaps.
|
|
#
?
Jun 10, 2019 17:25
|
|
|
If we were going to have largely self driving fleet in 5-10 years, we'd need to have finished vehicles on the market now. Instead of vehicles that ram themselves into fire trucks and the broad sign of tractor trailers, or vehicles that can't successfully react to a person with a bike crossing a wide street.
|
|
#
?
Jun 10, 2019 17:43
|
|
|
Surprise Giraffe posted:A mass transit solution that could completely replace cars is itself a techno-utopian idea I think. Systems like that will never cover the whole country, and never be as convenient as hopping in your car to go wherever at whim. I do prefer public transport but then I live in Europe, and I'd still love a driverless car for getting to out-of-the-way locations. we don't need it to completely replace cars. this is a standard reduction to absurdity reply, like "oh we can't make trains cover the whole country so what is the point of making trains cover the 10% of the country where more than half the population lives?" even in places with excellent public transit, cars still exist and coexist with a diverse selection of transportation modes. the problem is more that in the united states (and other places) the built environment is set up such that the ONLY feasible transit mode is an automobile, and that is Bad meanwhile, we don't need autonomous cars to fill the gaps. manually driven cars are fine. they work perfectly well, today, without having to wait for a technological solution (magic cars) to a social problem (built environment mandates car usage). the solution to a built environment that mandates cars is to change the built environment, over time, to permit alternative transportation modes, not to have sufficiently magical vehicles which may or may not ever appear. meanwhile american cities are becoming less car dependent all the time, right now, using solutions available to us right now Surprise Giraffe posted:The idea of trying mass transportize America sounds completely bonkers outside the cities, and even then a vast logistical challenge you'd have to develop tech so it would function efficiently. A mix of autonomous cars and public tranport is obviously the most sensible solution. You get a system to handle the majority of journeys with autonomous vehicles to fill in the gaps. it sounds bonkers, but it really isn't. it's a lot less bonkers than assuming self driving cars will somehow fix transportation problems, given that this all but guarantees car trips will go up without changing at all any of the underlying demand assumptions which generate car trips in the first place. furthermore it is happening today. many cities are building more transit. many suburban towns are building new denser, mixed use downtowns for themselves which will serve as transit hubs. it is the current Hot Trend in local governance. here's a few examples from notoriously sprawling atlanta https://www.suwanee.com/explore-suwanee/downtown/town-center https://atlanta.curbed.com/2018/7/25/17610340/lawrenceville-georgia-development-gwinnett-county-construction https://accesswdun.com/article/2018/12/739762/gwinnett-commission-approves-64-acre-buford-development-with-a-driving-range these sorts of things are necessary because when we talk about self driving cars and how they will be safer (they won't) or how they will solve congestion (they won't) it is essential to recognize there is a race against time here, in that urban populations are growing and will continue to grow, meaning more demand for transportation, more car trips being made, etc. so any efficiency gains made by self driving fleets or whatever will be overwhelmed simply by there being more and more and more people. so you must address the demand side of things (more people needing to go places) while also trying to address the supply (different modes of transportation)
|
|
#
?
Jun 10, 2019 17:44
|
|
|
Owlofcreamcheese posted:I still don't understand why all these years later the traditional taxi companies still suck so bad at modernizing. The thing people like about uber is the app, if it was the same style of app but it called yellow cabs everyone would have used that instead but it's 2019 and there are still cab companies that only take phone calls. Even the companies with apps are all limited compared to the feature set uber has and are near useless anyway because they only work in a single city. Uber exists because the taxis in San Francisco were at a level of awful that had to be experienced to be believed. In my case, I started traveling with a folding bike and nixed the issue altogether, but sometimes at conferences, the ability to summon a ride for 5 people at once was worth it.
|
|
#
?
Jun 10, 2019 18:03
|
|
|
Total Meatlove posted:Every cab company in the UK has an app, the issue is agreeing a national standard and interoperability. I once had a destination in London that the Black Cab driver couldn't locate, nor could Google Maps. Strangely Apple Maps did. A location had changed its name.
|
|
#
?
Jun 10, 2019 18:04
|
|
|
luxury handset posted:we don't need it to completely replace cars. this is a standard reduction to absurdity reply, like "oh we can't make trains cover the whole country so what is the point of making trains cover the 10% of the country where more than half the population lives?" even in places with excellent public transit, cars still exist and coexist with a diverse selection of transportation modes. the problem is more that in the united states (and other places) the built environment is set up such that the ONLY feasible transit mode is an automobile, and that is Bad I was replying to your accusation that I'm somehow 'techno-utopian' for thinking we'll have driverless cars in 5-10 years. You can have as much public transportation as you like, it won't get near to removing demand for cars so much that driverless ones are pointless unless fewer people start going fewer places. If anything, autonomous vehicles will just end up a de-facto aprt of public transportation, provided on the cheap by companies like Uber (which is blatently their longest-term idea for profitability, whatever you think of the plausability of Uber themselves succeedingin it)
|
|
#
?
Jun 10, 2019 18:09
|
|
|
Surprise Giraffe posted:Seems like there's at least a few driverless cars basically trundling around with human supervisors all the time now. Apparently weather conditions are hard to deal with and teslas occasionally try to kill people. It's not exactly discouraging for 5-10 years though. It's extremely discouraging if you expect that in a similar amount of time we'll have a viable commercial product, because the 'car that can drive itself under very controlled circumstances and only sometimes kill a lady trying to cross the street' isn't like 'this product is almost ready', that's 'okay now we have the basis that even lets us start on the hard part'.
|
|
#
?
Jun 10, 2019 18:12
|
|
|
luxury handset posted:yeah. free parking is a big weird thing in american society because in the suburbs it is simply assumed you can store your car wherever, for free, ether on your property or provided to you as a convenience by whatever business you are patronizing. parking minimums are a giant and secret driver of suburban sprawl in the united states, because in the sort of car dependent landscape built in the US and other nations (canada, australia, etc.) there's enough land to where this is not a big deal. but even in american cities, if you're poor or live in certain heavily urbanized locations, good luck getting a regular and inexpensive place to store your car. this can be an additional obstacle to car ownership The family business was parking garages in Manhattan (starting in the 1920's). Grandfather had one in Greenwich Village, my dad had one midtown. It was super expensive even in the 1970's. Just checked the price today (Icon owns it now), $38/day. One consequence of this, many NY'ers don't even own a car (my brother in Brooklyn for example). Free parking is bad for the climate.
|
|
#
?
Jun 10, 2019 18:14
|
|
|
Surprise Giraffe posted:You can have as much public transportation as you like, it won't get near to removing demand for cars so much that driverless ones are pointless i wish you luck on this argument you insist on having with yourself Surprise Giraffe posted:provided on the cheap by companies like Uber (which is blatently their longest-term idea for profitability, whatever you think of the plausability of Uber themselves succeedingin it) uber definitely will not be providing the vehicles. their entire business model is based on offloading the cost of vehicles onto drivers as a way of artificially reducing the cost of automobile trips. uber's self driving car research is oriented around developing patents that means uber cannot be sidestepped, as a way of getting their software built in to other manufacturer's platforms. individual car owners will also not be providing the vehicles - the idea of sending your car out to make money is a blatantly silly one, as you'll only need to clean vomit and used condoms from your car a few times before your morning commute to make you recalculate the worth of sending out your own property to drive strangers around with minimal supervision. it is more likely that middlemen firms would arise who own some dozens of vehicles and clean/repair/maintain them in a central facility to gain efficiencies of scale. but this puts uber in a dangerous position, as car-fleet owners could be swayed to use some other platform which takes less of a cut than uber, which is why uber needs to make their app the one that gets baked into ford's cars or toyota's cars or whoever
|
|
#
?
Jun 10, 2019 18:16
|
|
|
Surprise Giraffe posted:
I last lived in Los Angeles in 1994. Since 2018 I've been working there a few days a week. Mass transit has vastly improved. Didn't think it was ever going to happen, but it did. But yeah, I don't know if that's possible in many places in the USA.
|
|
#
?
Jun 10, 2019 18:17
|
|
|
VideoGameVet posted:I last lived in Los Angeles in 1994. Since 2018 I've been working there a few days a week. Mass transit has vastly improved. Didn't think it was ever going to happen, but it did. dallas texas, of all places, has the longest light rail network in america. they've been building it for over 20 years and they're still going
|
|
#
?
Jun 10, 2019 18:18
|
|
|
I thought mass transit couldn't be good until I moved to Europe. I don't know if it's something in the water here but the mass transit systems in European cities are darn good. Of course if you ask any European they'll say it sucks but they haven't seen how bad it can really get.
|
|
#
?
Jun 10, 2019 19:45
|
|
|
Cable cars are making a resurgence, although less for actual transportation and more it gives property developers immense hard-ons.
|
|
#
?
Jun 10, 2019 20:23
|
|
|
I used to live in an old street car suburb where it used to have a direct line into the metro center and instead it got torn out because cars hated the tracks and instead it was eventually replaced with a 30 minute walk to the nearest light rail station. Progress!
|
|
#
?
Jun 10, 2019 20:37
|
|
|
luxury handset posted:dallas texas, of all places, has the longest light rail network in america. they've been building it for over 20 years and they're still going Well a lot of other cities that would be in the position to do that instead have functioning commuter rail or rapid transit taking up system length that Dallas has to do all by light rail.
|
|
#
?
Jun 10, 2019 20:57
|
|
|
luxury handset posted:you could have driverless (i'm defining this as cars without steering wheels) for niche applications within certain areas that are relatively unchallenging, like well mapped and orderly urban areas with mostly predictable and favorable weather. los angeles is almost ideal. boston or pittsburgh would be hard mode. and then there are some applications which are unlikely to ever switch away from human control, such as emergency vehicles, off-road vehicles, service vehicles like tow trucks or construction vehicles, etc. Yeah, and pretty much every instance of that use case is better served by light rail and occasional subsidized taxi service, due to said congestion and the efficiencies due to reduced fuel usage and lack of need to store thousands of individual vehicles.
|
|
#
?
Jun 11, 2019 00:34
|
|
|
Surprise Giraffe posted:Seems like there's at least a few driverless cars basically trundling around with human supervisors all the time now. Apparently weather conditions are hard to deal with and teslas occasionally try to kill people. It's not exactly discouraging for 5-10 years though. The key is the human supervisors. I totally agree that we have cars on the road that would have stunned researchers just a decade ago. I think those vehicles are excellent test beds for what will likely show up as some really nice driver-assist features in a few years. But that's the big difference between Ford/Volvo/etc and Uber. Car manufacturers can all benefit from incremental improvements, Uber can't. A car that can drive itself in 75% of situations is valuable to Ford, but is worthless to Uber. What is Uber going to do, kick the passengers out when it rains or they hit an unplanned lane closure?
|
|
#
?
Jun 11, 2019 01:30
|
|
|
https://twitter.com/fchollet/status/1137788467591565312
|
|
#
?
Jun 11, 2019 08:59
|
|
|
Liquid Communism posted:Yeah, and pretty much every instance of that use case is better served by light rail and occasional subsidized taxi service, due to said congestion and the efficiencies due to reduced fuel usage and lack of need to store thousands of individual vehicles. Basically this but the taxis are driverless cars that are commercially available on 5-10 years from now.
|
|
#
?
Jun 11, 2019 10:08
|
|
|
even in a magic scenario with mass adoption of driverless cars, they will need places to park in the city center. being driverless doesn't mean maintenance free, and there is no driverless car scenario in which humans are not a diurnal species with set daily habits that generate peak travel times
|
|
#
?
Jun 11, 2019 12:20
|
|
|
Surprise Giraffe posted:Basically this but the taxis are driverless cars that are commercially available on 5-10 years from now. this is blockchain dreams, but for cars. "But, just think of all the applications of magical flying unicorn ponies! The demand is such that it's gotta happen!"
|
|
#
?
Jun 11, 2019 12:23
|
|
|
2020: "imagine... a fleet of driverless cars, never stopping, never ceasing, never needing to rest... a constant flow of human travel, like a great river of commerce. what efficiency!" 1900: "hold my digestif" https://www.youtube.com/watch?v=-X0xoUki128
|
|
#
?
Jun 11, 2019 12:27
|
|
|
divabot posted:this is blockchain dreams, but for cars. "But, just think of all the applications of magical flying unicorn ponies! The demand is such that it's gotta happen!" If demand made things exist like this we'd be knee deep in room temperature superconductors and cold fusion plants by now.
|
|
#
?
Jun 11, 2019 13:07
|
|
|
|
| # ? Jun 7, 2024 23:01 |
|
|
Pochoclo posted:I thought mass transit couldn't be good until I moved to Europe. I don't know if it's something in the water here but the mass transit systems in European cities are darn good. Of course if you ask any European they'll say it sucks but they haven't seen how bad it can really get. Cities aren't built around the car, which means that the distances are shorter between things and it's easier to build a network that actually goes places that are convenient for people to get to. There's an expectation that ordinary people are going to use it instead of just the poor and despaerate, which means that it gets funded. Building your environment with the assumption that everyone is going to own a car (which takes up space all of the time) had a big domino effect that makes it really hard to backfill transit.
|
|
#
?
Jun 11, 2019 13:16
|
|
