Autonomous Cars General Discussion

[ProjectWHaT]

With the rise of compute power and big data, car and tech companies have been developing complicated AI systems to drive cars, such as Tesla's AutoPilot, Cadillac's SuperCruise, Waymo, and Comma.AI.

As this tech develops, the general public understands less and less, the government understands less and less. There are barely any rules and regulations on AI cars. As these companies and media spew out more and more buzz words like "self driving", the understanding of what these cars actually are is become hazy.

The lack of understanding can be seen in the media's portrayal of the technology, making wild claims such as "semi-driverless cars". Or the many accidents that has happened involving Uber and their incompetent system ( "[Waymo's] cars went an average of nearly 5,600 miles before the driver had to take control from the computer to steer out of trouble. As of March, Uber was struggling to meet its target of 13 miles per 'intervention' in Arizona")

Many advocates for AI, autonomous cars, and human freedom, such as Alex Roy, have been trying to educate the public on these hot topics. Alex Roy has written a "Human Driving Manifesto". In the manifesto, he calls for the parallel development of autonomous cars and the people's right to be able to control the car. He also calls for the privacy of users from data collecting systems that will monetize your driving data.

Autonomous cars will most likely become a part of every day life in the near future. A better understanding of the technology is needed, especially by government officials, for the safety of the people.

What are your thoughts on autonomous cars?

 

[Eunos_Cosmo]

There's already a thread for this, more or less, here.

I'll summarize my thoughts though: Overhyped (particularly Telsa & Uber which seem to be irresponsibility overhyped) to the point of being dangerous.

 

[ToyGTone]

There's already a thread for this, more or less, here.

I'll summarize my thoughts though: Overhyped (particularly Telsa & Uber which seem to be irresponsibility overhyped) to the point of being dangerous.

I never understood why it's a such big deal. Couldn't people just use public transport (as longs as it's good and not awful like in bahrain)? Apart from taking a nap, I can't see myself doing anything while being driven around. Like, why do people really REALLY care about it so much?

Maybe because I just live in a tiny country and just travel around 62 KM to university and back every Sunday and Wednesday while others travel hundreds of miles. That I can understand.

 

[TexRex]

Couldn't people just use public transport?

The infrastructure for efficient (re: rail-based) public transportation is virtually non-existent in the United States. AVs make use of existing roads and don't rely on large groups of people that just happen to be going the same place with predetermined stops on a vehicle that smells like pee.

 

[Eunos_Cosmo]

The infrastructure for efficient (re: rail-based) public transportation is virtually non-existent in the United States. AVs make use of existing roads and don't rely on large groups of people that just happen to be going the same place with predetermined stops on a vehicle that smells like pee.

Yeah, but...

By the way, the US used to have a vast streetcar (horsedrawn initially, followed by electrified) network (yes, even in Texas), which was largely dismantled after cars really took over (many cities just literally paved over existing streetcar lines). A few cities have started to implement them again, like Portland & San Francisco. Our electrified-bus system seems to work pretty decently here in SF. Urbanistically, public transit is a better use of investment than AVs, which will essentially replace taxis. I don't see them doing anything other than marginally improving safety and causing a lot of taxi/rideshare drivers to lose their jobs. City congestion will probably either worsen or stay the same, especially if huge monetary incentives cause people to stop using conventional public transit. If that happens, it's likely those transit systems will fail completely. But the hype machine has too much inertia now, AVs are definitely coming.

 

[TexRex]

Yeah, but...

"This is how many cars are removed from the road if all these people are going the same route on a pee-perfumed bus."

Our electrified-bus system seems to work pretty decently here in SF.

Cool.

Urbanistically, public transit is a better use of investment than AVs, which will essentially replace taxis. I don't see them doing anything other than marginally improving safety and causing a lot of taxi/rideshare drivers to lose their jobs.

Urbanistically, lots of people still drive themselves and themselves only. Owning a car, be it low-tech or autonomous, isn't justifiable/practical/affordable for everyone, but rideshare/taxi services offer an alternative, and those that can't afford [or don't have ready access to] that alternative are left with public transit. I don't see this changing anytime soon as a result of autonomous vehicle development and yet I'm supposed to weep for the presumed job losses.

Oh and a pox on those venturing to advance personal vehicle autonomy instead of addressing public transit. May they all rot in Hell, amiright?

City congestion will probably either worsen or stay the same, especially if huge monetary incentives cause people to stop using conventional public transit.

How dare anyone implement such [huge monetary] incentives. A pox on them as well!

AVs are definitely coming.

 

[ProjectWHaT]

Autonomous public transportation will work in cities, but a lot of people also live in the suburbs or in rural areas. There is no public transportation in these areas. The average guy will sit in traffic for an hour in the morning and in the evening. With AVs, it'll take the stress away from being stuck in traffic.

 

[Eunos_Cosmo]

"This is how many cars are removed from the road if all these people are going the same route on a pee-perfumed bus."


Cool.


Urbanistically, lots of people still drive themselves and themselves only. Owning a car, be it low-tech or autonomous, isn't justifiable/practical/affordable for everyone, but rideshare/taxi services offer an alternative, and those that can't afford [or don't have ready access to] that alternative are left with public transit. I don't see this changing anytime soon as a result of autonomous vehicle development and yet I'm supposed to weep for the presumed job losses.

Oh and a pox on those venturing to advance personal vehicle autonomy instead of addressing public transit. May they all rot in Hell, amiright?




How dare anyone implement such [huge monetary] incentives. A pox on them as well!

View attachment 730770

That was a needlessly negative & sarcastic response. Its a discussion board man...

"Urbanistically, lots of people still drive themselves and themselves only. Owning a car, be it low-tech or autonomous, isn't justifiable/practical/affordable for everyone, but rideshare/taxi services offer an alternative, and those that can't afford [or don't have ready access to] that alternative are left with public transit. I don't see this changing anytime soon as a result of autonomous vehicle development and yet I'm supposed to weep for the presumed job losses."

I'll respond to this bit and ignore the rest of your sass. (Did you have a really pissy bus experience? )

My whole argument is that existing public transportation systems are chronically underfunded (NYC subway is a prime example) yet autonomous car development enjoys vast VC funding. It would be great if more money was invested into transit infrastructure, as unlikely as that is to happen, because it has the potential to make cities better which I don't think AVs will do (because of too many reasons to post). Also, in places other than Texas, it's not just poor people who take public transit. But even Dallas has pretty decent light rail & commuter rail that not just poor people take.

Ok I won't post anymore transit discussion in the AV thread.

 

[TexRex]

That was a needlessly negative & sarcastic response

directed at a needlessly speculative & condescending comment.

I mean...AV technology might put some individuals* out of work in the distant future, but that's no justification for laying a guilt trip right now on proponents of its advancement and implementation.

*Surely I'm not the only one who sees the irony in lumping ridesharing and traditional taxi services together here when proponents of the latter cite the disruptive nature of ridesharing platforms as putting people out of work...but I guess it's okay so long as those jobs are being taken by real living people rather than glorified toasters.

My whole argument is that existing public transportation systems are chronically underfunded yet autonomous car development enjoys vast VC funding.

Had it been seen as lucrative, surely it would have garnered the same sort of attention.

It would be great if more money was invested into transit infrastructure, as unlikely as that is to happen, because it has the potential to make cities better

There was plenty of opportunity before the AV "hype train" started chugging along [autonomously, I'd imagine]. There must be a reason for the lack thereof.

Dallas has pretty decent light rail & commuter rail

And I'm given to understand that CapMetro here in Austin has seen improvements since I was in college, but it still doesn't work for my present situation regardless of whether I want to make use of it, and I'm not so motivated as to seek a situation where it would work.

Ok I won't post anymore transit discussion in the AV thread.

So you address an open line of discussion and then shut yourself off from it? "It's a discussion board, man."

 

[Eunos_Cosmo]

directed at a needlessly speculative & condescending comment.

What was condescending? I certainly didn't intend to be condescending, I even acknowledged that you were correct in the post, but my intent was to argue that public transit is partially about reducing congestion. I could see the picture I posted as being a bit snarky, but I didn't mean to be.

I mean...AV technology might put some individuals* out of work in the distant future, but that's no justification for laying a guilt trip right now on proponents of its advancement and implementation.

No guilt trip intended. I stated that it could do that, but I didn't stake a position on it.

*Surely I'm not the only one who sees the irony in lumping ridesharing and traditional taxi services together here when proponents of the latter cite the disruptive nature of ridesharing platforms as putting people out of work...but I guess it's okay so long as those jobs are being taken by real living people rather than glorified toasters.

I don't have a horse in this race. Ridesharing (as well as Taxis) certainly has issues, but that is outside the scope of this thread.

Had it been seen as lucrative, surely it would have garnered the same sort of attention.

Public transit isn't sexy & new like AV development, but that doesn't mean it isn't important.

There was plenty of opportunity before the AV "hype train" started chugging along [autonomously, I'd imagine]. There must be a reason for the lack thereof.

Again, it's not sexy. It's a service. If some of these tech companies could be a little less obsessed with 'disruption' (looking at you, Bird) and a little more collaborative with cities (Uber appears to me heading this direction) than maybe infrastructure could become a little sexier again. Musk also seems to be heading this direction, albeit in a way that may or may not be aligned with the cities' interest.

And I'm given to understand that CapMetro here in Austin has seen improvements since I was in college, but it still doesn't work for my present situation regardless of whether I want to make use of it, and I'm not so motivated as to seek a situation where it would work.

I agree. I'm not sure why Austin can't get it's (transit) act together. I've never lived in Austin, but I have a few friends who do and I've visited quite often. Transit is definitely lacking. If it was better, I'm pretty sure people like yourself would use it. Hence the entire reason for my OP.

So you address an open line of discussion and then shut yourself off from it? "It's a discussion board, man."

I meant it's getting off topic.

 

[ProjectWHaT]

A car has crashed into a Waymo in autonomous mode according to the police, but Waymo claims it was in manual mode.

Waymo themselves have released a video of the crash.



https://jalopnik.com/google-minivan-in-autonomous-mode-involved-in-crash-in-1825786457

 

[cdubbbs]

I hope the technology advances quickly in terms of safety and function. Its a great option for people who really don't care about cars or the driving task. I'm all for it but I would never use it as I do enjoy driving and prefer to pay attention to what I am doing.

 

[Eunos_Cosmo]

An "Autopilot Buddy" has come to market allowing for theoretically attention-free driving on Teslas equipped with "autopilot". I'm becoming increasingly convinced that nothing between complete full on, "never need a driver ever" autonomy, and extremely basic "lane keep assist" is viable because humans are dumb and will use any opportunity given to them to be irresponsible.

 

[homeforsummer]

"This is how many cars are removed from the road if all these people are going the same route on a pee-perfumed bus."

I'm not a fan of buses (they come pretty far down my list of transport I'd use given the choice) but this is flippant to the point of gross inaccuracy.

Very few cars go directly point to point in the greater scheme of things. I'm driving the only car that leaves my own garage of course, but I'm one of probably a couple of hundred that ends up in the same car park at the end of the journey, and I certainly can't drive it right up to my desk once I get to work. And I use the same bits of road every day as probably thousands or tens of thousands of others going from vaguely the same point to vaguely another point.

Nobody claims buses or trains to be last-mile solutions but if implemented well they can be very good indeed for the large bit in the middle. Even in my position as an incorrigible petrolhead there are some journeys that I'd not even think about taking a car for unless the situation absolutely demanded it.

I've driven into London (not even central London) one single time over the past three or four years and regretted it massively, even though it was theoretically more convenient in terms of being able to park relatively close to my destination. I've taken the train in dozens of times though, and while it's not always perfect and certainly too expensive (thank god for work expenses) it's a far, far better solution than driving. If I'm going to the London office I have to park ten minutes' walk from it anyway so the 15 minutes' walk it takes from the station is hardly a poorer option. When the slave cylinder on my old MX-5 failed back when I lived in Newcastle many years back, I didn't get it fixed for two months because that city's public transport meant I simply didn't need to drive a car to get where I was going.

Buses are a similar story. I'm sure they can be quite grim in some places, but I can't remember the last time I took one that turned out to be massively inconvenient for where I was coming from or going to. That's why you have multiple buses with multiple routes and there isn't just one single bus that circumnavigates an entire city dropping people off door to door...

I shouldn't need to explain why I don't personally use the bus to get to work given what I do for a living, but it would be far from impossible for me to do so, and while I can certainly understand the benefits and appeal of using a car, the roads would be a much better place if well-designed public transport systems allowed fewer people to drive. In both practical and environmental terms it'd be several steps ahead of everyone simply jumping into autonomously-driven personal vehicles.

 

[TexRex]

I'm not a fan of buses (they come pretty far down my list of transport I'd use given the choice) but this is flippant to the point of gross inaccuracy.

Very few cars go directly point to point in the greater scheme of things. I'm driving the only car that leaves my own garage of course, but I'm one of probably a couple of hundred that ends up in the same car park at the end of the journey, and I certainly can't drive it right up to my desk once I get to work. And I use the same bits of road every day as probably thousands or tens of thousands of others going from vaguely the same point to vaguely another point.

Nobody claims buses or trains to be last-mile solutions but if implemented well they can be very good indeed for the large bit in the middle. Even in my position as an incorrigible petrolhead there are some journeys that I'd not even think about taking a car for unless the situation absolutely demanded it.

I've driven into London (not even central London) one single time over the past three or four years and regretted it massively, even though it was theoretically more convenient in terms of being able to park relatively close to my destination. I've taken the train in dozens of times though, and while it's not always perfect and certainly too expensive (thank god for work expenses) it's a far, far better solution than driving. If I'm going to the London office I have to park ten minutes' walk from it anyway so the 15 minutes' walk it takes from the station is hardly a poorer option. When the slave cylinder on my old MX-5 failed back when I lived in Newcastle many years back, I didn't get it fixed for two months because that city's public transport meant I simply didn't need to drive a car to get where I was going.

Buses are a similar story. I'm sure they can be quite grim in some places, but I can't remember the last time I took one that turned out to be massively inconvenient for where I was coming from or going to. That's why you have multiple buses with multiple routes and there isn't just one single bus that circumnavigates an entire city dropping people off door to door...

I shouldn't need to explain why I don't personally use the bus to get to work given what I do for a living, but it would be far from impossible for me to do so, and while I can certainly understand the benefits and appeal of using a car, the roads would be a much better place if well-designed public transport systems allowed fewer people to drive. In both practical and environmental terms it'd be several steps ahead of everyone simply jumping into autonomously-driven personal vehicles.

There are some good points made here, but what does it have to do with personal vehicle autonomy?

Are autonomous personal vehicles responsible for the decline in public transportation or does the blame lay on personal vehicles as a whole, for which autonomy is the next major technological step?

That said, I appreciate being called out on what I said by someone not involved in the original discussion. I know myself, and while I don't recall what was going on at the time of my response, I'm not surprised by the flippant nature of it.

 

[homeforsummer]

There are some good points made here, but what does it have to do with personal vehicle autonomy?

It's unavoidably linked because any discussion of autonomous vehicles is also one of personal future mobility, and mass transit has a fairly good claim already on covering that base.

It's arguably more relevant still when you remove the concept of actually driving from the equation of personal transportation, because in practical terms things like autonomous ride-sharing isn't far removed from mass transit. It's just a less efficient form of mass transit, albeit one that's less terrifying to car manufacturers whose business model depends on people still buying cars even if they aren't driving them...

 

[TexRex]

It's unavoidably linked because any discussion of autonomous vehicles is also one of personal future mobility, and mass transit has a fairly good claim already on covering that base.

It's arguably more relevant still when you remove the concept of actually driving from the equation of personal transportation, because in practical terms things like autonomous ride-sharing isn't far removed from mass transit. It's just a less efficient form of mass transit, albeit one that's less terrifying to car manufacturers whose business model depends on people still buying cars even if they aren't driving them...

The demand for personal vehicles is still one that exists, despite the number of individuals who utilize them regularly when they don't enjoy the act of driving as a part of daily life.

 

[homeforsummer]

The demand for personal vehicles is still one that exists, despite the number of individuals who utilize them regularly when they don't enjoy the act of driving as a part of daily life.

I'm aware, but then we re-arrive at the concept of there being very little societal benefit to autonomy beyond improving mobility for those who can't currently drive themselves and as-yet unclear estimates of reducing road deaths.

The problem with that is that very few companies boasting autonomous capabilities are designing cars for improving the mobility of the otherwise immobile - they're currently designing them to give regular people with driving licenses one fewer thing to do while they're texting as their mobile living room sits in the same traffic they've always sat in on a morning. Luxury pods with no need for a driver to control them.

There will no doubt be some small traffic benefit to autonomous vehicles, which may be less confused by the presence and actions of other AVs compared to the average human/human interaction. But if autonomy makes cars more viable for more people then traffic may get worse anyway as there'll be more vehicles using otherwise ill-suited roads - a bit like how adding an extra lane on a motorway doesn't cut traffic as more drivers feel compelled to use it.

So we're back to a conglomeration of vehicles sitting in queues every morning and evening; the only difference is that their occupants have more time to do stuff that isn't driving. This doesn't solve the key problems with modern transportation: There are too many cars in the first place, and too many unnecessary journeys are made.

The former issue is best solved by public transportation, which is a more space- and resource-efficient (building an autonomous car ultimately still requires the building and powering of a car) way of moving lots of people from place to place than individual vehicles. Public transportation is also more efficient in usage - it spends a greater proportion of the day moving people about, whereas a car spends most of its day parked (and parking is a whole other issue).

The latter has several options - staggered working/schooling hours to spread out traffic rather than focusing it into a morning and evening rush hour, encouraging people to use more appropriate methods of transport to get to work/school/etc, or the one that makes the most sense in the modern day, which is remote working.

The last thing clearly isn't appropriate for everyone (indeed, there is no silver bullet for any of this stuff), but imagine the traffic benefits if everyone whose job involves sitting at a computer all day could do so from home instead of driving to an office?

Full disclosure in all this: I want fewer vehicles on the road because I like cars and I like driving.

 

[TexRex]

I'm aware, but then we re-arrive at the concept of there being very little societal benefit to autonomy beyond improving mobility for those who can't currently drive themselves and as-yet unclear estimates of reducing road deaths.

The problem with that is that very few companies boasting autonomous capabilities are designing cars for improving the mobility of the otherwise immobile - they're currently designing them to give regular people with driving licenses one fewer thing to do while they're texting as their mobile living room sits in the same traffic they've always sat in on a morning. Luxury pods with no need for a driver to control them.

Are individuals and groups seeking to address convention as a means to make money obligated to do so in a manner that benefits society?

I'm sure that there are tech companies using autonomy to serve those with mobility difficulties, just as there are tech companies garnering bipartisan support in using their expertise to make hearing aids less expensive. Even then, they're encountering pushback, not only from existing companies concerned about their profits but also political groups perturbed by their representatives' cooperation with the other side. How is that to society's benefit?

And if people are going to continue driving distracted be their little screens, which they clearly are despite doing so being illegal (to say nothing about it being dangerous), isn't addressing the issue by requiring less of them to the benefit of those around them?

 

[homeforsummer]

Are individuals and groups seeking to address convention as a means to make money obligated to do so in a manner that benefits society?

Whether they're obligated to or not will become moot if the individual's impact increasingly affects those around them negatively. A world where everyone sits in nose-to-tail traffic isn't beneficial to anyone, whether they're doing so themselves in something that burns petrol or whether they're doing it autonomously and in an EV - the end result is still clogged roads and low quality of life.

People are of course free to make their own decisions, even if those decisions are idiotic. But there's worryingly little emphasis on giving people options beyond just "car you don't have to drive".

I'm sure that there are tech companies using autonomy to serve those with mobility difficulties, just as there are tech companies garnering bipartisan support in using their expertise to make hearing aids less expensive. Even then, they're encountering pushback, not only from existing companies concerned about their profits but also political groups perturbed by their representatives' cooperation with the other side. How is that to society's benefit?

Could you rephrase that? I'm not quite sure what you're asking.

Unless you're looking for the obvious answer, "it isn't" - if I do understand what you're asking, then of course private companies fighting against projects that might improve society isn't to society's benefit. But then that seems like a redundant question ("is doing bad things bad?"), which is why I'd like you to try and rephrase it as I might have misunderstood.

And if people are going to continue driving distracted be their little screens, which they clearly are despite doing so being illegal (to say nothing about it being dangerous), isn't addressing the issue by requiring less of them to the benefit of those around them?

It's addressing the issue but that doesn't make it the best way of doing so.

Cutting off my foot would address the issue of me stubbing my toe but it wouldn't help me walk any better.

 

[TexRex]

Could you rephrase that? I'm not quite sure what you're asking.

Your response indicates you understood, but I'm willing to amend the [rhetorical] question to:

"Is that pushback, be it in service of profits or holding political party lines, beneficial to society?"

It's not, and in the case of the latter, actions are supposed to be for society's benefit.

It's addressing the issue but that doesn't make it the best way of doing so.

Cutting off my foot would address the issue of me stubbing my toe but it wouldn't help me walk any better.

These are automotive and tech companies using their expertise to address an issue in a manner that fills a perceived need with hopes of making a great deal of money, and investing a great deal of money (in the form of time, resources, permissions and, unfortunately, compensations) as a bet that they'll see substantial returns.

It's been indicated that public transit systems have worked well in the past, but there was a desire for personal vehicles and companies willing to provide them. The result has been a decline in public transit use, and therefor profitability in the venture.

Remember what I said about using expertise to fill a need? Are you more inclined to see a butcher or a trained medical professional to mend that lame appendage? [rhetorical]

What you're saying isn't lost on me though.

 

[homeforsummer]

Your response indicates you understood, but I'm willing to amend the [rhetorical] question to:

"Is that pushback, be it in service of profits or holding political party lines, beneficial to society?"

It's not, and in the case of the latter, actions are supposed to be for society's benefit.

Thanks for clarifying. Now I'm just not quite sure of the context you're asking it in

The post of mine you initially quoted suggested there's little societal benefit to AVs, and that companies developing AVs seem to effectively be making them for lazy people.

To which of these points does the pushback apply (rhetorical or not), and whom exactly is doing the pushing?

It's been indicated that public transit systems have worked well in the past, but there was a desire for personal vehicles and companies willing to provide them. The result has been a decline in public transit use, and therefor profitability in the venture.

They didn't just work well in the past, they work well today - but only in certain places.

The logical thing to do would be to see why they work well in those places (London being one, for all its faults) and apply that elsewhere. Naturally, it'd be impossible (and pointless) to apply it everywhere, but when public transport doesn't work in places like Los Angeles, I suspect as much of that is because it's not implemented well enough to work, as it is a kind of rejection of the system by those that might benefit from it.

And it has to be said, a good public transport system is also appropriate for those who can barely afford any kind of car in the first place, let alone a shiny new autonomous one. A lot of people close to the breadline spend a lot of what they earn on the shoddiest of cars just so they can get to work etc. I don't doubt their lives would be improved if something let them get to work at much lower cost - and that something won't be AVs.

 

[TexRex]

Thanks for clarifying. Now I'm just not quite sure of the context you're asking it in

The post of mine you initially quoted suggested there's little societal benefit to AVs, and that companies developing AVs seem to effectively be making them for lazy people.

To which of these points does the pushback apply (rhetorical or not), and whom exactly is doing the pushing?

None of those points. It was in response to the apparent implication that companies should be doing things for good over money. While I think that would be great on the face of it, I can practically hear my portfolio's death rattle at the mere mention of such a thing. It sounds nice, but it doesn't work. (Kind of like what may prove to be the case with AVs.)

As I indicated, there are companies trying to do good (for example, make inexpensive hearing aids using tech experience) while trying to make money. Now, I think they should have free reign to do this so long as everything is above board, but, and here's the bit about pushback, there have been groups--largely the existing providers of such products--fighting legislation that allows these tech companies to operate in this market. It's these [profit-serving] tactics that I believe are to the detriment of society.

I understand the concerns for what it may bring about, but much of my end of this recent discussion has been regarding the apparent implication that I referenced at the start of this post.

 

[homeforsummer]

None of those points. It was in response to the apparent implication that companies should be doing things for good over money.

Ah, that. Well any company will have its own motivations. My commentary is more theoretical - AVs are the answer everyone seems to be focusing on, but they're not really addressing the root problems they're purported to solve.

A less ludicrous comparison than my earlier toe-stubbing example might be that everyone seems to want to make all urban traffic electric. It's a noble aim, and would improve urban air quality, but it's not really a solution to the root problem, which is that there's urban traffic in the first place. Give people options other than cars and you improve the air quality and reduce the time people spend sitting doing nothing, rather than just improving air quality alone.

The AV argument is very similar. We already have methods of transportation that allow you to scratch your arse or stare at your smartphone (or both simultaneously), and they're vastly more efficient in many of the ways that matter - but instead we (as a market) are trying to make our existing inefficient methods of transport drive themselves. Market-driven it may be, but as a species we're too dumb to focus on the bigger picture.

 

[GBO Possum]

Based on my experiences in Assassin's Creed: Origins, the ancient Egyptians had autonomous driving in development. As you can see from the video, it wasn't entirely bug-free

 

[Danoff]

Urbanistically, public transit is a better use of investment than AVs, which will essentially replace taxis.

Public transit can make good use of AVs as well. No driver pay means more bus routes, which has to be non-linear for ridership. Also it can mean smaller busses for those additional bus routes. I'm imagining a car that doesn't exist right now, but it would look something like a minivan with sliding doors for 4 sets of 2-rider compartments. You could fit 4 people completely isolated or up to 8 people in 2-person pairs (or anything in between) (keep in mind, no front seat for the driver). I could see a version for 6 sets of 2-rider compartments as well with larger size of some of the vehicles today.

Busses are the size they are because of an optimization between number of routes, vehicle cost, and cost per mile overhead (including driver's wages). Take the driver out, and the optimization changes, with more vehicles, smaller size, and more routes. Something like driver's wages, which goes up linearly with number of vehicles, takes a MASSIVE pressure off of the calculus, and can result in a big shift in the optimization point - to the point where there are no set routes, just real-time optimization across the fleet of vehicles constantly adapting to the daily demand. If you start having trouble satisfying demand, you're doing well so you add vehicles.

It does have the potential to put mass transit out of business entirely. But of course it also has the potential to reduce traffic by a factor of 2, 3, 4? That's a big difference in congestion and fuel consumption. If enough people adopt it, traffic could decline to the point where commute times drop significantly, opening up cross-town job access for people that didn't use to have those options. Overall that improves employment. Not needing to own a vehicle to commute reduces the number of garage spaces needed, which reduces the footprint of houses built and improves the desirability of higher density housing.

I still think the primary driver for AVs has to be fleet-based transit. The fact that it's marketed as a cool gadget for super expensive vehicles is an important foot in the door.

 

[Eunos_Cosmo]

Public transit can make good use of AVs as well. No driver pay means more bus routes, which has to be non-linear for ridership. Also it can mean smaller busses for those additional bus routes. I'm imagining a car that doesn't exist right now, but it would look something like a minivan with sliding doors for 4 sets of 2-rider compartments. You could fit 4 people completely isolated or up to 8 people in 2-person pairs (or anything in between) (keep in mind, no front seat for the driver). I could see a version for 6 sets of 2-rider compartments as well with larger size of some of the vehicles today.

Busses are the size they are because of an optimization between number of routes, vehicle cost, and cost per mile overhead (including driver's wages). Take the driver out, and the optimization changes, with more vehicles, smaller size, and more routes. Something like driver's wages, which goes up linearly with number of vehicles, takes a MASSIVE pressure off of the calculus, and can result in a big shift in the optimization point - to the point where there are no set routes, just real-time optimization across the fleet of vehicles constantly adapting to the daily demand. If you start having trouble satisfying demand, you're doing well so you add vehicles.

It does have the potential to put mass transit out of business entirely. But of course it also has the potential to reduce traffic by a factor of 2, 3, 4? That's a big difference in congestion and fuel consumption. If enough people adopt it, traffic could decline to the point where commute times drop significantly, opening up cross-town job access for people that didn't use to have those options. Overall that improves employment. Not needing to own a vehicle to commute reduces the number of garage spaces needed, which reduces the footprint of houses built and improves the desirability of higher density housing.

I still think the primary driver for AVs has to be fleet-based transit. The fact that it's marketed as a cool gadget for super expensive vehicles is an important foot in the door.

I like the idea of AVs applied to public transit. To me it makes more sense than more simply automating everyone's commuting vehicle, without any sort of paradigm shift in how commuting works. Wasn't Uber exploring this? Or in the very least, using their transit-optimization technology to help cities improve public transit?

 

[Danoff]

I like the idea of AVs applied to public transit. To me it makes more sense than more simply automating everyone's commuting vehicle, without any sort of paradigm shift in how commuting works. Wasn't Uber exploring this? Or in the very least, using their transit-optimization technology to help cities improve public transit?

I think that one of the things that these concepts seem to be lacking is the recognition that people don't actually WANT to share space with other people during their commute. They want to be by themselves, maybe on the phone, or having a personal conversation that isn't overheard. Compartmentalizing these vehicles (with separate doors, because weather), makes a lot more sense to me. Maybe one large compartment for a larger group on some of the vehicles. But surely for commuters a 2-person compartment would be one of the biggest draws. I'd say 1-person but it seems a little claustrophobic.

 

[Eunos_Cosmo]

It has begun.

Sort of.

Waymo, Google’s self-driving car project, hit another milestone on Wednesday by launching an Uber-like service in Arizona that’s driven by a fleet of autonomous taxis. But there’s some caveats to the service.

The service, dubbed Waymo One, will ferry a group of riders that have participated in the company’s autonomous vehicle pilot program since 2017 around a roughly 100-square-mile zone across four cities in the Phoenix area—Chandler, Tempe, Mesa, and Gilbert. The catch is that, now, they’ll be paying fares. And as Waymo describes it, the service works a lot like Uber and Lyft.

Riders download an app, provide a credit card number, and plug in a destination. Images of the app, published by Waymo on Wednesday, show a setup that’ll be familiar to users of the other, better-known taxi startups.

Also, and I didn't see this before yesterday, Waymo has apparently given up on autonomous driving level 3 as they say it is impossible for the same reason I thought it was impossible.

Google/Waymo ceased the testing of so-called Level 3 self-driving vehicle systems — where drivers are expected to take control quickly when the need arises — after some drivers literally started falling asleep during testing, the company’s CEO John Krafcik has revealed.

 

[Eunos_Cosmo]

Very interesting article published by waymo regarding the limitations of deep learning as applied to autonomous driving. Waymo seems to be so far ahead of everyone else...yet they seem to acknowledge there is still so much further to go. Makes me wonder how safe the consume-level systems can possibly be right now.

While deep learning has enjoyed considerable success in many applications, handling situations with scarce training data remains an open problem. Furthermore, deep learning identifies correlations in the training data, but it arguably cannot build causal models by purely observing correlations, and without having the ability to actively test counterfactuals in simulation. Knowing why an expert driver behaved the way they did and what they were reacting to is critical to building a causal model of driving. For this reason, simply having a large number of expert demonstrations to imitate is not enough. Understanding the why makes it easier to know how to improve such a system, which is particularly important for safety-critical applications. Furthermore, if such improvements can be performed in an incremental and targeted manner, a system can continue learning and improving indefinitely. Such continual lifelong learning is an active field of research in the machine learning community.