Lol is this your first day? You don't get to be in the office less because you worked during your commute.
Self Driving Car thread
- Thread starter ChaoticUnreal
- Start date
Oh I'm not actually expecting it to change to accommodate that but if that was the option I wouldn't mind working on my way in.
People said that about work at home too. Now we track hours and just fill up the 38 per week and no more. Once SDC is common you'll be able to clock extra hours during commute at better companies, and the ones that require too many hours in the office will be at disadvantage for recruiting.
Yeah, I'm repeating a few others in the thread but it really is important to keep hammering down on that argument. No, you will not get an appreciable amount of work done in an SDC and no, you will not be allowed to work less because you worked in the car. This aspect of time efficiency of SDCs is entirely fictional. It falls apart on every facet:
SDCs aren't some potential hidden productivity-increasing factor, if anything they're just going to mean more expensive commutes for everybody as they crowd out proper transit.
- Only a small fraction of people get any work done in transit, even if they have ample seating, a desk and power.
- The whole point of going to a workplace is to benefit from economies of scale and the network effect as well as associating a place with a task. Your productivity in a car on the way to work is not going to be significant
SDCs aren't some potential hidden productivity-increasing factor, if anything they're just going to mean more expensive commutes for everybody as they crowd out proper transit.
Even if I can't count work hours while commuting as actual worked hours the SDC would still allow me to do other things besides driving.
I carpool and regularly grab a laptop on the passengers seat, put phone on hotspot can work just fine. Either work locally (browser, IDE all available) or phone home via Remote Desktop.
You can do more on a laptop than on a phone, there catching up on mail, light research and some light ticket triage is all you can do.
SDC on private cars make it so I can do this 100% of the time instead of only when I'm not driving.
In public transport there actually is a bit difference between max density screw comfort mode like the busiest metro lines and 1st class trains here. You have a decent tray table and there are even some power sockets available. It just depend on what you aim for.
You could go for maximum capacity on a short line, so mostly standing room only, and yes that's an actual production self driving shuttle with a call button at stops:
Or equip it like economy aircraft cabin and have only seating room and a tray table for everyone, for longer trips. This is required for productivity and probably costs extra, like Uber vs Uber Black.
Last edited:
I would expect to get the same amount of work done as on an Uber because that is functionally what the types of SDCs that actually work are. They're an Uber where the vendor is trying to invest in machinery to avoid paying the driver. I get work done on trips when I take an Uber somewhere, but not everyone works in cars.
When I have taken Waymo vehicles I could have pulled out a laptop, but did not because I was watching to see how the car would handle intersections, bikes, etc and not really thinking about work. I assume if this technology becomes mundane I would probably answer emails.
When I have taken Waymo vehicles I could have pulled out a laptop, but did not because I was watching to see how the car would handle intersections, bikes, etc and not really thinking about work. I assume if this technology becomes mundane I would probably answer emails.
I view the benefit of self driving like a blue light filter on a computer screen. It reduces stress and allows you to be more alert once you get to the destination
With Waymo apparently preparing to start highway driving in the southwest, I'm curious about long haul trucking. If you can drive freeways in California, Texas and Arizona, you could do a fair bit of interstate through the southwest that connects the middle of the country to Atlantic and Pacific ports. Has anyone looked at the economics of that recently? Compared to Uber drivers, long haul (or better yet team drivers since there are two of them per truck) are much better paid, but I don't know how the costs break down there.
Aurora is actually launching no-driver, on-highway operations in the next few months.
An argument against self driving haulers is the weight, it can get very very messy if it fucks up. Plus the driver / captain is responsible for securing the load and checking the vehicle combination for safety issues.
The liability if anything goes wrong is a huge concern. How would you set insurance rates?
What do they mean by "cyber incident?" It seems like an intentionally vague term when you apply it to self-driving trucks. It could mean anything from a denial of service attack or ransomware on the conventional end to the more applications specific attacks or fuckups including misrouted traffic, electronically enabled theft of cargo, intentionally or unintentionally crashed vehicles, or blocked roadways. Or murder.
There's things like container hauling where the port loads the truck via crane (probably not operated by the driver). Google says ~ 1.5-2 million containers per month moved via truck. Seems like a promising market for automation but I bet theres all kinds of nasty details in the way.
Waymo already has 3rd party insurance on their taxis. Probably not that hard to negotiate when you're google and have more liquidity than the insurance provider. Overtime as you get data on accident rates presumably the market can set prices, same as they do now for truckers. I'd expect this to be quite competitive since trucking insurance also have to cover catastrophic events already:
https://wilsonlaw.com/blog/walmart-settles-with-comedian-tracy-morgan-for-estimated-90-million/
Over $100m to settle a single accident due to an under-rested walmart driver.
Around here (NL/DE) with platooning, the big problem is still just cost. Self-driving trucks are considerably more expensive and have a lot of extra (hidden) labor involved in them, so there's currently no cost or labor savings. It's also not the purpose of these trials and licenses right now - the first step is to gain trust in the system and figure out the kinks. It's a very, very slow process of getting companies to have confidence in the concept, developing systems to incrementally remove all the non-driving labor and end up with a system that does autonomous driving the way it has been promised. A bit like the Rotterdam harbour authority: they took about 20 years to go from on-dock cargo starting to get automated to having reduced >99% of the labor force for - basically - longshoreman-type tasks.
There's also been a lot of discussion academically about the real achievable cost reduction. Drivers aren't that expensive, nor are there that many opportunity costs to using human drivers on most routes. Most arguments around automation don't apply to trucking, e.g. most trucks can't drive when the driver is sleeping... well, yeah most suppliers and customers are also closed at those times. Only trucker shortages are a good argument right now.
If that were taken as a standard for accidents involving shipping trucks, companies would stop shipping stuff by truck.
I think it really depends on the location and type of drivers being replaced. Tandem long haul drivers in the USA for example make 100k+ and you're paying two of them. Can pay off the lidar pretty quickly at that rate.
View: https://www.reddit.com/r/Truckers/comments/1aqmp97/what_is_the_expected_avg_salary_for_married_team/
But a local delivery driver for Amazon is making a tiny fraction of that. I think it boils down to what fraction of routes you could service are actually paying a lot for labor. I have no idea how to figure that out.
Well here DHL driving for Amazon spends a lot of time moving his van a few meters and getting out again. A SDC limited to 20 kph would do just fine, stop, let guy grab the next few packages, close door, part ways, guy goes to the next few doors while the van moves itself a bit further.
The more relevant point is that it likely isn't worth automating that even when the tech is ready because the labor savings is so small.
Because he has to or because he'd rather drive his van a few meters than walk a few meters?
edit: see @redleader's point above. You're describing saving a fraction of the labor of one person. My guess is a small fraction.
I agree with most of what you wrote but this line (I personally likely wouldn't work much in an SDC unless I was commuting during a meeting - I'd probably sleep). From a US perspective, SDC should have enormous benefits to commute time and cost due to crowding out idiots/incompetents. I'm guessing you're in Europe, where density allows public transportation / mass transit options to be viable.
Here in the US, having orderly masses of SDCs taking people to work/school would result in far better traffic flow and dramatically fewer accidents.
So the mainstream traffic engineering thinking is that that doesn't actually work that way. The existence of a hypothetical future cheap SDC service would mean the induced traffic will always outpace whatever labor or time saving is caused by a new car technology.
What people choose to use for transit is determined almost analytically by the time and money cost of the mode. So if SDCs truly become cheaper in that sense, even if they're a bit more expensive in money but cost way less time, people will move there up to the point where the cost is at parity with other modes. That's how induced demand works.
And from a traffic engineering point of view, SDCs are still cars, and cars are very inefficient users of infrastructure. They're going to gum up the streets just as bad as regular cars, but at a higher financial cost (because they can trade on time savings).
That's roughly the idea behind skepticism towards the promises of SDCs from every traffic engineer I've heard on the topic.
What people choose to use for transit is determined almost analytically by the time and money cost of the mode. So if SDCs truly become cheaper in that sense, even if they're a bit more expensive in money but cost way less time, people will move there up to the point where the cost is at parity with other modes. That's how induced demand works.
And from a traffic engineering point of view, SDCs are still cars, and cars are very inefficient users of infrastructure. They're going to gum up the streets just as bad as regular cars, but at a higher financial cost (because they can trade on time savings).
That's roughly the idea behind skepticism towards the promises of SDCs from every traffic engineer I've heard on the topic.
Once again, from an EU perspective this makes perfect sense.
From a US perspective: What other modes?? There are no 'other modes' here, outside of a few places like NYC. Everyone already drives everywhere. We've already built the car infrastructure. Everyone has a car. I fail to see the downsides of replacing idiodic / incompetent / sleepy / distracted / drunk / etc bags of wetware with SDC tech (assuming mature tech, of course.)
You're putting the SDC transit spectrum up against humans doing the same thing. There are no other alternatives in the US to that, and outside of pure science fiction there's not going to be - the cost is simply too high in this huge, sprawly place. Cars are going to 'gum up the streets' regardless - do you want them driven by humans or machines? Humans are collectively terrible at it.
OK I get that it's hard to imagine other modes, but SDCs and human-driven cars are two modes. Those will be competing. Likewise, you can actually introduce successful transit in the US as well. If we're considering SDCs as the future, it seems kind of weird not to consider a similar investment into transit as well, especially considering SDCs will not solve most of the issues plaguing US traffic infrastructure.
I don't want to be too academicbrained but consider that cars also cost money, a whole lot actually. Mass transit is almost always a cheaper option, so if people are given an informed choice, it's not that hard to transition over to transit-oriented urbanism. Not in short order, this takes decades, but it's perfectly possible to be done. The Netherlands did the switch in about 20 years.
And to your last question: with the technology we have right now, I have no expectation that SDCs will arrive anytime soon. Traffic interactions are a social system, and removing the human element from traffic means you're losing a whole lot of what manages smooth and effective traffic flow. Maybe a utopian perfect SDC would be preferable to humans, but I'd be very interested to see if that has any cost advantage anymore. Framing it as cars being driven by either humans or machines is too simplistic, it doesn't work that way at all.
I don't want to be too academicbrained but consider that cars also cost money, a whole lot actually. Mass transit is almost always a cheaper option, so if people are given an informed choice, it's not that hard to transition over to transit-oriented urbanism. Not in short order, this takes decades, but it's perfectly possible to be done. The Netherlands did the switch in about 20 years.
And to your last question: with the technology we have right now, I have no expectation that SDCs will arrive anytime soon. Traffic interactions are a social system, and removing the human element from traffic means you're losing a whole lot of what manages smooth and effective traffic flow. Maybe a utopian perfect SDC would be preferable to humans, but I'd be very interested to see if that has any cost advantage anymore. Framing it as cars being driven by either humans or machines is too simplistic, it doesn't work that way at all.
Sorry, but I find your position on public/mass transit vs SDCs in the US to be completely unrealistic. Yes, if the US becomes something other than the US, and completely changes it's character, outlook, and national culture, mass/public transit might become viable for more than 1% of the population. You're talking about a transformation that's so breathtaking it's simply hard to even fathom. The US is a huge place that's mostly very rural, and land in most of the US is so plentiful and cheap cities don't massively urbanize - it's much, much cheaper and simpler to just build out suburbia, and put in more roads. And that's what happens every day, 365 days a year. Putting in effective mass transit in any city that doesn't have it now is TRILLIONS of dollars, and decades of time and investment to get even minimal coverage for a small percentage of the local population due to the sprawl.
Don't get me wrong personally, here - I much prefer your vision. If I could afford to move to London I'd be there yesterday. Same for Paris, Amsterdam, etc. I adore urban living and not having to own a car, and in a few years (unless the current admin destroys us all here financially) I hope to transition to that in retirement.
It'll be in the EU, though. The people here in the US have voted for this car-based sprawl repeatedly and fanatically. They vote for it formally with urban mass-transit projects being incredibly difficult to get passed and funded, and they vote for it with their buying/living behavior.
SDCs are expensive, sure, but in the US there's no way they're more expensive than mass transit without radically changing how US people live and work. I also completely concede your social interaction point about traffic - successful integration of large numbers of SDCs in US urban/suburban environments will mean changing how we do road construction, signage, repair, and other things. SDC tech and adoption of SDC-friendly road design in major cities is a matter of 20-30 years, though.
Mass transit benefiting more than 1-2% of the US is a matter of centuries without radical cultural change.
Don't get me wrong personally, here - I much prefer your vision. If I could afford to move to London I'd be there yesterday. Same for Paris, Amsterdam, etc. I adore urban living and not having to own a car, and in a few years (unless the current admin destroys us all here financially) I hope to transition to that in retirement.
It'll be in the EU, though. The people here in the US have voted for this car-based sprawl repeatedly and fanatically. They vote for it formally with urban mass-transit projects being incredibly difficult to get passed and funded, and they vote for it with their buying/living behavior.
SDCs are expensive, sure, but in the US there's no way they're more expensive than mass transit without radically changing how US people live and work. I also completely concede your social interaction point about traffic - successful integration of large numbers of SDCs in US urban/suburban environments will mean changing how we do road construction, signage, repair, and other things. SDC tech and adoption of SDC-friendly road design in major cities is a matter of 20-30 years, though.
Mass transit benefiting more than 1-2% of the US is a matter of centuries without radical cultural change.
The year is 2025. SDCs are serving plenty of satisfied customers in my area. They're not everywhere but it's not like they don't already exist and work.
I don't think self-driving with no human driver will ever be a thing. At the very least, a human driver also provides security for the cargo and he helps with mechanical breakdowns which an AI cannot do. I don't think an AI could winch down a load.
What AI self driving can do is substantially reduce driver fatigue. This improves safety and allows the driver to work more hours.
Same for rideshare
Even driving 40 minutes to work in hands free bluecruise was nice and relaxing.
I can see fully autonomous large trucks with no drivers under very specific circumstances, but as you and demultiplexer points out it might not be worth the cost play. Long haul interstate runs could have a form of truck-only ZOV lane where the truck doesn't mix with other traffic. This would functionally be like a light-rail transport line, just without the rails. Implementation cost would be much lower than new rail lines.
To your point, though, it would still likely be a better to have a human on board, even if he/she isn't driving the vehicle. The truck could drive to the limit of it's batteries or whatever fuel it's using if the onboard 'captain' is just doing oversight and isn't required to be hands on. From a quick bit of research it appears trucker salaries are as much as 40% of the operating expenses of commercial transport lines, so if you could stretch that over a lot more miles driven using SDC tech and a hands-off crew it might make the investment worthwhile.
Ooh, if time not driving at L4/5 counts as rest time you can like double the distance travelled in a day using it. Sleeping and eating in a moving truck is a bit more challenging but surely they will find a way around that.
The van has to move. Pick up the next few, ding dong x3, move truck, repeat. If you skip the drive the walk distance back and forth will keep increasing.
This < 20 kph manoeuvre is easily done by even something like a parking assist but the challenge now is to move it from L2 (driver required) to L3, dual 12V power bus, redundant brakes, and a program smart enough to know its limits / ODD. It's fine if it decides the next hop is too difficult and you need to do so yourself, or too much traffic to do it safely. Or stop prematurely. It's just an assist function like the summon feature on Tesla (was it ever moved out of beta / released?)
Ugh. Can you image a L4 Amazon van that has no driver and a giant package carousel inside. It pulls up to your house and you have to go get it from the pickup slot within 5 minutes or it leaves.
At least the van probably won't lie about where it is, or whether it attempted delivery.
Amazon's totally-not-an-employee delivery people around here span a range of diligence and reliability, from "puts each package in the locker, and photographs it there", to "throws it over the courtyard fence and calls it a day", to occasionally "eh, check it off now, and maybe I'll drop it off tomorrow".
Thankfully we get none of that nonsense here but we're deep in suburbia with a private road but not gated neighborhood.
I kinda like delivery to package lockers, a moving one would be funny but an unnecessary addition as I have one at 5 minutes walking distance.
Tiny packages that fit through the mailbox gets delivered at home, small and light ones packages: locker pick up please. Large ones, at home please, cba to drag a 2 months supply of diapers home on foot, or things for work, that's almost always bulky and/or heavy.
But I can see the self driving package locker being a great addition if you are outside of a densely populated area. If you can follow it on the map, you can catch it either at your stop or before/after if you missed it. Sure, you would need to walk/drive to it but if you have time, it beats waiting for round 2 tomorrow. It could even park at popular locations like village/town center with shops and library for a few hours a day, for places that don't have enough packages to get a stationary locker.
It's a learned skill. I recall when learning to drive, they showed us films -- it was that long ago -- where you're suppose to be on the lookout for kids who may chase balls that bounce out into the street in front of traffic or dogs also running out, with maybe kids chasing them.
That was long before all the various diversions which lead to distracted driving.
Completely fails at any apartment complex that has controlled access.