Robo-Cab Predictions

[Brumby]

I don't think the EV market will go away like that.  Certainly, it will lose a high profile player, but there are other manufacturers who are in the game to some degree - and with some bold legislative targets in place in some areas and the general tendency away from the love affair with petroleum fuels, it will continue.  There will be an impact, but it won't be terminal.

[Red Squirrel]

Yeah hopefully.  I really want to see EVs become the future.  I don't think it will do much to solve climate change on it's own, but it will at least help.

[james_s]

I also think fully autonomous vehicles are a long way off. Modern computers still have a difficult time of identifying an email as spam, something that a human of average intelligence can determine at a glance with nearly 100% certainty. They work on carefully mapped urban courses but the real world is full of edge cases. Every day I see debris in the road, spilled paint, sand, worn off markings and all manner of other things that will confuse the heck out of an autonomous system. That's before you even consider all the impatient/malicious people who will do things like not allow a vehicle to merge. You have to be pretty aggressive in a lot of areas to merge onto a highway, I can picture it now, this poor self driving car inching timidly forward waiting for a gap in traffic until it runs out of lane and helplessly stops.

[magic]

That's before you even consider all the impatient/malicious people who will do things like not allow a vehicle to merge. You have to be pretty aggressive in a lot of areas to merge onto a highway, I can picture it now, this poor self driving car inching timidly forward waiting for a gap in traffic until it runs out of lane and helplessly stops.

A simple solution is to kick humans off the roads and equip cars with some "fair" protocols to negotiate merging. Futuristic crazies in Silicon Valley are already talking about banning normal driving because robots will ultimately become safer, never get tired or drunk, can be made not to speed
I don't think it's coming anytime soon, but in states saturated with Tesla snobs, who knows

[Algu607]

The argument that autonomous cars don't get drunken, tired or drugged is a weak argument. I don't want a car that statistically crashes less often then humans but fails to handle rare situations a human would have handled. Imagine a tree just about to fall on the street during a storm. A human would see that tree leaning over and stop. But what about an autonoumous car? Autonomous cars need to be able to deal with all situations a human can and more before they can be accepted.
So it is not a numbers game. The benchmark must be an intelligent, not speeding, fully aware and sober human driver.

[Algu607]

We also have to consider how incredibly good and fail save the human neural network actually is. Most (probably 99%) of drunken drivers don't crash. So we humans are still able to drive even if our system is severly impaired. Imagine a self driving car with ram chips falling of the board, lens of the camera defocussed and lidar giving wrong distance readings.

[wilfred]

Uber is going out of business and they will finish doing so before they develop an auto. That's my prediction.

At least with Uber you know why an auto is a good idea. They have a business model that exploits the drivers and they know it.

[Electro Detective]

I'd like to see how Robo-Cab Inc. deals with car jackers stripping them for parts and or fun,
and how that will impact fare prices

and how a halfassed AI box on wheels deals with back seat rapists, muggers, pedo-nappers and heist getaways, just for starters     

Too much automation = unnecessary complications, drama, unemployment,
less human to human interaction, 
and or a lazyass population that can't be rooted to turn a steering wheel and watch where they are headed. 

Bring back the horse and carriage and lose these wet dream tech headaches that are unlikely to ever get 'smart enough' to be SAFE safe

All that aside, I can't imagine anything more boring than riding in one of those POS   

[Ice-Tea]

Anyone talking next year is kidding themselves.

Weed smoking Musk has promised exactly that, putting them on the streets next year. Yeah, another visionary thing. With cannabis the term "visionary" takes a new meaning.

He promised 1 000 000 of them in 2020. Which is kinda funny because he's a long shot from producing even 1mio regular Teslas/year.

[EEVblog]

I would argue that real autonomous driving (in a real world traffic situation you could encounter today in a car) could only work reliably with real artificial intelligence (and not some brainless neuronal network giving somewhat nondeterministic results) so it could react to unexpected situations at least as good as a human driver.
So even ten years from now sounds pretty much optimistic unless there is some unexpected breakthrough in AI.

Yep, in all the hype this gets forgotten, they aren't even close to what a human driver can do in terms of the unexpected. Won't even happen in 10 years, I'm calling that one too.

[apis]

Unless there is some major mishap I predict Waymo are going to begin driving without safety drivers in Phoenix within the next 5 years. It will take much longer until we see them here in Sweden though.

This is where they drive today (with safety drivers):

Waymo One riders are currently able to take rides within parts of the Phoenix metropolitan area, including Chandler, Tempe, Mesa and Gilbert. Our service operates around the clock, seven days a week. Riders can travel anywhere in our territory day and night. Our goal is to give our riders everyday access to our cars, and we encourage them to take rides as frequently as possible.

[splin]

'Driverless Cars: On a Road to Nowhere (Perspectives) Christian Wolmar'

Driverless cars are the future - just around the corner. That is what the tech giants, the auto industry and even the government want us to think. But closer inspection reveals that we are much further from that driverless utopia than we are led to believe by newspaper headlines and by press releases from firms with vested interests. Christian Wolmar argues that autonomous cars are the wrong solution to the wrong problem. Even if the many technical difficulties that stand in the way of achieving a driverless future can be surmounted, autonomous cars are not the best way to address the problems of congestion and pollution caused by our long obsession with the private car. This entertaining polemic sets out the many technical, legal and moral problems that obstruct the path to a driverless future, and debunks many of the myths around that future's purported benefits.

Preview: https://books.google.co.uk/books/about/Driverless_Cars_On_a_Road_to_Nowhere.html?id=aZCCDwAAQBAJ&printsec=frontcover&source=kp_read_button&redir_esc=y#v=onepage&q&f=false

[apis]

Double hell - I'm still waiting for a house robot that will fetch me a beer and find the TV remote after 50+ years of promises from 'Tomorrows World' etc. How hard can that be?

It's pretty straight forward actually, and we're getting there. The biggest problem until recently have been computing power, but that has changed now. The problem now is only the massive development cost and the cost of all the electromechanical parts, servos, and sensors. It just isn't worth it simply for a curiosity that can fetch you a beer. But there are other areas where an agile robot would be very useful (like military applications) so it is under development. Once the robotic platform has been developed and is being mass produced enough to get down costs we might begin seeing domestic robots doing simple tasks.

I'm sure you have seen these before:



"We have begun field testing the Spot robot for commercial usage around the world. After an initial mapping run, Spot autonomously navigated two dynamic construction sites in Tokyo and used a specialized payload for surveying work progress. An additional camera in its hand lets Spot do even more detailed inspection work on site. The Spot robot will be available in the second half of 2019 for a variety of applications."

For as long as I can remember there has been no shortage of headlines trumpeting major scientific breakthroughs every few months. But almost invariably, after the initial excitement, it turns out that they've only developed the first 90% and all it requires for the last 10% are a few minor wrinkles to be ironed out, somebody to decide the best shade of blue to paint it, a solution to the Goldbach conjecture and the discovery of some new materials which can withstand 5000C whilst being a bit stronger than graphene at half the weight.

The difference is that this technology have been demonstrated for many years now, there are several different companies developing such self driving systems and companies with very deep pockets are pouring a lot of money into it. I find it a bit weird so many are in denial about it.

You might not care but almost everybody else will. People tend to get very excited about deaths caused by things out of people's control and very quickly start baying for blood/something to be done/heads to roll. Rational consideration of cost - benefit tradeoffs won't be at the fore-front of most observer's minds (unless they happen to be engineers). When a human f**ks up and kills someone it's an accident. When something manufactured f**ks up and kills someone because of a design defect it's a company/corporation head that has to roll. Laws will have to change radically to allow autonomous vehicles to operate without the threat of company destroying liability costs arising from every incident involving human injury or death.

You might be right about that, people are irrational, and I won't pretend I can predict what the psychological reaction will be.

There are plenty of youtube videos showing the difficulties and problems that self driving cars get into for those interested.

Really, they are not showing up in my feed. I would love to see videos showing what Waymo are having the most problem with today.

[apis]

I also think fully autonomous vehicles are a long way off. Modern computers still have a difficult time of identifying an email as spam, something that a human of average intelligence can determine at a glance with nearly 100% certainty.

The border between spam and not spam is not black and white, what you might consider spam someone else might consider a treat. So identifying spam is a very different problem, reading and understanding text requires a semantic understanding of human life on whole different level, and neither natural language processing nor AI is good enough for that yet. Even so, using surprisingly simple methods, I find that Thunderbird is pretty effective at classifying spam, as are gmail.

They work on carefully mapped urban courses but the real world is full of edge cases.

Does Phoenix count as the real world?
"Waymo One riders are currently able to take rides within parts of the Phoenix metropolitan area, including Chandler, Tempe, Mesa and Gilbert. Our service operates around the clock, seven days a week. Riders can travel anywhere in our territory day and night."

Every day I see debris in the road, spilled paint, sand, worn off markings and all manner of other things that will confuse the heck out of an autonomous system.

Last year, Waymo's prototypes drove the equivalent of a year (for an average american) between disengagements in California. I'm convinced none of the things you mentioned would be a problem for their cars. Undoubtedly there are still some unpredictable things that will cause issues, but It's really hard to come up with an example since if we could they wouldn't be unpredictable. What matters is that the problems are rare enough that they are at an acceptable level. Perfection is impossible, also for human drivers.

That's before you even consider all the impatient/malicious people who will do things like not allow a vehicle to merge.

People are nuts, and maybe malicious people will be a problem, but there will be real people in self driving cars as well, and most people aren't sociopaths.

You have to be pretty aggressive in a lot of areas to merge onto a highway, I can picture it now, this poor self driving car inching timidly forward waiting for a gap in traffic until it runs out of lane and helplessly stops.

Actually that kind of thing has been mentioned before. Similar problems arise in many situations, e.g. when making a turn in an intersection you can't just timidly sit and wait since people will never let you drive, you have to start driving and signal that you are going before others let you pass. But they have solved such problems before (e.g. with intersections) so even if merging onto a highway is problematic (I don't know if it actually is, they wouldn't tell us if it was) I think we can expect it to be solved relatively soon.

There's another problem though, sometimes driving in the real world requires you to break the law. I.e. the law might dictate that you drive very timidly but in practice people routinely break it or they wouldn't get anywhere. What should the programmer do in that case? Program the cars to break the law routinely?

Something they have admitted is very difficult is figuring out the best place to stop to pick up/drop of passengers. They can't just stop and block all traffic on a crowded parking lot for example, or other drivers would become pretty annoyed.

[nctnico]

I would argue that real autonomous driving (in a real world traffic situation you could encounter today in a car) could only work reliably with real artificial intelligence (and not some brainless neuronal network giving somewhat nondeterministic results) so it could react to unexpected situations at least as good as a human driver.
So even ten years from now sounds pretty much optimistic unless there is some unexpected breakthrough in AI.

Yep, in all the hype this gets forgotten, they aren't even close to what a human driver can do in terms of the unexpected. Won't even happen in 10 years, I'm calling that one too.

I think you are wrong here and it is time you back these kind of claims with some factual information. You should also look at how human drivers can't cope with unexpected situations as well so there has to be at least a single standard as a benchmark. Fortunately there is one... In another thread someone pointed out the SAE has defined 5 levels of autonomy and Waymo (and at least one other but certainly not Tesla) is already at level 4. And they are going to start doing tests without the safety driver which means getting at level 5 (which is full autonomy) is very close. Several years at most and certainly not a decade.

[nctnico]

I don't think the EV market will go away like that.  Certainly, it will lose a high profile player, but there are other manufacturers who are in the game to some degree - and with some bold legislative targets in place in some areas and the general tendency away from the love affair with petroleum fuels, it will continue.  There will be an impact, but it won't be terminal.

Tesla isn't the biggest player in the EV market today. Not by a long shot. And I'm pretty sure EVs will go away in a few years. They only sell in artificially created markets. In the Netherlands the politicians are slowly waking up to recognise the fact that EVs simply aren't affordable to most of the people and the CO2 reductions are minimal at best. The latter means that EV subsidies (which already have been reduced) are a large waste of tax payer's money. In the Netherlands sales of the  relatively expensive Tesla Model S & X, Jaguar I-pace, etc have already dropped to near zero because the subsidies on these have been cancelled. What is left are cars with a very small range. On the topic of Tesla: Tesla has released a very short range (cheap battery) version for (AFAIK) the Canadian market in order to have their cars eligible for subsidies. Needless to say that a limited range doesn't help broad adoption of EVs due to lack of usefullness as a replacement for a regular car.

Edit: link to article about Tesla's short range model: https://www.cnet.com/roadshow/news/tesla-model-3-93-miles-range-canada-ev-incentive/

[apis]

'Driverless Cars: On a Road to Nowhere (Perspectives) Christian Wolmar'

Driverless cars are the future - just around the corner. That is what the tech giants, the auto industry and even the government want us to think. But closer inspection reveals that we are much further from that driverless utopia than we are led to believe by newspaper headlines and by press releases from firms with vested interests. Christian Wolmar argues that autonomous cars are the wrong solution to the wrong problem. Even if the many technical difficulties that stand in the way of achieving a driverless future can be surmounted, autonomous cars are not the best way to address the problems of congestion and pollution caused by our long obsession with the private car. This entertaining polemic sets out the many technical, legal and moral problems that obstruct the path to a driverless future, and debunks many of the myths around that future's purported benefits.

Preview: https://books.google.co.uk/books/about/Driverless_Cars_On_a_Road_to_Nowhere.html?id=aZCCDwAAQBAJ&printsec=frontcover&source=kp_read_button&redir_esc=y#v=onepage&q&f=false

I'm not allowed to read that. I doubt self driving cars will be something you own privately any time soon, so it seems like he got the premise wrong. They will be operated as robo-cab services first, but the same computer system could drive normal busses if you think busses are the future. It certainly will be driving trucks as well as vehicles in dangerous locations like mines, etc. I really don't get why some people think automation is a bad thing.

[schmitt trigger]

People will use them for sex, drinking, and leave all sort of garbage behind.

There is actually a viral video of two people  doing exactly that on a Tesla on autopilot.

Won't post it here, it actually qualifies as porn.

[schmitt trigger]

Yeah hopefully.  I really want to see EVs become the future.  I don't think it will do much to solve climate change on it's own, but it will at least help.

There is an overlooked pollution type:  noise pollution.
EVs will definitively help.

[apis]

Tesla isn't the biggest player in the EV market today. Not by a long shot. And I'm pretty sure EVs will go away in a few years. They only sell in artificially created markets. In the Netherlands the politicians are slowly waking up to recognise the fact that EVs simply aren't affordable to most of the people and the CO2 reductions are minimal at best.

All the well-to-wheels analyses I've seen shows that the CO₂ reductions are substantial except if they are purely charged by coal power plants, but most countries have a lot of nuclear and renewables in their energy mix.

There are actually many benefits with EV's for a robo-cab service. If they mainly operate in a metropolitan area the shorter range isn't a problem, and it's also not a problem to take cars out of service for charging regularly. They will get lower fuel costs, and be environmentally friendly which is good pr (besides the obvious advantages for the environment and smog levels). You also utilise the cars better, they won't just collect dust in a garage but will be used until they wear out, so fewer cars can service more people.

[james_s]

The Waymo routes in Phoenix are carefully curated routes that are mapped to incredible levels of detail and Phoenix is flat with a dry climate. It's nothing like being able to drive to arbitrary locations throughout the nation. The area where I live for example is mountanous with lots of twisty crowded roads and complex confusing intersections that don't meet at right angles. There is a lot of inclement weather, rain, hail, the occasional dusting of snow. The roads are very crowded and getting worse all the time, I think there's a good reason none of the self driving testing is going on here despite the tech industry being second only to Silicon Valley.

[james_s]

Tesla isn't the biggest player in the EV market today. Not by a long shot. And I'm pretty sure EVs will go away in a few years. They only sell in artificially created markets. In the Netherlands the politicians are slowly waking up to recognise the fact that EVs simply aren't affordable to most of the people and the CO2 reductions are minimal at best.

All the well-to-wheels analyses I've seen shows that the CO₂ reductions are substantial except if they are purely charged by coal power plants, but most countries have a lot of nuclear and renewables in their energy mix.

There are actually many benefits with EV's for a robo-cab service. If they mainly operate in a metropolitan area the shorter range isn't a problem, and it's also not a problem to take cars out of service for charging regularly. They will get lower fuel costs, and be environmentally friendly which is good pr (besides the obvious advantages for the environment and smog levels). You also utilise the cars better, they won't just collect dust in a garage but will be used until they wear out, so fewer cars can service more people.

No sense in debating that with some folks. He has a pathological hatred for EVs and shows up in every single thread discussing anything to do with them and relentlessly poo-poos them with a never ending arsenal of strawmen, false equivalency and baseless arguments.

[Electro Detective]

Speaking solely as a realistic tech fanperson:
I reckon all the '..way of the future' dreaming and waste of R+D dollars isn't going to bring about the advanced products seen on sci-fi shows 

Yes, a few novelty clunkers may come about that will come and go (probably have already? )  and be forgotten

Money better spent on endeavors to benefit humanity, shareholders, CEOs,   
and upgrades/updates/patches and psychiatry R+D for the current robo-zombies,
walking around with phones glued to their faces   

 

[apis]

The Waymo routes in Phoenix are carefully curated routes that are mapped to incredible levels of detail and Phoenix is flat with a dry climate.

They don't have special routes, they operate everywhere within "the Phoenix metropolitan area, including Chandler, Tempe, Mesa and Gilbert." (It's supposed to be about 100 square miles or 259 km²). They use detailed maps but they claim they are able to create those maps and Alphabet certainly knows a thing or two about creating maps, so why not?
Phoenix is apparently flat with a dry climate and sympathetic laws. It makes sense to launch your pilot project in an as ideal location as possible.

It's nothing like being able to drive to arbitrary locations throughout the nation.

I feel you are moving the goalposts now. They have to begin somewhere.

The area where I live for example is mountanous with lots of twisty crowded roads and complex confusing intersections that don't meet at right angles. There is a lot of inclement weather, rain, hail, the occasional dusting of snow. The roads are very crowded and getting worse all the time, I think there's a good reason none of the self driving testing is going on here despite the tech industry being second only to Silicon Valley.

It's not the twisty crowded roads, it's the rain, hail and snow. The lidar is famously blinded by heavy precipitation.

[wilfred]

The difference is that this technology have been demonstrated for many years now, there are several different companies developing such self driving systems and companies with very deep pockets are pouring a lot of money into it. I find it a bit weird so many are in denial about it.

I'm as puzzled by those who seem to think autos are anywhere close to reality. Even if they can drive well and if they can park in a space, how will they go finding a space to park? Is this even part of the programming? I know I use lots of clues to find a spot including looking to see if someone is in the drivers seat and may be about to put their seat belt on. I also look at people with a shopping bag and stalk them in case they are leaving. The cars have to be able to recognise parking signs to ensure they don't park and get a ticket. Who gets the ticket?

On the issue of who gets the ticket, that extends to who is going to accept responsibility for loss in the event of an accident? That's the road you have to navigate successfully before autos gain mass acceptance.

Going back to the parking issue. Yesterday I was getting into a lift (not Lyft) and the doors were painfully slow to close, or so it seemed because no-one else was around. In reality it was about 5 seconds. A lift still can't even tell no-one is waiting. It still takes a timeout to elapse on the door entry beam not being broken. How long will someone tolerate driving around whilst the car seeks a spot and is outbid by human drivers better able to find a space? Take the steering wheel out of a car that drives well enough to not require a passenger with a drivers license and the car will have to park itself somewhere. This aspect of driving has to be dealt with. It just has to be, but I have never seen any auto do it. Or even attempt to do it.

Talk about denial that's a place to start.