We are at the peak of the autonomous driving hype cycle, and the trough of disillusionment is going to wipe out a few players, and see the massive consolidation of others. These start-ups are very capital intensive and once the funding environment shifts, it's going to be a big problem for a lot of them.
The initial launches of autonomous taxi services will be underwhelming - relative slow, sometimes frightening (there will be accidents for sure), and with limited deployment areas in limited weather. The novelty will wear off quickly, and then the real work is going to start.
There are basically going to be two ways to survive it - be part of a larger organization that can shoulder a long-term R&D burden to go from 'toy' to 'real infrastructure'. I can see Cruise and Waymo making it that way, unless GM gets cold feet.
The other survivors will be either super lean like Voyage, who went straight to revenue generating niches like retirement communities and rely on being downstream of the technical innovation being done elsewhere (no in-house vehicle or sensor development to cut down R&D costs and move quickly).
Or else they'll be in niches like logistics (e.g. Peloton, Nuro) where the parameters of the game are different, and the structure is more B2B than B2C.
One other play is the autonomy-tech licensing structure like Aurora is trying, but that's a hard sell, especially since they're dealing with German automakers who (from first hand experience in this domain in particular) are clueless about autonomy.
This is an incredibly exciting, risky time to be a part of this new, emerging industry. It feels in many ways like the very very early P.C era, where everything is very much still in play. I'm glad I made the career shift to get there.
Without revealing too much, German automakers have expertise primarily in logistics, with some expertise in automotive design.
There is a vast pyramid of Tier-1 and Tier-2 suppliers that feed into that logistics chain. A lot of those suppliers are promising automakers that they can deliver either components that feed into autonomy, or else autonomy in full. This fragments the effort across literally hundreds of small teams, many of whom have little to no expertise at all in the necessary disciplines.
Additionally, every German automaker also has its own in-house autonomy team, and the relationships between in-house, contracting, and supplier teams is chaotic. Many managers see autonomy as another way to build a small empire and make a name for themselves, and it just results in a huge organizational snarl.
The fact that this is an area no one knows how to execute on yet since it's brand new, coupled with these gross inefficiencies, will mean autonomy efforts from the automakers will be stillborn. They will eventually learn from it though, acquire the right teams, and get on with putting it into their cars just as they're doing with EV.
You're mentioning startups that will be wiped out. The big players in this scene are companies like Google, GM, and Uber. Not startups by any stretch of the imagination.
There are many many startups working on some fraction of the autonomy stack (e.g. Aurora) or the entire vertical (e.g. Zoox). I call out several specific ones in my comment if you care to read it through.
And in 5 years they will be "just around the corner" again.
There's so many hurdles before it becomes something that is commonplace or accepted, that really it's decades around the corner.
The technical hurdles to get to level 4 and 5 are just the start. Once you get past that, there's going to be a whole slew of legal and public opinion challenges.
The first time an autonomous vehicle is involved in an accident that kills a few kids, you'll have a Challenger level probe to figure out what happened and probably a whole revamp of safety measures.
>> The first time an autonomous vehicle is involved in an accident that kills a few kids, you'll have a Challenger level probe to figure out what happened and probably a whole revamp of safety measures.
Although I have serious reservations about the safety of autonomous vehicles, or at least our ability to predict it (see earlier comment referencing a RAND corp study of the subject) I'm afraid that running over kids with an auto-car is not going to be such big news as it perhaps should be.
For instance- can you remember the first time in history when an automobile ran over a child, or even several children at once? Rhetorical question- you most likely can't. Because it was never such a big issue to go down in history as the "Minors Motor Massacre" or whatever. And there most likely was no Challenger-level probe into the whys and wheretofores of that first time, either.
Besides, remember that, for many, many decades people drove around in cars that had no provision for driver safety. For instance, seat belts were first adopted by big car manufacturers in the early '50s and made compulsory by law for the first time in the '70s. It's clear that most of the world cares a lot more about fast, efficient transport than about safe transport.
This transition, for most, is as casual as two people deciding who is going to drive. There’s simply a different driver available and the convenience of that will make the acceptance of autonomous vehicles happen very quickly
Gearheads will refuse to let go of their cars. They love driving and maintaining their vehicle. They'll argue that no robot can drive better than they can, no matter how wrong they are.
Imagine a world where police no longer pull you over for speeding, running red lights, or other traffic infractions. The legal fight will be huge, as those easy revenue sources suddenly disappear.
There are hundreds of more fights and scenarios like this we haven't even begun to imagine, that will tie up progress in the courts for decades.
We're going to need a concentrated force focusing on this tech as the next thing we are going to do, and able to get out in front of these issues. Right now, I see autonomous driving as a thing that a few groups are seriously working on, while everyone else takes a passive stance, or dips their toes into the idea as a passing fancy and not a serious milestone to tackle. We need to move it past a novel toy, and in to a serious undertaking that we all need to be vested in shaping the future of.
Because in my opinion, the automobile is the most efficient and prolific human killing machine we have ever invented. I feel autonomous driving is a step toward fixing that issue.
Actually, as someone who loves to drive and race (& has won regional championships in road racing, ice racing, autocross, karting), I can't wait for self-driving L5 cars to become the norm.
Driving on the public roads has long since ceased to be much fun. Traffic density is ever-increasing. The average drivers' skill level has deteriorated noticeably and their distraction level has increased such that 60% of them might as well be actively trying to kill you. Meanwhile the guys with the blue lights have much denser coverage.
Occasionally, far from the cities, I manage to find a brief stretch of fun, but they are ever more scarce. I've also given up my road bike and stick to mountain biking. And my observations aren't unique among those I know.
At this point, I'd much rather just be able to hop in my car and read as it gets me to my destination. Save the fun driving for the racetracks or off-road.
And yes, the loss of revenue to the Police Depts who spend their time catching the easy to measure infractions instead of the ones who actually create hazards will be delicious.
"Imagine a world where police no longer pull you over for speeding, running red lights, or other traffic infractions"
That's an interesting point I hadn't heard before. Are there stats for what kind of revenue various little cities, counties, etc, in the US are generating with speed traps, red light cameras, and similar tactics? Those tactics obviously lose to self driving cars.
No, it’s just the most boring conspiracy theory ever.
There are a few hamlets out west where the local sheriff and justice-of-peace got into the business of modern highway robbery, extracting fines from unsuspecting travelers. For other cities, fines aren’t nearly as significant as people love to think.
Case in point: fines make up $140 million of LA’s $9 billion budget. That’s 1.5%.
That’s not even factoring all the positive changes, financially and otherwise, cities will see from autonomy: parking can happen off-site, freeing up two lanes on a typical street for pedestrians, bikes, or public transport. Smart algorithms can make far more efficient use of existing infrastructure, reducing the need for construction and maintenance. Fewer accidents means less need for emergency response...
That's an interesting point I hadn't heard before.
Exactly.
Because why would they? Why would we even need speed limits, if the computer it in control and proven safe at X speed, because it has a Y reaction time vs a human's Z reaction time.
Well, reaction time is not all you need to drive safely. You also -and in fact, first- need to be able to generate the right reactions. Otherwise, you'll just end up reacting in the wrong way very quickly. And causing accidents very quickly.
In fact, if you have an autonomous vehicle with the wrong reactions and very short reaction times, it will pose a much greater risk to agents with slower but safer reactions (like humans, or just safer auto-cars) who won't be able to get out of its way fast enough.
If people see a vehicle has stopped for them to cross, they oftenbassume it's safe for them to cross. Unfortunately they are then less attentive to see if other lanes are following suit.
Pedestrians are constantly being hit in my "friendly" town where the culture has people stop for pedestrians even if the vehicle has right of way. But then the stopped vehicle blocks the view of another car in a neighboring lane, which keeps on coming.
Not a problem. That's not even a niche case and would be easy to train. The simple examples you're coming up with are the 99% situations that are readily trainable and have been. Literally the only thing stopping Waymo from production is the 1%... the situations that are so niche and unique that the Waymo AI hasn't seen them.
Yes, and when we get around that corner, we will be told the same again. And when we get around that corner we will find a lot of car wrecks. And the next corner will be some auto insurance companies, lawyers, and politicians saying "Not on public streets".
Eventually enough people will figure out and/or admit that some of driving a vehicle on current public streets with current traffic occasionally but too often requires real, actual, full, no compromise, human intelligence complete with ability to read, hear and understand, and speak and be heard, to comprehend situations and plan and execute solutions. That's essentially full AI, no compromises, parlor tricks, claimed grand successes in narrow contexts, driving on roadways defined by electronics, etc.
You're assuming the driving environment will remain the same.
It's possible the convenience and cost savings of autonomous transport would be great enough that society would be willing to change its relationship with roads.
But that can't possibly happen overnight. Until every car on every road is an auto-car, we'll have a mix of auto-cars with ordinary, human-driven vehicles. And in that time, what happens?
We spend trillions of dollars over the course of a couple of decades, with major disruptions of our transportation network, along with the overall economy.
This isn't a stand-alone article. It's the introduction to The Economist's "Reinventing wheels" special report issue. There's no timeline given for "around the corner" in the headline. Instead:
> This special report will assume that the technological hurdles to full autonomy can be overcome. It will consider the implications of autonomous vehicles (AVs) for personal mobility, car ownership and carmaking, but will also look at the wider economic, social and cultural knock-on effects. How will everyday activities be transformed? How could AVs reshape cities? And what lessons does the rise of the car in the 20th century hold for driverless vehicles in the 21st?
I doubt you’ll be able to buy a personal level 4 anytime soon, it doesn’t make sense to use them as anything but fleet/taxi vehicles since they can just come to you whenever needed.
Buying a L4 car may be not possible. Because I don't even think they will start selling them initially. May be we can bet on number of L4 cars in operation after 5 years.
Any likelihood of this vehicles being at times(say on the highway) autonomous and at times remotely driven ? Does this work technically , and maybe legally ?
Similar background and profession - putting aside whether an individual would ever want to own one, general purpose passenger autonomy won't happen for probably another decade. Limited domain autonomy will be here in before 5 years, though. It'll just suck.
But the first step to getting good at something is sucking at it.
>> BCG, a consultancy, reckons that by 2030 a quarter of passenger-miles travelled on America’s roads will be in shared, self-driving electric vehicles, reducing the number of cars on city streets by 60%, emissions by 80% and road accidents by 90%.
Besides the fact that it's obvious how these nice, tidy, exact-multiple-of-ten values are unlikely to be accurate, that last 90% in particular was totally pulled out of thin air and without any experimental, theoretical, or other justification whatsoever.
That's according to a study on the safety of self-driving cars, by the RAND corporation, Driving to Safety; How Many Miles of Driving Would It Take to Demonstrate Autonomous Vehicle Reliability:
Quoting from that (bullet points, sidebar on first page):
Autonomous vehicles would have to be driven hundreds
of millions of miles and sometimes hundreds of billions
of miles to demonstrate their reliability in terms of
fatalities and injuries.
Under even aggressive testing assumptions, existing
fleets would take tens and sometimes hundreds of years
to drive these miles—an impossible proposition if the
aim is to demonstrate their performance prior to releasing
them on the roads for consumer use.
Therefore, at least for fatalities and injuries, test-driving
alone cannot provide sufficient evidence for demonstrating
autonomous vehicle safety
In other words, forget about knowing with any degree of certainty how much safer autonomous vehicles are than humans (or even if they are safer at all) before they actually hit the road. And let's all admit that if we're eager to see self-driving cars in mass use, it's because we think the tech is cool and not because of any justified belief in their safety.
I'm sure people are thinking/working on this but it seems that nobody is talking about the infrastructure improvements that would likely be necessary to support fully autonomous vehicles (smart roadways, etc). To me, this should really be a two-pronged approach; the vehicle and the environment. Trying to engineer a vehicle to successfully navigate an environment soley designed for human cognition seems like a losing battle. But, if roadways were fitted with a myriad of sensors and communication networks which couldn't be thwarted by a can of spray paint, I think the success rate of the actual autonomous vehicle would be much better.
Not sure why you're getting downvoted, this is a legitimate take.
My thinking is this is two-pronged. It won't make sense to make the infrastructure investment until there are sufficient proof points that this is 1.) something people want and 2.) something that's economically viable.
It's likely that the early geofenced version of passenger autonomy will demonstrate what further infrastructure is needed and how much it would cost.
One interesting side point; dedicated lanes for autonomous traffic are already being proposed on some roadways, particularly interstates in the U.S and highways in Europe. The economic benefits from autonomous logistics (e.g. trucking) are more readily capturable, so the infrastructure investment might make more sense there up front.
He’s getting downvoted because not many people talk about “smart roadways” since it became clear that sensor-based machine learning autonomous vehicles were commercially feasible, and dare I say inevitable.
Good autonomous vehicles don’t work primarily by sensing and detecting the roadway, they work by sensing the environment and having a complete, detailed 3D map of drivable areas, matching their environment to the map as they go along. This only requires annotation of the terrain, not clear road markings or anything like that.
The 3D detail of the environment is the smart roadway. With always-on GPS and an incredibly detailed map of the environment stored on a honking hard drive in the trunk, there’s never a question where you are in the world.
In fact, now that I think about it, this makes reduced infrastructure possible, since you don’t even need things like signs or streetlights in an all-autonomous world, and can even get by with narrower lanes.
I work in autonomous vehicle R&D, and if we had an environment filled with reliable sensors it would considerably simplify our problem.
There is no such thing as a _complete, detailed 3D map of driveable areas_. Most current efforts rely heavily on high resolution, large scale, semantically labeled maps, but on any given stretch of road this is only a first approximation of the environment. Live sensors embedded in infrastructure that could pass along real-time information and updates, particularly from directions our on-board sensors can't capture, would be very useful indeed.
And yes, road markings are part of the semantics we use.
So how much will it cost to upgrade 4.7 million miles of roadway in the US? How long will it take? What does the technology look like? What impact will it have on the economy to close huge swaths of road?
These are all questions to be addressed, along with 'how many of those miles should we upgrade', 'what will we get out of it', and 'how do we make a self-driving car'.
Autonomous vehicles will make it much safer to be a cyclist, a pedestrian, or a child near a road. These are some positive changes I am really looking forward to.
Imagine knowing that if your child runs in the road, the cars will stop? Safety is a great feeling.
Autonomous vehicles are already here and have arrived a couple of decades ago. They are referred to by the umbrella term "rail transport". There are whole transportation networks all around the world, including the US, that are entirely driverless.
Very good point. If the car industry was really aiming for the benefits of transport without the need for a human at the wheel they could invest in rail transport, a technology that we already know works just fine. But of course the car industry makes... cars. So the only thing they really care about is in selling cars. Therefore, they instead invest in developing a new technology that may take years to be as beneficial as rail transport. And that will probably never be as beneficial as rail transport in many important ways, including very notably its impact on the environment and safety.
The initial launches of autonomous taxi services will be underwhelming - relative slow, sometimes frightening (there will be accidents for sure), and with limited deployment areas in limited weather. The novelty will wear off quickly, and then the real work is going to start.
There are basically going to be two ways to survive it - be part of a larger organization that can shoulder a long-term R&D burden to go from 'toy' to 'real infrastructure'. I can see Cruise and Waymo making it that way, unless GM gets cold feet.
The other survivors will be either super lean like Voyage, who went straight to revenue generating niches like retirement communities and rely on being downstream of the technical innovation being done elsewhere (no in-house vehicle or sensor development to cut down R&D costs and move quickly).
Or else they'll be in niches like logistics (e.g. Peloton, Nuro) where the parameters of the game are different, and the structure is more B2B than B2C.
One other play is the autonomy-tech licensing structure like Aurora is trying, but that's a hard sell, especially since they're dealing with German automakers who (from first hand experience in this domain in particular) are clueless about autonomy.
This is an incredibly exciting, risky time to be a part of this new, emerging industry. It feels in many ways like the very very early P.C era, where everything is very much still in play. I'm glad I made the career shift to get there.