"I have this new auto-mobile concept which doesn't require a horse and which can go fast; I envision up to fifty miles per hour. It will require a smooth, hardened road surface, but that will be achievable someday."
"Forget about it. We already have millions of miles of roads, which are bumpy, made of dirt, and hard to build and maintain. Who will do the extra work to smooth them? Harden them? Maintain them? What if some jerk digs a hole in one as a prank? Maybe this could happen in a limited way in cities, but this is overall a pipe dream."
My point is, it isn't a question of whether this is feasible, it's a question of incentives. If the incentives lead society in this direction, it will happen.
I'm sure it'll be possible some day, but I'm not positive that it'll be possible in the general case before we reach the Singularity. And when we get there self-driving cars are going to be a small side-effect of this unprecedented revolution.
I can't help to draw a parallel between these threads about self-driving cars where many people are saying that it's basically a done deal and we just need to wait a few years, and that thread I read a few hours ago about the Boing 737 Max re-certification being delayed once more. I know it's a bit of a fallacy to treat HN has a singular entity but when I read the threads about the 737 the consensus in here is that "it's a death trap and I'll never fly one of these planes ever again" but at the same time we're totally optimistic that the industry will have perfected self-driving technology in our lifetimes? The industry has been cutting corners on planes that cost a fortune and didn't manage to make them safe to operate in unobstructed airspace because of a minor sensor dysfunction but they're totally gonna nail the incredibly complex task of operating a 1+ton vehicle at highway speeds in much less controlled environment?
I can totally see an ever-increasing amount of driver assistance in the future. But fully autonomous driving everywhere at all times? I'm really not so sure.
What happens to all these markings when it snows a couple centimeters?'
Fully self driving cars won't happen in our lifetime, probably not this century.
Autonomous vehicles will only ever truly exist upon infrastructure literally designed to aid them, greatly simplifying how they need to interact with the environment, thus making the problem tractable with code we can prove works. I really think it will take more than putting markings on existing roads. It is going to take new roads full stop, probably with various wireless checkpoints built into them.
How will it deal with an accident up ahead where some drunk bystander is trying to direct traffic? How will it know to ignore the drunk guy? What if it isn’t a drunk guy but a sober person directing traffic? Does the car obey in that case?
None of those are edge cases because every time it drives it will encounter some novel edge case that has never happened before and it will have to perform better than a human.
Don’t even get started with liability. Once you take away the steering wheel the manufacturer is on the hook for every single mistake and every single accident. You’d be insane to be a manufacturer and sign up for that.
Sorry, but self driving cars are a complete fantasy.
[citation needed]
It's possible, sure. But it only seems likely to me because I grew up in an age of rapid progress in information technology. History, though, has plenty of examples of technological plateaus and regressions. To people in the 1970s, it seemed obvious that by 2000 they could vacation on the moon. But the rapid progress of the space race quickly dwindled; the problems were harder than we thought and the rewards smaller.
The notion that we can make a computer as smart as a human is one of those things that seems like it will be just around the corner. But it seemed that way 50 years ago, too. E.g. HAL from 2001. It's perfectly possible that humans aren't able to make anything smarter than themselves. Judging by most of the software I use, we're barely able to make things much dumber than ourselves.
I feel like the correct way to describe this future isn't "self-driving cars", but rather "personal autonomous trains." In est, the road system described here would just be a rather clumsy railroad network.
I interpret the goal of having "self-driving cars" as referring to the ability to have a passenger vehicles that can autonomously navigate (wayfind?) off-road, i.e. what the aim of the DARPA Grand Challenge would eventually evolve into.
Think about attending a festival or fair with grass parking. You follow a line of cars, pull up to a guy who's standing out in the field. He looks around and says, "Why don't you go park next to that red Toyota two rows over?" Sure, that part is not "on the road," but certainly I had to take highways to get there.
Maybe I, as an urban-dwelling American, only need functionality like this a few times per year. But there are significant chunks of this country and the world in general where this is part of daily life. Adopting fully self-driving cars without manual driving modes is going to take extreme amounts of change and adaptation, not only technologically but also culturally. I would recommend spending a few weeks in the deep country if you want to fully understand some of the difficulties in reaching level 5.
If the time scale you're talking about is on the order of 50 years, I could maybe see it. But I do think there will always be a need for personal vehicles with some level of manual control.
Beyond all that, however, this article to me seems like 90% clickbait. The statement merely was "Maybe it will never happen," and it was stated in the context of a discussion of the difficulty in reaching level 5 autonomy. But now we have articles throwing headlines up saying "VW Exec admits fully self driving cars may NEVER happen." Feels a little disingenuous.
We’ve done self-driving as a POC for a decade in Denmark, and it hasn’t really improved much, to the point where we too are considering, that it’s probably never going to work in the real world.
Don’t get me wrong. Self driving already works, it just doesn’t work on roads. Roads where you’ll suddenly have a bunch of leaves flying around. Roads where the paint job is cracking, perhaps even missing. Roads where the street signs are old and faded.
In ten years of testing with some of the best in the business, we’ve had maybe two days worth of self-driving.
new technologies are over-estimated in the short term, and under-estimated in the long term.
Decades ago our computers were "soon" to be voice controlled, listening to our speech and doing our bidding. That was a big load of hype. However, over time and below the radar it became true as computers first answered phones, then took limited commands in cars and smartphones and now it is basically true (without all the hype).
Also, I wonder if these kinds of comments risk becoming
"I think there is a world market for about five computers."
or
"640K of memory should be enough for anybody."
In fact, the new generations of Volkswagen Golf, Škoda Octavia, Seat Leon, and Audi A3 (released between October 2019 and March 2020) will already be able to communicate with each other[1] and with the traffic authority[2] in real-time to prevent road accidents.
Volkswagen Golf, Škoda Octavia, and Seat Leon are #1 best-selling car models in multiple European markets, and Audi A3 is one of the most popular and affordable premium cars in Europe.
[1] https://innovationorigins.com/nxp-makes-volkswagen-golf-cars...
I'm presonally bullish on fully autonomous transport because of the industry interest shown in it and the potential demand for it. The latter I don't see going away unless something comes around that makes it redundant.
We have already done this, they are called railways. Some of them evening have self driving trains.
To get self driving cars on roads you need a human level AI. Trying to get away with less intelligence by restricting the environment will never get to the point where it will be safe to have automated vehicles - you can’t provide infrastructure clues to the vehicle to help it tell when a child is about to run out into the road for example (and there are many other examples too) so you would have to segregate them physically from everything else and then you might as well stick to railways
Having cars move slowly but steadily does not make trips longer on average, at least in areas prone to traffic jams. Even though human drivers can find it frustrating, it's actually the opposite. If vehicles are consistently slow enough, you can even get rid of traffic lights and stop signs.
- researching it may not be feasible
* because it becomes a waste of engineering resources to continue the research
* because almost self-driving cars may be good enough
* political resistance or liability laws prevent mass rollout
* the cost of building and maintaining required infrastructure is too high
* not enough car purchasers may want itI think when someone says 'will never happen' they mean 'not in the foreseeable future'.
Obviously the environment, usage and science can change such that full self driving could happen. I mean in the 1600s nobody could imagine a Boeing 747 flying loaded with hundreds of people, sure.
But the hype around self driving (by all the dreamers) has been that it's more or less 'right around the corner' not in 30 or 40 years or even 20 years.
How resilient would such a system be.
How easily can it be attacked to fiddle with the ability of the car to navigate safely?
Also, and much longer term, isnt it kind of weird to think that human infrastructure can only be seen to massively scale i. The planet so long as we turn the planet into a cyborg?
Isnt it weird and wasteful to build a planet wide cage of tech infrastructure for the evonomies of the world to survive and for human civilization to operate along the trajectory we are currently pointing?
Seems dismal to me.
Or even easier, platooning. I don't understand why autonomous cars is a bigger thing than platooning. I mean, platooning solves 95% of use cases of self driving cars [1] and is orders of magnitude easier problem to solve.
[1] At least for me. I do not mind driving in the cities myself, but if I just could nap or watch a movie on the highway part, that would have some utility
Multiple important stake holders seem to have significant incentives to make this happen. Local/state govs want less traffic jams and crashes, auto insurance companies would love to collect premiums and not pay out, Uber, Lyft would love to get rid of their drivers
Never means never, not even in 145 million years. But he probably meant not in our lifetimes and with current urban planning. When all cars are self-driving it will probably be better
My guess is that interstate type highways will be instrumented for trucks and cars will benefit.
So this means that self-driving cars will have to safely handle these cases and that also means that it is likely that cars will have to still be able to be driven manually.
Bottom line: self-driving cars will have to handle absence of smart infrastructure (in which case do we really need that infrastructure? I think we'll still need it, though, to guide and improve traffic)) and/or cars will continue to be driven manually at least some of the time.
Just like horses. Am sure there was a time no one ever imagined cars would completely replace horses on roads.
Here's a grim reality: at ~3 million per death in a car crash (that was a reasonable estimate of insurance cost, overall, a decade ago or so), with ~37k deaths/year in motor vehicule accidents in the USA for instance, that's roughly $100 billion / year — a mere 0.5% of its $20 trillion GDP. So I'm not holding my breath for public or private action at a massive scale (think that fracking alone was orders of magnitude more profitable for the US, and came with a strong geopolitical advantage to boot with).
Do the math for your country, $3M/death over GDP, it's usually negligible compared to "the big thing" that your local politicians and corporations keep talking about.
Even in Western Europe, where
- regulation is people's #1 method for solving everything and anything,
- "the value of life" is emphasized every other speech and publication and actual social security systems, free medical care, free education even, etc. (a few hundred bucks away from actual UBI, for real),
- companies could actually compete (Europe has 0 tech giant, but several big car manufacturers),
you don't hear a lot of political or popular or private (business) support for Level 4 infrastructure (L4: roads dedicated to self-driving cars, likely to kill ~1000x less than human-driven roads, not to mention the economic gain of time while commuting and travelling by road, which whether work or leisure is a net psychological gain).
Actually L4 is not even a "topic" in many such countries (let alone L5), it's a curiosity, a funny segment to wrap up the news. Even though L4 is totally doable NOW. What you actually hear is much fear about tech — as usual. That's about it for self-driving cars.
I have no idea why, it makes no sense to me, but even if rich cosy comfortable life-adoring 35h/week western Europe doesn't want it bad, I don't know who does/will, in the short/medium term.
The above "grim reality" is just my way of fishing for answers, really. I don't know. I'm just skeptical that self-driving cars are a thing that people or leaders (public and private) actually want. I hear much, much resistance to the idea and very little interest for the upsides from the mainstream. I sees smiles and eyes rolling, and 10 years later there is still no decent infrastructure to charge EVs except Tesla's — a foreign entity, by far the biggest promoter of it all, but can they do it? Can they reach L5 or politically negociate L4? Back to the above concerns, or absence thereof really, of the mainstream.
It's like space, basically: it would be incredibly little of the world's GDP to put massively more effort and shorten industrial-scale space activities dramatically — like if it's 30 years away at current rate, we could make it by 2030 really easily, without pushing it far (nothing like a war effort for instance). And the benefits are so immense it's basically stupid to argue against, the question is how to do it best. And yet it's still anecdotal in most countries budget, it's mostly just PR. Even as we speak, a "prime time" for space as a topic of (positive) interest for the mainstream. Go figure.
Self-driving cars, it seems, are met with even more political and social resistance than they are made impossible by idealistic goals, because the former is a current showstopper whereas our current technological capacities are not.
And sing kumbaya, holding hands together....
What about adversarial car AIs? Malicious actors, etc.?
I reckon motorways could be handled easily enough, and basic dual carriageways and normal intersections, but once you start mixing up multiple modes in inner cities, some tough decisions need to be made.
So it's like the IPv6 problem. If we could all coordinate at once, it would be easy-peasy. In reality, it's virtually impossible.
Edit: a commenter pointed out that pedestrians are also a major problem; even in a far-fetched imaginary scenario I don't know how you would remove those from the equation.
Most of the value in being able to travel by auto without attention is in trips longer than 15 minutes. Commutes, tourism, vacations. Shipping. Most of which is certainly highway driving.
If it's possible to automate highway driving under most conditions (and safely transition either to a stop or human-piloting when those conditions aren't met), then at least 80% of the value is there.
Wrestling with the harder edges of the problem is still the right thing to do for tolerance reasons, but I hope we don't have to see last-mile problems solved before we start reaping the benefits.
Behold! https://c1.staticflickr.com/8/7177/6894934663_0619c8bea3.jpg
Bonus: the vehicle can be resized at will, depending on the number of passengers.
Of course, I live in a city, where parking is a major pain. But the other things still hold. It it far more efficient of resources to have cars be used most fo the day, rather than sitting idle.
https://edition.cnn.com/2019/08/22/us/traffic-commute-gridlo...
They can also, if Level 5 becomes a thing, reduce traffic collisions. And if they all end up electric, through route optimization they can significantly reduce energy consumption requirements for vehicle usage.
The value in the 95/95 autopilot is because it will make commercial drivers be able to never need breaks (they sleep or rest until the autopilot is in trouble, and they can even remote control cars). And of course because a lot of people seem to hate driving and would pay to avoid it.
I don’t - I hate owning a car.
Given that planes, which rely on a class of specifically trained professionals with minimum hour requirements, still have issues with this from time to time, I don't think this is a realistic thing for cars.
Driving standards, at least in the US, are very low, and raising them is a hard problem.
that to me is something automakers can easily cooperate with governments to get done and done well. they already mark HOV and Express lanes and with federal legislated lane markings it would make automating those drives pretty simple. Then as the number of cars increase the number of free drive lanes is decreased until the vanish.
using metro Atlanta as an example express lanes have their own entrance and exit points so they are already separated. One set for I75 is wholly elevated two lane bridge work nearly the full length of its run.
Tesla Model 3 owner, I think even Musk is walking back what he "implies" as self driving. As into self driving with human oversight
This way, you can set out home with cars and you only operate the last part going to Cities for work. And when you leave you drive out to high way and then every full else will be automatics. It also partly solve housing problems and people can now live further away from work.
Surely this is a solution that is simply enough for the short term. What am I missing here?
A 90% autonomous trip where I have to pay attention 100% of the time gets me no real value. It's just a slightly nicer, maybe, ride.
It seems unlikely you commutes would completely autonomous even people commute on freeways.
Self-driving cars that only go 80% of the way would compete against those forms of transportation which have established business models, networks and in the case of trains are subsidized in a way that self-driving cars likely won't ever be.
Additionally the key advantage of cars is flexibility, not having to figure out how to get to or from the train station or airport. In your scenario self-driving cars would lack this advantage.
Now granted this may still sound like an appealing proposition in countries with poor public transportation infrastructure but that substantially reduces the size of the overall market. Sure this might be attractive for the US but what about Europe? Asia?
Now of course there is an obvious answer to this problem: Keep the steering wheel and drive the rest of the way yourself. That just means you'll have a lot of drivers who won't gain experience at the current rate though. Not sure I would like to share the road with those people.
It would be the end of owned cars, huge reduction of car production, millions of jobs in the transportation business amortized, to name just a few.
Spoken like someone who does not drive regularly in snow.
However that would mean that everyone would have access to it.
The idea of self driving cars now is, a winner take all situation where whichever funded effort that succeeds, generates outsized profits from licensing or going public at a high valuation.
The tags, whatever form they may take, will designate the official lanes. But people in snowy climates don't drive in the official lines because they can't be seen. They drive an emergent set of paths where everyone else drives. It's often, if not most of the time, that these paths human drives take don't follow the actual road markings.
Now imagine we're doing that 1000 times per mile, like what you're suggesting. Even if the device is free (both initial cost and maintenance) it's just too cost inhibited.
I think we shouldn't assume that the reason is some malicious intention behind "However that would mean that everyone would have access to it." It could simply be that the idea is not economical.
I can see this happening gradually, beginning with docking stations, followed by parking spots with wireless charges and then downtown roads and highways. With each step reducing the required amount of space destined to parking.
That will affect how outsized the profits will be.
Also, because of politics, I expect there to be separate geographical winners, at least in China and not-China.
People claiming most of AI hype barely know what the "AI" people mention is.
And people who go as far as drawing rosy pictures of human like general AI being your personal chauffeur are past ridiculous.
The entire idea of human-like general AI for practical applications is like trying to make people using horses for transport in 21st century, by trying to make a horse than if better than a car.
Also, while claiming these innovations are only a few years way may be insane, saying they will never happen feels probably just as non-sense. History hasn't been on the side of such absolute claims.
It will take time, but Waymo already has working 100% self-driving cars out in the wild. Yes it's in a small area, with perfect weather, and with technicians ready to jump in, but it's still a good start and easily paves the way for it to extend over the next decade or so.
Driving a vehicle is not an AI problem and none of the serious autonomous vehicle teams are using AI.
Unless you are using a super loose definition of AI that includes anything with an if statement.
Insurance solved that problem, both by eliminating the possibility of a catastrophic financial loss, and by creating a buffer between me and all those lawyers.
I predict that insurance will solve the lawyer problem for self-driving cars as well. At some point, it will cost me $5,000 a year to drive my own car, and $500 to let it do all the driving.
And on top of all that, if my car drives itself into an accident, the lawyers will talk to my insurance company, not to me. I see the insurance companies as the enabler for this tech. And they will want to enable it, it will put them in control of the market.
However, I think the profit incentives of the trucking industry will manage to carve out some regulatory exceptions; something like “freight trucks can self-drive between 11a and 5a on these specific Nebraska highways, with warning signs on both roads and vehicles”. This sort of lobbying will be the thin end of the wedge for both iterating the tech, and normalizing its acceptance.
With a self driving car you know exactly what happened in an accident as video and telemetry is recorded. This is a tremendous advantage over having to reconstruct it without this data.
On the contrary I expect legislation forcing every car to include telemetry like the Chinese are forcing every car to be connected.
This accident's data is evidence, not opinion, not a belief, not a prejudice.
The usefulness of this has already been proven with airplanes.
He's poo-pooing level 5 autonomous driving and says that they just about have a level 4 autonomous vehicle?
This article makes no sense.
Honestly, the CEO of my company has very little idea of the details of the technology that we produce. If you picked some cutting-edge technology that isn't key to our market-share yet, he'd have even less of a clue.
Maybe he's right, maybe he's not. But either way it'd be dumb of him to claim level 5 autonomous cars were right around the corner.
Nuclear power, let alone fusion. What happened to “too cheap to meter”?
Supersonic travel: raynier already pointed out what a bust the Concorde was. Nothing else on the horizon now.
Automation giving me a 20 hour work week. Nope, capital holders just skim that efficiency right into their pockets.
I’d go on, but suffice it to say that about the only thing much different than my childhood in the 70s are computers in our pockets. Revolutionary, no doubt, but we are still burning oil for our energy needs, our cars don’t fly, and I still show up and do my 40 hours. And healthcare has gotten worse in the U. S., not better, if you can believe that.
So, yeah, when the head of the autonomous driving division of VW says Level 5 ain’t gonna happen, I don’t immediately jump to doubting him/her and attacking their resume.
[1] Yes, I know about the accidents
It's personal & subjective, and nobody will persuade anybody in a thread, but FWIW I absolutely positively can.
- I'm unlikely to utilize self-driving any time soon
- The stories of their firmware terrify me. And I don't want a DAW and computer games running on my ECU :O
- More pragmatically though, the Tesla UI paradigm is completely foreign to my way of driving/thinking.
I'm looking for a "HOTAS" type UI, where I can do anything I want without taking my eyes and focus off the road. A UI that's one giant screen, that may change position of buttons from minor firmware to another, is basically as scary and alienating concept as I can imagine.
I get that I am in a minority nowadays - a lot of manufacturer's are replacing switches, buttons and knobs with a touchscreen and deep menus, but not thankfully all just yet :|
Aside from the fact that the Tesla is not available where I live, if I was only going to own one car, it would not be a Tesla. A Tesla (or any other EV, really) does about 80% of what I need a car for, so it would be perfect to buy as a second car for me, to be used most of the time and have the other car be a backup for when I really need to.
If I had to settle for only one car, it would probably have to be something like a Chevy Volt, which I am quite sad got discontinued. Even then, that would get me about 95% of the way there.
If my non techy family see the Tesla performing what they deem to be self driving then they will quickly trust it 100%.
> "This is one of the hardest problems we have. This is like we are going to Mars," Hitzinger said in a comment. "Maybe it will never happen."
First of all, it seems obvious that we are going to go to Mars, eventually. Maybe not any time soon, but never? Seriously?
But the bigger thing is that there is about 1000 times more economic benefit to self driving cars than of going to Mars, at least in the near term. To think we'd just give up on it seems absurd.
Given the questionable economics underlying humans having a presence on Mars and the extreme toll on the human body this will have, foremost by radiation I think it's actually a fair comparison in particular because full autonomy like mars colonisation is constantly being overhyped mostly by a very small group of very affluent individuals who seem to be more inspired by sci-fi than engineering.
Let's take going to Mars. We can't reliably go to the moon. We can't even go everywhere on the Earth, where we have every possible advantage. Spend a year on the ISS and you'll develop all sorts of health issues. Spending time on Mars isn't likely to be less hazardous. We may send humans to Mars but it is by no means guaranteed. (And I would also ask, what reason do we have to go to Mars? Probes do a better job of exploring, and I doubt we could colonize it.)
Until someone builds a real self-driving car it is just an idea. On today's roads, it is probably not possible to safely implement a fully self driving car. Driving is not only a technical exercise but a social activity that involves communicating your intentions to other humans, and interpreting the intentions of others. That is something that humans do far better than machines. (Edit: To clarify, humans are better at communicating with other humans. Machines do a great job of communicating with each other, but have mixed results communicating with humans.)
If every car was self-driving and the roads were remade from first principles, then sure, that seems feasible. The degenerate case here would be a self-driving train, which seems perfectly reasonable. But the technical challenges are the easy part of that endeavor. Funding such a project, developing the political will to see it through, and organizing the logistics are far more difficult. Consider for instance; what will happen to the legacy vehicles? Will it be illegal to drive them? Will there be a massive government buyback? Who will fund that? Where will the cars go? How will we organize the logistics of moving hundreds of millions of vehicles? How long will that take? What other matters will we need to turn our attention away from to accomplish that task?
A much more likely scenario is that companies continue to come up with partial, ad-hoc solutions, driving gets more automated, ride sharing becomes more popular and car ownership less so, but that humans remain in the loop for the foreseeable future. What happens outside the foreseeable future is something we can't and shouldn't pretend to know with any degree of certainty.
There are a million ways in which we could never go to Mars or build a self-driving car. We could get a better idea for how transit should work, making self driving cars superfluous. We could discover life on Mars, and make the decision that it would be too dangerous for us to visit. The superpowers of the world could go to war with each other, and our infrastructure could be devastated to the point where space travel is impossible. Climate change could drive us to extinction. Something could happen that we cannot predict or imagine, that we have no precedent for, that completely changes our situation and outlook.
Some of these are more likely than others, but the point is that it would be folly to take the future as read. And frankly, a couple of them are more likely than us ever going to Mars.
We've had autopilot for decades, we're certainly not flying planes without pilots. Or even getting them from runway to gate without humans. Nor is anyone claiming that pilot-less planes are coming soon.
The thing is, there are some conditions where people probably just shouldn't be driving. Like we shouldn't be driving in a heavy snow storm, at least not for most trips where you can just wait it out. Planes will wait out storms, even with human pilots. I don't see a problem with "level 4.9" cars that occasionally say "I'm going to wait out the storm" or that will avoid certain routes (such as "crazy mixing bowl" intersections where you can instead just take a side street)
I imagine there’ll also be manual modes to get the car into trucks and ships, or even just out of the factory.
And I can't see mass transit being viable when our biggest cities with the money and initiative to improve it plan these projects over the scale of decades and billions per mile.
Self driving cars or trolleys on the existing road surface is probably the most realistic form of mass transit out there. We have the vehicles and the infrastructure already.
It doesn't have to be self driving at first either. Uber's forays into mass transit with shared rides and fixed pickup locations was always pretty solid for me and reasonably priced.
I can see that this isn't a future a VW exec would dream of, since it would mean far fewer cars per person.
The situation may be different in Europe, but I think self-driving cars could be useful for navigating the sprawling American suburbs or less-dense cities. But yes, we shouldn’t attempt to replace perfectly functional and more efficient public transport within cities, only supplement where public transit isn’t (yet) practical.
Why can't we have both, and sizes in between, and then dispatch whichever vehicle size is most appropriate for the situation?
... while Musk promises "you'll be able to self-drive, coast to coast, this year" (haha, no).
At some point in the not-to-distant future I expect them to bundle most of what they currently package as FSD and sell it as enhanced autopilot again, for a whole lot less than $7K. Then I might buy it.
This just reflects VWs ideas to accomplish the goal... ie: they have none.
Let me know when Google says it's not possible.
https://www.cnet.com/news/alphabet-google-waymo-ceo-john-kra...
And since he recently worked there I am sure he knows quite a bit about where the industry is at.
I were a VW engineer, why would I continue working in the autonomous driving division, if my CEO says the team would never succeed in the end game ?
I'd rather work in a company where the leadership is actually leading instead putting on the brakes.
I'd prefer an executive to be level-headed than spouting dreams that will never come true and expecting engineers to just make it happen.
If anything I respect him as a leader more than "visionary" executives.
This is incidentally one reason for Tesla's huge market value. The company actually has a plan to transition from individual ownership to fleet, so when this happens it will be prepared to deal with a new manufacturing reality.
Just try to imagine VW without all those ads to sell a positive self image because you drive a sexy cool car they make.
Cuts both ways. Tesla execs have the same reason to overhype autonomous driving, because their market value depends on being perceived as a hypermodern tech company.
So far reality has proven VW right. Elon himself had to walk back on grandiose claims about full autonomy and robotaxi fleets, or advertising cars with slogans like "the driver is only there for legal reasons, the car drives itself" a bunch of years ago.
Full autonomy on a human level requires human levels of intelligence and 'common sense', it's a ridiculously hard problem that requires several leaps in AI and plenty of other fields.
Why on earth would anyone use this over public transit? That would be a hugely expensive way to travel.
Highway driving is the most constrained normal driving problem, and this is solvable in 99% of cases. But there are so many things that can happen in most other driving situations that make me think that model-based approaches (Tesla) are doomed to fail... Go ahead, train a classifier for every situation you can think of - I guarantee you that you've missed many things.
Elon tweeted the other day that FSD is "coming soon". Either I'm totally wrong about this, and of course I hope I am, or Karpathy + the dev team should be tempering his expectations.
That isn't to say that there isn't immense value in L2/L3, there totally is. But I think that solving driving (being able to drive any situation a human can) is pretty much the same thing as solving intelligence generally.
If you go up or down, the number and kind of obstacles reduces. The location of interactions between the vehicles is reduced, and the interactions with other classes of vehicle are zero, so you can negotiate.
Solve a simpler problem, if you can.
Is the term "Grade" here used in the "Slope" sense, as in don't run the cars up and down the hill?
And if that's the interpretation, I don't necessarily agree with the next point that obstacles are reduced on slopes/hills... so I probably am not following correctly :-/
Thanks!
What I mean is it's a fascinating but daunting task.
And flying would be superior to driving (faster, no traffic, no need to maintain infrastructure, etc..) so there’d be little incentive to continue even trying to get to level 5 for self-driving.
Why are "admit", "confess", "announce", "reveal" etc used instead of the usually more accurate "says"?
(rhetorical question)
This is an off-the-cuff remark turned into a clickbait headline for a fairly bland article.
The article describes Level 5 as 'full computer control of the vehicle with zero limitations' but what exactly does that mean?
(a) Does it mean a vehicle available for sale to the general public that can drive at night when it is snowing on a new road that has not first been mapped by a human?
or
(b) Does it mean that there is at least 1 city of a million residents where at least 50% of vehicle journeys did not require any occupant to have a driver's licence?
To me (and Volkswagen) it's far more useful to forecast when the second criteria will be met than the first.
1. First, it begs the question. By saying "X admits Y" we are assuming that Y is true. This is very different than writing "X says Y", or "X believes Y" etc etc
2. Then, they carefully chose Y to include the word "may" so that they can claim it is true regardless of its content. As long as the rest of Y isn't literally a tautology, you're good to go. "A teacup may be orbiting Jupiter" etc etc
3. But step 2 is sneaky. The way that the word "may" makes the overall statement true is different than how it functions in the quoted speaker's sentence. In his sentence, he means like "there is a good chance it will never happen." When they use "may" to make Y true, it is invoked in a much weaker sense.
Basically they are reusing the word "may" in two different sentences to achieve a misleading headline. They obviously have skin in the game and I don't expect this to be a fair article.
I see the future diverging into two paths, fully autonomous commercial vehicles, like taxis, delivery vehicles, semi-trucks, etc. that work within urban areas or other designated, mapped and specially prepared areas. This will possibly involve a centralized system of control and communication, something like an ATC but for cars. These will be owned by corporations and only used by people. It makes no sense for a person to buy one of these fully autonomous vehicles, although I imagine some people would pay extra to have priority access so they always have one available.
The other side of the coin will be privately owned cars that have autonomous capability, or autonomous cars with override. These will be able to go anywhere the driver wants, including unmapped villages, small towns, off road tracks, etc. I suspect these will be the domain of enthusiasts, people who really need them for work (ranchers and farmers, for example) and people who choose to live away from urban centers. They will be more expensive than cars now, but the need for these vehicles will never go away. Even if we get true AI capable of driving anywhere with only the sensors aboard the vehicle, there will still be the need for a human to override it, even if that involves just authorizing a risky maneuver or putting the AI into "unsafe driving" mode.
I think the movie I, Robot (with Will Smith) got the future of autonomous cars surprisingly right, autonomous inside cities and on highways, and using the manual override comes with penalties (higher insurance, being at fault in an accident, etc.).
On a personal note, and this may sound bad, but I would never buy a car which I cannot use to break the law. Even if I never plan to do it, being able to speed, jump the curb, intentionally crash into a wall (or another car) or even run over a person (for example, in self defense) may be at some point required or the least bad of many bad options. In this case any consequences should fall on me, but I don't think a thing I own should be designed to prevent me from breaking the law or doing something stupid if I really want or need to, although providing warnings or an optional safe-mode is fine. I suspect many people feel the same way, even if they don't put it in such an extreme way. This can be seen by the fact that a lot of cars, particularly those focused on performance or off-roading, come with switches to turn traction control off, and if they don't, it will get mentioned as a negative in any review done by publications focused on those audiences.
Can one take an "autonomous car" up a barely visible car wide path (not a real road) which squiggles through a forest (up to a cottage)?
Can one make an autonomous car understand a free-form textual sign when there's a roadwork or accident?
Drive in a place without marked roads?
There are plenty of edge cases and difficult situations.
Hitzinger says that level 4 might be achievable. I agree, it is conceivable that some day a lane of a motorway could get reserved for semi-autonomous vehicles; those vehicles can communicate with each other and are allowed to drive much faster (say, 250 km/h) since it's mostly a straight way, computers have faster reaction times especially given early warnings from cars ahead in the chain and so on.
Sure, technological advances will boost the vision resolution or interpretive rigor to insane levels (compared to present day), and this will be used to acquire increased confidence from the public, but is that enough for you to sit your child in the street and trust the driverless car to steer around it? This is the question I would ask of any proponent of any model-interpreting public technology.
For example, there's a snowstorm out here today. Unless they really need to, people aren't going out displayed their incredible skill at navigating through snowsquals with centimeters of snow on the ground. They just stay home.
What will determine the success of self driving cars is not philosophical musings but their usefulness in day to day life. And if you can spend 10k on a system that'll work most of the time, but refuse to go out in snow squeals, it'll sell very well. I'd buy it.
The space is incrementally improving. Lots of new ideas.
I’m very bullish on systems that improve human driving and slowly move to more and more autonomy than “replace humans”
Anyone claiming full self driving is around the corner is prolly lying. I have no idea when it will be here. The edge cases are enormous.
But self driving on selected city routes and highways in good weather is here.
I would say Highway lane following and distance keeping is already better than human.
We aren’t great at holding monotonous focus. Machines on the other hand excel at that. We do excel at edge cases though. Marrying the two seems like a good bet.
When a work colleague drove home completely wasted and his Mazda CR-V nearly drove itself, you could not tell that his driver input was erroneous. Actually, it felt very safe. It dawned on me that these systems will mitigate a great deal of preventable accidents due to human stupidity.
But I am pretty sure it will happen within 20 years. 5 years? Unlikely.
In any case, it will require much better testing than just letting it learn on the street. A vehicle should be able to deal with a snow covered road in the dark going down a hill with babies crawling across the road.
And EMP-induced total electrical failure. No sense in a million people getting injured at once, should that ever happen. Emergency brakes should quickly engage in a predictable way when power is lost.
I'd expect level 5 autonomy to be slow, because level 4 delivers 80% of the benefits for 20% of the cost.
BUT "never" seems crazy unless you've got a very pessimistic outlook on humanity's medium-term future. Other than extinction or the collapse of civilization, what could cause it to "never" happen?
I hope that by the time I'm too old to drive - a few decades from now - self-driving are available, but I'm not betting on it.
There's video on YouTube from three years ago with George Hotz predicting that Level 4 or 5 would never be reached without artificial intelligence. It made sense to me then and it still does today.
Volkswagon can speak for its own company. It can't speak for the industry. If it were to speak for the industry the word admit can't be used. It would be a different word, like "believe".
Volkswagon can certainly admit that "they" won't have full self driving capability, but it can't "admit" that for other people.
The cost to upgrade roads appears to be a significant hurdle. A google search says:
"There are approximately 4,071,000 miles (6,552,000 km) of roads in the United States, 2,678,000 miles (4,310,000 km) paved and 1,394,000 miles (2,243,000 km) unpaved."
And there's this:
>> einrealist 18 hours ago [-]
>> I am more fearful of trolls, tricking the technology.
That's even more difficult to address.
We have to evaluate the cost/benefit of implementing this once we get the tech to the point of near total awareness of real world conditions. It might make sense to implement it on major highways, but probably not on rural roads and neighborhood streets because it's not something we can skimp on.
To make it cost effective the roadway tech has to be for the most part "dumb". We cannot rely on "smart" tech that requires complex communication systems or dumb tech that's easy to hack, like lines painted on the road or stop signs and traffic lights that AR can "see".
I think we'd be better off working on implementing assistive safety technologies for automobiles and mass transportation infrastructure like high speed railways.
What won't happen is cars without the ability to have a human take over. There are too many fringe cases to allow cars without steering wheels.
https://en.wikipedia.org/wiki/Hype_cycle
The big challenge for self driving cars is that they are held to a standard of perfection. No human driver can meet that standard either, but they are exempted due to incumbency bias.
The benefits of self driving cars are so phenomenally huge that if they are possible they will happen. Apart from the costs of paying people to drive, there are the time benefits of not wasting time driving cars - is there anything more boring and tedious? - and the fact many people, the young, the old, those with poor eyesight cannot drive.
I don't see any evidence it is not possible within 10-20 years. Cars will get smarter and easier to drive and the final step will not seem large. For the edge cases, there is always the possibility of having the vehicle temporarily being taken over by someone remote. Remote driving is already done in some mines.
This is definitely not consistent with statements Alex has made in the past. Seems like more of an off the cuff remark a German engineer would make while confident that the fully realized result is right around the corner.
If you want to know about the future, don’t ask the incumbents. They will only tell you about the status quo.
admit (v): confess to be true or to be the case.
Correct use in a sentence: "VW admits guilt and pays $4.3bn diesel emissions scandal penalty"
The word admit doesn't apply to predictions of the future.
Compare driving somewhere vs taking a ride share and how much more prepared/present you are when you arrive.
Electric, sure, bring it on.
Let's suppose it will happen. Do you have a proof that such "good" tech does not develop a will to one day run over every human it can?
Updated to clarify message a little.
Level 5 means that the car can drive autonomously > 95% of the time, and it will absolutely be possible. Level 4/5 distinction doesn't really make any sense once the autonomy gets beyond certain percentage. It doesn't mean that it's level 4 until it hits 100% (which is impossible).
Level 5 car is designed to drive in all conditions, but there are always statistically unlikely corner cases or situations that require high-level decision making, which the car can't handle by itself. A single driver may never hit such case, and for them the experience is full self-driving.
It's likely that if humans ever developed the sort of AI necessary for full self-driving, this AI technology would radically transform the world. Using it for cars would be one of the least important applications.
