Self driving cars and beyond | Hacker News Reader

Self driving cars and beyond

a16z.com

145 pointsyogrish116 comments

It’s great to explore this topic. You can have a lot of fun in picking a random industry and imagining the effect that self driving will have.

Health - beyond the obvious effects on the emergency room, what else? Ambulances willl quite clearly be affected. Why would someone wait minutes for one to arrive when they can just hop in an autocar to go the hospital? Or would the entire emergency room itself be on wheels? (Less likely.)

Will seniors stay at home for longer? Absolutely. How will this affect the need for aged care nurses? Will their doctors come to them?

But we can go even deeper. How will autocars affect people’s desire to exercise? Will more people ride bikes as a result of the increased safety of the roads?

Education - at first glance, it’s hard to see. But this is the one I’m currently most excited about. Primary and high school education are both built on cars - we just don’t realise it, but a crucial underpinning factor for parents using a local school is its convenience for them in dropping their kids off and picking them up. When that goes away, what happens next?

The demand for better public high schools will explode as the friction in going to a high school 25 minutes away in an awkward direction evaporates...

This makes for much more fertile territory for new schooling concepts - if I want to start a new kind of high school, one where the kids spend a bit of time each day tending to farm animals or perhaps studying entrepreneurship, it now becomes 10x easier to build critical mass. And given how much people care about their kids educations, I expect the movements in this space to be highly disruptive.

Those are two. On health, I’ve barely scratched the surface.

Then there is - tourism - logistics - distribution (imagine Amazon’s FBA on steroids) - housing (what happens to all those garages? Do we build shipping-container sizes homes that fold it wherever they park?)

The best way to understand the depth of the upcoming changes is to pick a seemingly unrelated industry and explore how current car-related assumptions underpin it. It becomes clear, to me, that the internet of vehicles will be even more disruptive then the internet itself.

People already take.an Uber rather than wait for an ambulance if they are stable enough. You call an ambulance when moving is dangerous.

>Health - beyond the obvious effects on the emergency room, what else? Ambulances willl quite clearly be affected. Why would someone wait minutes for one to arrive when they can just hop in an autocar to go the hospital?

Besides ambulances (which should be used for emergencies where medical/paramedical assistance is needed on the spot),I mean you can take a taxi (or Uber) to go to the hospital if you can move and don't need immediate assistance, consider the amount of "periodic" (daily, weekly or bi-weekly) transporting of (otherwise healthy, but not enough to be self-sufficient) patients that cannot drive or walk for long distances, like - as an example - people under dialysis or chemioterapy.

Right now (at least here in Italy) usually the service is assured by volunteers that drive an ordinary car supplied by the (public or private) assistance, in most (not all) cases the patient can get into the car (and get out of it) without any assistance, and the time the volunteers provide could well be used for a number of other (needed) chores in assistance.

Every and each of the proposed benefits of vast public policy initiatives supported by a generous public purse in support of autonomous cars could be as easily derived from a similar public policy initiative to create public transit infrastructure.

I’m not convinced that’s true. The friction involved with public transit (limitations schedules, routes, and sharing of vehicles) are enough to limit much of the usage outlined in the above comment.

Vehicle fleets don't solve the spatial and temporal realities that are reflected in the schedules and routes of public transit. A fleet of vehicles has to be staged to meet fluctuating demands.

Mathematically, on demand vehicles are probably harder to schedule because of random variation in queuing. Consider a population of workers each of which sometimes works from home. Some days an unusually high, but not statistically unlikely, number of riders all stay home: the idling vehicles need to sit somewhere. Other days, a statistically likely but atypical number of riders don't work from home and there are not enough vehicles to go around.

Public transit addresses fluctuation with a low nominal cost of providing surplus capacity to normal operation in the form of standing room. Public transit can be sized more closely to peak demand. On the other hand, from 20:00 to 06:00 the peak demand capacity of an autonomous car fleet needs to be parked somewhere. Between 06:00 and 08:00 and again between 18:00 and 20:00 the fleet needs to be moved to storage.

Sure, in a city. But I really question that in places likes Southern California with its endless, winding suburbs. The place is built for cars.

The A16Z utopia proposes strong public policies based conclusions drawn from Denmark (about the size of Rhode Island) and the Netherlands (smaller than Delaware). Anywhere a car can go, a bus can go.

To put it another way, a premise that autonomous cars will work in places like Southern California is that an autonomous car will be at your doorstep whenever you need to leave. The ideal fleet to achieve this is one car dedicated to only serving your schedule. Anything else has all the drawbacks of party line telephone service.

At least where I am in the US (SoCal) distance to school is not a primary concern. The barrier to choice is not distance but school district boundaries and assignments.

I remember seeing a concept video ages ago that touch on many the things that self-driving can change(from Audi I think).

I always think in a fully self-driving world there would be huge shifts in the real estate market, living further from your workplace/school won't be such a problem if the commute is easy and the time is consistent, but also in a way that is more convenient than public transportation, similar to how there's higher demand for living close to good transportation links.

Also, if the car can drive itself then you don't have to park it close to you when you don't need it, it can stay at a remote location while you work then "pick you up" when needed.

I also wonder if that could have any effect on car ownership, one of the biggest hassles of living just with uber is that it might just fail you when you need it, either because of surge pricing or lack of drivers, such as late hours or major events, if Uber eliminates the need for drivers it can deploy its fleet at 24 hours a day besides maintenance and therefore could guarantee it's users a minimum supply in a certain area at all times.

Then again, would Uber own those cars? Does it make sense for them to lease unutilized car time from owners similar to what people do with idle CPUs for crypto mining? Thus allowing owners to have some passive revenue.

> Does it make sense for them to lease unutilized car time from owners similar to what people do with idle CPUs for crypto mining?

I wouldn't want to find a used condom on the floor of my car when it picks me up, or vomit after a night of transporting party goers.

Yeah, I imagine using shared self-driving cars would be as fun as using public toilets.

Lyft cars do not feel like public toilets.

That's because you can down-rate the driver.

That’s because they’re not Lyft’s cars. They’re personally owned vehicles with the owner in the vehicle monitoring its condition at all times.

As a single, twenty something, I would likely buy a luxury self driving RV. It would pick me up wherever I want and I'd fall asleep as it drives into the boonies somewhere.

When I wake up, brush my teeth, shower, it's already outside my office.

After I'm out it goes to a full service cleaning institution, they dump all the tanks, tidy up the space, clean where needed, so I don't have to.

It might also make the trip to the grocery store, and someone would load my groceries into my fridge for me.

And it comes back to me later if I need something, in a shorter time than my commute home would have been.

Assuming "hard" self-driving...

Education: more kids "driving" themselves to school. There's an age where a kid could be trusted to walk alone, but not ready to "drive" there. Get in car, car takes kid to school & back. The one thing stopping my kids from going to a preferred school is we can't drive 'em there at that time; give me a trustworthy car and I'll let it take them there.

Play: like above, parents will be more likely to let the kid travel a non-trivial, non-walkable distance to play with friends. No longer bound by parental availability, parents can just authorize the kid to get in and be taken somewhere.

Parking: there will be a "Napster"-type stage of anarchy where self-driving cars won't park, they'll just slowly roam around until the owner calls it back. I've spent a whole day in NYC driving in circles (while the family enjoyed scenic stops) in lieu of finding & paying for expensive parking; would be nice to just get out and let the car wander, using roads as "parking in motion". This will lead to legislation & new solutions requiring aimless vehicles to actually get off the roads, albeit into more compact storage.

Tourism: lots of people dislike dealing with taxis. I'd certainly be more comfortable with an automated "pick your destination on a map" UI, evolution of getting rid of elevator operators.

Travel: huge hit on travel industry. My family rarely flies anywhere, opting instead for painful overnight 1000+ mile drives. Many people will be far more likely to do such trips when they can get on a major long-haul freeway and completely ignore the driving process. No more trying to stay awake on a boring overnight drive, just doze off completely, waking up only for refueling or nearing destination. I'll happily drive the complex parts at start/end, but it's that 700 mile stretch that's agonizing as driver.

Housing: longer commutes will be more tolerable. I will live in an acre+ site, and let the car deal with traffic while reading/working/napping. So long as I can ignore the commute, I'll put up with a much longer one.

Similar to "parking in motion" we might even see "housing in motion"

Good point. Short term parking gets leveraged, car moving to a new spot every few hours while occupants sleep.

> Health - beyond the obvious effects on the emergency room, what else? Ambulances willl quite clearly be affected. Why would someone wait minutes for one to arrive when they can just hop in an autocar to go the hospital? Or would the entire emergency room itself be on wheels? (Less likely.)

Well, the car can't go through red lights or overtake in certain locations to get to the hospital more quickly. That could be an easily ignored issue in some cases (if traffic is light or if it's easy to get to the hospital from your current location), but good luck if there's a traffic jam or the car needs to ignore certain laws to save your life.

That depends on how far out you're projecting. In a hypothetical all-self-driving future, a car with a sick occupant could broadcast its emergency mode to the rest of traffic to let it through.

Ambulances do a lot more than transport patients. It's already the case that you can usually get someone to the hospital faster than waiting for an ambulance. Typically if you're using an ambulance you're going to need someone to provide medical assistance on location and/or help load you onto a stretcher.

Ambulances do a lot more than just move people around, they carry doctors who stabilize the patient prior to transportation so that a relatively small problem does not turn into a critical one.

> Why would someone wait minutes for one to arrive when they can just hop in an autocar to go the hospital?

If you can just hop in and go to the hospital by yourself you don't need an ambulance.

You can have a non-life threatening injury that prevents you from operating a car yourself, and without access to something like Uber or having another person to drive you, could necessitate calling an ambulance.

A study was done (I haven't read the whole thing) that seemed to indicate cities with UberX had a reduction in ambulance use: http://www2.ku.edu/~kuwpaper/2017Papers/201708.pdf

(Though the wording of the GP was slightly ambiguous and seemed to compare a time-sensitive emergency to being able to take some other vehicle to the hospital)

Or use Uber or a taxi

>> It’s great to explore this topic. You can have a lot of fun in picking a random industry and imagining the effect that self driving will have.

Yes, it will be interesting to see the public's reaction in a few years, when the promised revolution fails to materialise.

______________

[1] I say this with some insight borrowed from people in the automotive industry. A commenter in another thread has recommended the following Medium post:

We’re still very much in the early days of making self-driving cars a reality. Those who think fully self-driving vehicles will be ubiquitous on city streets months from now or even in a few years are not well connected to the state of the art or committed to the safe deployment of the technology. For those of us who have been working on the technology for a long time, we’re going to tell you the issue is still really hard, as the systems are as complex as ever.

https://medium.com/self-driven/a-decade-after-darpa-our-view...

A couple of years ago there was a new article on the front page pretty much every day about the great new thing that will change our lives--3D printing. I'm worried that whatever happened to 3D printing is about to happen to self driving. What did happen to 3D printing?

I have a 3D printer and never use it simply because I have more money than time. It can take hours to print anything useful. One problem and you have to start all over.

It’s the same thing that’s happening to VR and lots of other things that are getting hyped before the hard problems are solved.

There’s been a serious epidemic in the last decade or so of mistaking solving the low hanging fruit for being “almost there.”

3D printing works ok for prototyping in certain scenarios, but it got hyped up for things it was nowhere near doing. Similarly everyone jumped all over VR assuming solving the hard problems that have always plagued it were just around the corner. Willfully ignorant over-optimism has been way too prevelant and I hope it is replaced with measured, informed pragmatism with lofty goals instead.

> There’s been a serious epidemic in the last decade or so of mistaking solving the low hanging fruit for being “almost there.”

Wonder how much can be blamed on 24/7 news cycles, and how much can be blamed on pump and dump VCs (And now i noticed how close VC is to WC...).

I agree that VR got a bit hyped. AM, however, is alive and well. The biggest issues are cost and time - AM takes a long time and requires an expensive machine. Thus you can print super-intricate structural parts of an airplane, but not brake pedals for common cars.

I'm sure self driving will have a MUCH bigger impact on our lives than 3D printing. But it could take a pretty long time (really curious to see how long).

How? We already have means to get from point a to point b without driving ourselves. Sure, technology would be amazing, but besides that?

Autocars have the potential to reduce the price of this kind of transportation by an order of magnitude.

> Autocars have the potential to reduce the price of this kind of transportation by an order of magnitude.

Could you cite this? I find this hard to believe. I currently pay around $70/month for an "unlimited trips" Marta transit card in Atlanta.

It gets me anywhere I want to go within the city. Getting out of the city is another story.

What's the recovery ratio of Marta?

EDIT: It's 34%. And the cost per bus passenger trip is $3.63. Yeah, I could see automatic cars (and buses) being cheaper that that, unless the average trip is very long.

http://stip.gatech.edu/wp-content/uploads/2017/10/MARTA-AVs-...

The IRS reimbursement rate for auto expenses for business is about $0.50/mi. Uber rates (likely subsidized) are in the $1 - $1.50 range. You can argue about the details but your floor for self-driving cars is around what the cost to operate a vehicle is. Fleet operations have some economies of scale but they also have management and staffing costs.

Electric vehicles may someday bring down costs across the board.

Bottom line is that autonomous driving isn't a path to "too cheap to meter" transportation especially as it increases congestion in various ways.

On the plus side, much less stress and much more time to do things we like/find useful; many less accidents. On the minus side, many, many lost jobs. In any case, considering how important cars are in our lives, it's a trasformative technology.

Cheap 3D printers produce only useless plastic trinkets. There isn't a big market for replacement parts that can't take any load, because those things don't break very often. Very few people need to print cases or the like in exotic forms. Buying traditionally manufactured stuff is cheaper than printing it yourself.

However, 3D printers are used in the industry. Not only for prototyping but also for producing parts that can't be machined easily.

So I'd say 3D printers are pretty successful.

Not disaagreeing. It is possible to print useful plastic items. But what I found interesting when doing the research for a product we where producing, was that the price per item on a €1800 3D printer was about the same as producing them in quantities up to 2500-5000 units in India. The difference obviously was how many you could produce in a month, and higher quality from the injection moulded stuff.

I had expected the price difference to be more. In the end we didn’t go to larger volumes due to weak interest in the product.

The main benefit, as you point out, was the fast turnaround in prototyping.

Yes, I agree that you shouldn't take the premise for granted (that self-driving cars will be here soon).

A high-profile Uber investor has said that level 4 self-driving won't be in the U.S. for 25 years:

https://www.cnbc.com/2017/04/06/bill-gurley-uber-investor-se...

That means they can service people without driver's licenses and people who can't drive -- the old and the young.

Also, Chris Urmson said the transition will take 30 years:

https://spectrum.ieee.org/cars-that-think/transportation/sel...

https://mondaynote.com/autonomous-cars-the-level-5-fallacy-2...

Here's my negative scenario: self-driving is "AI-complete"; you can't really hit all the edge cases without solving AI in general, which is more than 30 years away (Kurzweil is the wild optimist and predicts 2045).

You CAN use self-driving in limited circumstances, but those limitations are precisely the ones that make driving yourself around more attractive. The expense doesn't go down as quickly as anticipated because of this. They are a niche technology for DECADES.

Just like 3D printing -- the tech obviously works to an extent, but it's just not appealing for many use cases. Same with VR -- the tech works to an extent, but the content is not appealing to general audiences.

I'm happy to be proven wrong though, because I would be the first one to buy a self-driving car where I could fall asleep behind the wheel. I would love to take a trip from SF to Portland or LA like this.

Another good link: https://medium.com/self-driven/a-decade-after-darpa-our-view...

Even if self-driving is AI complete , a taxi that knows how to self drive in the highway and easy areas of the city, but is remotely operated otherwise is extremely useful.

Add the ability to remotely drive a platoon of cars and it's even more interesting.

What are people doing abut this?

https://www.wired.com/story/self-driving-cars-power-consumpt...

The Wired article you link discusses Nvidia hoping to deliver a system with lower power draw.

I think the intro kind of overstates the downside. Lots of people will be happy to pay for more fuel if it means they don't have to drive.

Extra 2.5kW for prototype cars?

I think it will halve as the tech progresses, just due to optimization. Also, no windshield means better aerodynamic possibilities that could compensate for the loss.

> because I would be the first one to buy a self-driving car where I could fall asleep behind the wheel. I would love to take a trip from SF to Portland or LA like this.

You likely wouldn't be buying a car anymore when this becomes possible... this entire video series is about the replacement of individual car ownership with fleet based systems.

This is the big thing people don't factor in:

- the AI "problem" gets easier and easier the more cars talk to each other

- the average person won't need to 'afford' self-driving cars or buy add-ons for their cars, they'll just use Uber/Waymo and save money

- mass produced smaller fleet cars designed specifically for the job reduce costs per car

- fleet subscription prices to become cheaper and cheaper than owning a car (via competition, technical progress, etc), further increasing adoption

- 10x more developers/designers/traffic engineers start focusing on the problem (with real life/death consequences)

- 100x more real-time data is shared, thousands of new unexpected events become optimized each day

- AI cars get safer and safer, creating great social and altruistic pressure to not drive non-AI cars

- creating public health incentives as less traffic fatalities and law enforcement reduce expenditures

- countless businesses get wealthy off it (Waymo/Amazon/etc) and reinvest in the industry and R&D

- hundreds of billions in capital gets poured in rather than tens of billions

- traffic/infrastructure begins to be designed around them, dedicated lanes are built, further increasing the utility of using fleets/self-driving > manual

Once it hits a certain inflection point of adoption and utility it will be explosively exponential rate of adoption, just like smartphones.

Self-driving trucks that make 95% of the route on highways automatically, and only take a driver for a couple last miles and the loading-unloading operations would still be a huge disruptor.

>What did happen to 3D printing?

Basically "nothing".

Sure 3D printers came out much cheaper (talking of FFF Fused Filament Fabrication or FDM Fused Deposition Modeling ) and much more precise/accurate, but still they are way too slow for any practical "final customer" use, outside the pure fun of it.

Other technologies are still too expensive (or again still too slow) to get anywhere soon in "every household".

In the professional world, set aside the "quick prototyping" field, and possibly a few other "narrow" fields, like jewelry and dentistry they can make "products" that either miss the mechanical capabilities of the "real" thing or that - if you can make them with suitable characteristics or materials - still cost tens of times as much as a same product manufactured industrially with more traditional means.

3D printers aren't material reassemblers. They replace milling and lathing when those methods take a day or more per part. You won't see it in your cupboard but rather a flight engineer might use it.

You mean - I believe - a (mechanical) engineer working in aerospace industry, a flight engineer does other things:

https://en.wikipedia.org/wiki/Flight_engineer

And yes, there is some use for such printers, but given their cost (and the cost per piece manufactured) they are limited to other very narrow and "rich" fields like aerospace.

My friend bought 3D printer to print a plastic kitchen cabinet hinge part he could not find elsewhere. After few days of ramping up he produced a near perfect replacement for a broken plastic part. That was a bit of an overkill, but it was more fun than anything. I don't think he printed anything useful since then.

I cannot think of anything that I would need to print. If anything, I need fewer plastic things in my life, not more.

Last I checked there was a pivot to a (not exactly) new term - adaptive manufacturing.

On the printing side, I remember some companies getting into metal 3-D printing but I guess they are not exactly unicorn material.

What happened to 3D printing (ran and sold a small startup in this space):

We realized that the vision of a 3D printer in every house was absurd. Consumer ≠ designer. Amazon is cheaper, faster, easier, higher quality unless I am putting on my design/engineering hat and doing R&D/NPI.

We realized that 3D printers are just manufacturing tools, albeit really interesting ones. They belong in places where you might find CNC machines, laser cutters, etc.

3D printing is quietly revolutionizing small areas of R&D/NPI/MFG in every field. For instance, small to medium volume end part manufacturing: https://www.jabil.com/insights/blog-main/3d-printing-reality...

Lots of technologies are maturing and costs are coming down as the ecosystem moves from a high margin low volume prototyping mindset to a low margin high volume manufacturing mindset (look up MJF, EBAM, DMLS, EBM, CLIP, LOM, etc. in addition to the well known FDM and SLA technologies).

I second this. I'm part of an university team building a racecar and we mostly use additive manufacturing for carbon fibre parts that are difficult to lay by hands, such as the intake and certain wings. For this we just need basic polymers. It's a cost issue.

However, this year we are starting to build functional components. We are looking at printing the brake cooling ducts - they need to survive 200 C and 400 bar.

Some better funded teams print their uprights (the part joining the wheels to the frame) in titanium (https://www.reddit.com/r/FSAE/comments/6acr6d/amazing_additi...). The part is strengthened by annealing it.

The biggest problem in AM right now that I hear about is "how do we scale up"? Batches are very small and take a long time - to be truly useful, one needs to print a part in minutes, not days.

> 400 bar

Say what now? Unless it's a misprint, that restrictive a cooling system sounds like it flows basically no air?

That'd gave to be 400psi, or maybe kPa. 400 bar is more than most hydraulics run at.

Most race car brake cooling air ducts are made of silicone and rated to at most 20psi (125kPa). Brake cooling ducts are powered primarily through ram air effect, and cost a few bucks a foot.

Not sure what OP is referring to, but if they're talking about a replacement to the 10-12cm high-temp diameter silicone air hose that typically gets routed to just front of brake rotors (and rated to 230C sustained, 260C intermittent), those measurements are MUCH higher than anything I've seen on genuine, real race cars, including WEC prototypes.

Now, the initial ducting may be made of a composite (including carbon fiber), which obviously can be tremendously strong, but the last bit of it is usually silicone.

3D printing is alive and kicking, and is a fantastic tool for the hacker/maker to have in his workshop. I use my cheap basic FDM printer (DaVinci 1.0) on a regular basis to make all sorts of things for my hobbies and business.

The manufacturing capabilities it has given me has absolutely revolutionized the way I think. I think back to before I purchased one, and I was so skeptical about whether it could produce any value at all. I assumed everything it would produce would be too weak to be of any value, and that the limitations of FDM would make parts too difficult to design. Well I was completely wrong and I only wish I had started down this path sooner.

I think my fantastic experiences with 3D printers is in part because I'm an experienced maker and a competent 3d modeller. I can knock up a prototype (like the Lipo balance charger circutry cradle I just finished) in an hour and have it on my desk in another half a dozen. Once I'm happy with that I can have another half a dozen ready for when I wake up in the morning.

Best of all it's just a tool. I never wanted a 3D printing hobby, and I haven't had to take one on. I put filament, power and my STLs in, and parts come out. I've had to repair the wiring loom on one stepper motor as it's a basic unit with no cable chains, but other than that it has been operated like a hammer or a soldering iron, on when I need it, off and on a shelf when I don't.

3D printing is here, just not in the form envisioned by the media. Autonomous vehicles may be here soon, or they may take decades to be ready for prime time. There seem to be indications that autonomy, real autonomy, is hard and distant.

What happened to 2D printing? If you want a paperback copy of a book, do you download a PDF and then print it on your laser printer, or do you order a copy from Amazon / a store?

It went back to being an industrial tool, which is what it was before the "replicator" fad. There are now quite good 3D printers that can print strong parts. They're industrial machines found in factories. 3D printing as a job shop service is doing moderately well. The home 3D printer remains a niche product.

Could be worse. Could be 3D TV.

3D printing is a real technology that just doesn't make sense for a consumer product. Autonomous vehicles make a little more sense as a consumer product, but aren't anywhere close to being a real technology.

In particular, I'm always confused about how autonomous vehicles are supposed to solve congestion. There wouldn't be fewer cars on the road. Actually there would be more, because some percentage of the fleet would be en route to pick someone up rather than in use. And I see no chance of them reducing total trips.

But it's impossible to respond to a string of videos, so I don't know if this is relevant to the posted "article", which, it should be noted, is from a VC group that's _heavily_ invested in this tech and benefits from prolonging the hype.

The presentation doesn't take long to leap from inevitable market forces to totalitarian restructuring of public life in the US by city planners to provide those market forces with a chance of inevitability.

All those cell phones didn't require drastic changes to public space. They didn't come about by massive new principles of public policy. They came about because cell phones layered on top of existing lives and policies in a generally unobtrusive way. The cell phone created market forces that didn't require the displacement of large swaths of humanity. It was a rising tide that lifted ships other than those of venture capital.

This is kind of a meta comment, but I’ve been lurking on HN for a while on car posts in particular.

I have the impression that HN in general doesn’t like cars all that much, and in fact we would be much better of without them.

I can’t argue with the facts: congestion and pollution in large cities, such as London, have never been higher where almost all on-road public transport is still run on diesel.

However, I think driving for a lot of people is a liberating and pleasurable experience. and, before we go there, I’m not talking about being stuck in traffic, which is often a issue with poor planning and lack of infrastructure investment rather than an issue with cars themselves, but the pure pleasure of driving.

I’m quite concerned that self driving cars will take this simple pleasure and satisfaction from us, and perhaps that’s just a selfish opinion. I don’t know?

Being stuck in traffic is a problem with cars, we can't build our way out of that problem, more roads just generate more traffic as longer commutes become viable.

I suspect self driving cars will make this worse not better, if a 60 minute commute is OK with you then perhaps 90 minutes is if you can read a book on the way.

I agree, more roads are not the solution. That said, there are definitely poorly thought out roads in cities all over the world that actively create traffic because they unnecessarily disrupt traffic flow.

But, more generally I believe the solution to traffic is to provide compelling alternatives. It’s all about creating a balance. In London we have the opposite issue, there are too many people trying to take the tube and mainline rail, leading to massive overcrowding.

Now, is waiting 20 minutes to get on a Northern Line train, because the last four have been packed to the doors, a problem with people, or is it a problem with planning and infrastructure?

> I’m not talking about being stuck in traffic, which is often a issue with poor planning and lack of infrastructure investment rather than an issue with cars themselves

I disagree with this statement. I believe that cars are not a transportation solution that scales like public transportation, for instance, does.

For these same reasons I believe Musk's Boring Company's plans are exactly the wrong kind of urban planning ideas that will just create a "fast lane" for those who can pay, instead of a solution that works for all.

Maybe cars could come with this romanticized "freedom" and "pleasure" when they're not prevalent enough to cause the issues that make driving undesirable. But I think the last time that happened was when cars were luxury items, reserved for the rich suburban white America and I don't want the future to involve making transportation more under-democratized.

> I believe that cars are not a transportation solution that scales like public transportation, for instance, does.

One thing I would like to add is that public transport is too very difficult to scale. For example, the Northern, Central and Piccadilly lines on London Underground are massively overcrowded to the extent where it’s borderline dangerous.

I don’t really know what TFL can do about it, because the lines already operates at maximum capacity. They could build more lines at £15Bn each (http://www.cityam.com/279301/gbp148bn-crossrail-project-face...), but even if money grew on trees these projects take 5-10 years to complete, by which time population has grown so much that they too are overcrowded by the time they open.

Having lived in London I am starting to wonder if really large cities aren’t going to suffer under overcrowding and increased housing prices to the point where it very painfully (maybe) self regulates?

It already does. Far more people would live in the desireable areas of London and Manhattan, among others, if the housing were available at half the current prices. The effects in the greater Metro areas are less—people trade off commute time for lower prices/more space—but people still absolutely factor in CoL when making location choices. I certainly would never consider a number of areas for that reason.

> One thing I would like to add is that public transport is too very difficult to scale

A quick check show the Piccadilly line is doing 21 trains per hour and they have a capacity of 684 people. Around 14,000 people per hour.

That is roughly a 7 lanes of highway traffic under optimal conditions. Not even thinking about the other direction.

I think you find that it is easer to build a new underground line though central London than a 10+ lane highway.

London was recently faced with exactly that question: how to transport more people across the city efficiently without paying the astronomical sums required for either another Crossrail or an unthinkably damaging motorway.

The solution is in place now and is working. It’s the North-South and East-West Cycle Superhighways.

Any estimate of how many people these can move at peak volume? I'm curious how it compares to the Underground.

So I think the insight here is that, as with everything, it can only scale so far.

Here's a question though: is it easier to build a new underground tunnel now compared to 100(?) years ago? While previously there was a lot less existing infrastructure to have to worry about disturbing, these days we have a lot more advanced technology to make the process easier.

Chinese cities have grown tremendously large metro systems in a very short amount of time. China has 5 of the 10 longest metro systems, and 4 of them were started operating in the 90s or later. The Nanjing metro is the 6th longest in the world and was started running in 2005.

https://en.wikipedia.org/wiki/List_of_metro_systems#List

So the answer is a resounding yes, it is much easier to build metro systems than it was in the past. As long as you do so in China.

I am quite sure that people who will want to drive cars will still be able to - but not necessarily on the same streets as self driving cars. Isn't it the same now with horses and carriages?

Not sure where you live, but over here, horse and carts can indeed travel on the same roads as cars/lorries/buses/bikes/pedestrians/whatever. They're obviously a lot less common than they used to be, but every now and then you'll see a horse and cart go trotting by like it's still the 1800s or what not.

https://www.gov.uk/government/publications/code-of-practice-...

Being stuck in traffic is a problem with cars. Driving can be a liberating and pleasurable experience. I do enjoy driving on empty country roads. I don't enjoy driving to and from work every day. If the time comes and driving by humans will be banned, I will happily accept that deal. But I don't think that day will come in the next at least 20-30 years. Biking (without any engine) is also a liberating and pleasurable experience.

Yes, this is a selfish opinion. Traffic kills a lot of people, and it's not just the ones doing the driving. Traffic is terrible for the environment, causes noise pollution and respiratory illness.

I can understand that some people need to drive. But driving for pleasure is not something that should ever happen.

What is your opinion on flying for pleasure (i.e vacation)? We know that flying too is very damaging for the environment and creates terrible noise pollution. Would love to hear your thoughts on this!

That isn’t the same thing at all. When you fly for a vacation, the flying is a means to an end (transportation), not the goal in itself. This discussion is about driving for fun, in which the driving is the goal in itself, not transporting yourself to a “fun” destination.

So, I understand you'd be OK with driving to get somewhere (anywhere?) as the ends justify the means. And driving around a race track just for the purpose of driving sounds really bad? Or is the purpose of racing somehow different than someone taking a Sunday drive to nowhere? The race car also produces a lot more particulate emissions than a highway cruiser, both from fuel combustion and from the friction wear of tires and brakes.

What about taking the "scenic route" to get somewhere instead of the most direct route available? You still have a destination and purpose, rather than driving just for the sake of operating the vehicle, but you might drive a little further and take the car's emissions into a nicer area. What about driving a big, gas-hungry SUV or luxury vehicle to be comfortable instead of the most economical compact car that is feasible? What about people operating off-highway vehicles instead of hiking and using mule trains to get deep into the woods?

Starting with the same initial argument that it is wrong to operate a vehicle for pleasure because of its impact on others' health, why do we allow people to create filthy camp fires in the very locations where the most people will be exposed in national parks? Compared to a modern car, the fire is pumping out horrendous unregulated emissions...

Not necessarily, you could be doing both at the same time. This is actually quite popular, as a kind of Grand Tour.

For example, a driving tour of the French Alps would, I think, count as driving for pleasure, especially when you could take a coach and do the same trip. You can even pay to go on organised tours, such as http://petrolhead.tours.

We get it, you take the bus.

To not be completely facetious though, it's going to be hard if not impossible for you to understand someone's passion for something, if you don't have any concept of enjoyment for it yourself.

A great many people find enjoyment in driving, for she sake of the activity itself. There are inherent risks and environmental downsides to it, for sure, but the same could be said for many things.

Your comment is however, the sort of thing that the grandparent comment was about - this attitude is expressed a lot on DN, and while this is an assumption on my part, it makes me feel that these people (such as yourself with your comment) must live in an environment where you haven't had the chance (or necessity) to use vehicles in the capacity that gives you an appreciation of the leisure aspect to them.

Even if that's not the case, however, it does pay to at least recognise that while you're strongly against something you can at least be objective in how you express your opinion.

BTW I think cars are super fun, I'm actively involved in motorsport, and I can completely understand when people aren't interested in it.

Why don't you think driving for pleasure should happen?

It doesn't have to be damaging to the environment, or dangerous if you're on the open road.

Let's assume your worst-case scenario comes to pass, and politicians decide (or are influenced) to outlaw manually-driven vehicles from public roads.

First, you would still be able to drive on private tracks; plenty of car enthusiasts already do this today. If your enjoyment is derived from driving fast, or taking curves, this should be able to satisfy your requirements.

Second, if your satisfaction is derived from aimlessly driving, simply enjoying the scenery and the zen of the road, this is not something that is incompatible with self-driving cars. Hell, there could be a "Scenic drive" button that randomly takes you on different routes every time.

If you are concerned that you will no longer be able to combine these activities (driving aggressively on long, rambling courses), I think this is something you are likely to lose out on. On the other hand, if this is the case, is it even appropriate for you to be doing this on public roads today?

If your argument is that you don't want to invest the time or money to use tracks or private courses for this purpose, then I think you should reexamine just how much enjoyment you derive from driving, and where it comes from specifically.

Ultimately, it goes to a core discussion of whether roads are for transportation or for recreation; I imagine similar discussions have occurred in the past between wagons and horseback riders, for example, or later, between horse-drawn vehicles and cars. Today, I would argue that more than 50% (and probably much more) of the cars on the road are being used for transportation rather than recreation, which suggests that modifications that streamline that use will eventually crowd out the remaining percentage through the tyranny of the majority.

Let's assume your worst-case scenario comes to pass, and politicians decide (or are influenced) to outlaw manually-driven vehicles from public roads.

There seems to be some confusion around HN about politicians signing/voting for laws and the people just magically acquiescing because a hundred guys decided it was for the best.

There's consequences for everything and the people can push back hard if bureacrats aren't thinking clearly.

>First, you would still be able to drive on private tracks; plenty of car enthusiasts already do this today. If your enjoyment is derived from driving fast, or taking curves, this should be able to satisfy your requirements.

This is false choice. It's like saying you can still drink beer and watch sports in a bar even if you can't buy beer and can't buy a cable subscription.

> I have the impression that HN in general doesn’t like cars all that much, and in fact we would be much better of without them.

In a discussion forum, people are free to form their opinions and rebut anyone, who they think is being incorrect. So, I don't understand why do people say this forum is unfair to x,y and z, stuff like HN in general doesn't like "cryptocurrencies" or "cars" etc doesn't help the discussion.

As for self-driving, when I read stuff like this:

But the final one speaks loudly about what city roads of the future could look like: “that autonomous vehicles (AVs) in dense urban areas should be operated only in shared fleets.”

https://www.popsci.com/self-driving-car-fleets#page-2

I can't help but get worried.

I agree, but it's more than just that --- self-driving cars will take more control away from individuals and put that in the hands of corporations and governments, who will have a much easier time tracking --- and even controlling --- the large-scale movement of the population. When they decide to outlaw non-self-driving cars almost completely, "the frog will have been thoroughly boiled."

Of course, it's all done under the argument of safety. But I don't think a perfectly safe world, one with no risk at all, is one worth living in either. There's enough dystopian scifi around to predict rather accurately what will happen.

Regular cars are already being tracked. Real-time license plate scanning is a solved problem.

You're probably more anonymous buying a bus ticket with cash than driving your car.

You already got one or more GPS wireless tracking devices in your car, though.

I have a GPS but I can turn it off when I don't want it, and I'm pretty sure it's off because it ceases to function when it's unplugged.

Do you also turn your phone off, because phones can be triangulated by cell towers.

The response-by-analogy is that people still ride horses for pleasure, and we should expect to see the same thing with cars: a (comparatively) small base of enthusiasts practicing the activity in a way that doesn't endanger others.

Broadly speaking, it seems like there are an infinite number of potential "simple pleasures" out there that people could enjoy, and if we have a chance to transition lots of people from a simple pleasure that endangers others to other pleasures that don't, we should probably do it.

I think a better way to reason about the tradeoff (since reasoning by analogy is unreliable) is to suppose that we were in a world where cars had always been autonomous, and imagine how it would play out–specifically, would enthusiasts in such a world try and build human-driven cars (maybe?), would they drive human-driven cars on public roads (probably not?), and would the non-enthusiasts feel a sense of longing for the joys of the road (for the most part, no?).

The future we should try to reach is the future that we would aim for regardless of our starting position. To me, that seems like a future where no one has to drive cars, and some people who want to do drive cars, but no one has to be impacted by the manual-driving of cars if they don't want to. So I think the response-by-analogy is probably correct in this case.

Well, beyond that, I have to wonder if self-driving cars are really the solution we're looking for, rather than more mass transit.

Mass transit is clearly the solution we need, but it is tricky to convince people since they are so used to private transport.

Right, like, OK, maybe mass transit isn't the solution for rural areas. But most Americans do not live in rural areas; they live in metro areas, which can probably be served by mass transit. Even the suburbs, I think, could plausibly be served much more by mass transit.

I am an advocate for autonomous cars, but I have to wonder if tech behemoths aren't forgetting a huge portion of people live outside of cities. It's important to note that rural people will be needing and buying new non-autonomous cars for decades yet and I don't know that there is actually a way to integrate autonomy into rural driving habits.

I'm not so sure. I bet the suburbs will spread out into rural areas when people don't have to drive and can just hang out and watch Netflix on the way to work.

America is huge, it would take hundreds of kilometers of suburbia to reach many towns. And there are tens of thousands of small towns that are themselves hundreds of kilometers away from the nearest burb.

Part of my comment about rural driving habits involves lifestyles in towns too, getting to work isn't such a big deal. Work is only 5 minutes away but they will still drive because after work they might need to drive 30km down some dirt track to get the kids from a friends house. Parking at the football club which is actually just a grass patch requires communication between humans about who is leaving first or maybe Stuart needs to get out to the jetty by 5 so we need leave room for him at the back. That kind of stuff is hard to solve autonomously. Driving in the country is just a bit more socially aware I guess.

The thought of autonomous cars on thin country roads is hard too. Sometimes there are dips or parts of the road that you know are dangerous through tribal knowledge and experiencing it previously. The car would have to have full geometry of the road for one. Second it would have to know that people sometimes pull out of this blind corner driveway and that if you continue doing 110kmph past it assuming it's a sealed highway you will one day have a head on no matter how quickly your car can react.

The conclusion I come to is that fully autonomous, no controls in the car, is unlikely to happen at all, as there will always people needing non-autonomous or semi-autonomous driving.

The most interesting phenomenon is "sleepers": people who sleep Friday night as their car drives to some scenic overlook sunrise and drive them home from retreat on Sunday night. Rural back roads will be moving at full stream in the dead of night with 'ghost cars' when they were once were completely quiet; zoning will limit noise, speed, illumination, toll's win go up, but airbnb's will flourish.

In the urban zone, there will be competition for scarce roadway as businesses no longer paying a driver, go increasingly JIT, and elite commuters value door to door service. No city gov't will be able to resist the right to tax per mile, tax rush hour surge charges, and tax luxury idle vehicles. This will push the middle and lower market into on demand rentals. And without the driver as an informal superintendent of the vehicle, there will be a very annoying problem of getting into a car that a gross person knew they were never going to be in again. Enter the eye in the sky. Except for customer with a 4-star reputation or higher. This will create a race to the bottom for anyone who can't or won't establish reputation to get passage on an autonomous vehicle.

I think we can't even fully contemplate yet how alienating and de-humanizing self-driving cars are going to large parts of our society, and how amazing they're going to be for others.

I think the part about gross cars is a trivial problem. You pay a 100$ deposit for the account and if you puke in the car you lose the deposit. If that happens to often you get banned or whatever.

The problem is you can't cut people off of the only form of urban transportation available. Who is banned from the subway?

And the other problem is there are a lot of people who want to get around without getting their identity associated or reputation checked by a centralized system.

On the Tesla earnings call last week, Elon Musk stated that a software update coming in 3-6 months will allow Tesla cars with Autopilot HW2 to drive fully autonomous. At least on highways.

He's predictions are always late, so it may not be ready in August. But his predictions seem to come thru in end. If he can deliver by the end of the year or even next year, he would still beat all other car manufactures schedules.

We'll probably see self-driving before widespread electric car use. Self-driving is an auto accessory. It's going to come from companies like Continental and Delco, which make auto parts for manufacturers. The big players are targeting 2020 as the year self-driving becomes a more or less standard option. That may slip; Volvo's big Level 4 demo with 100 real users has been pushed back to 2021. But it's getting close. The hardware is here; it's now a software problem.

Electric cars need more infrastructure. Many more charging stations. Cheaper batteries. More lithium and cobalt mines. That will take a bit longer.

Self-driving cars and electric cars are independent technologies. Uber's self driving cars are gasoline-powered Volvos.

True, this (great) video series doesn't make a clear distinction between electric/AI. But the key evolutionary shift in the next 10yrs is the massive shift towards fleet-based cars, delivery drones, and the post-car-ownership era... it's not just about plugging sensors+AI into today's cars:

- self-driving fleet cars can drop you off and go automatically pick up next person OR recharge at base stations eliminating need for most parking spots (currently 14% of city real estate)

- reduction in average car size, as cars with single occupants no longer need 4-5 seats + cars can drive very close together + synchronized higher speeds... meaning less roads per car & less cars per road

- cars automatically linking up like trains on highways to conserve energy

- repurpose above unused lanes for dedicated high speed delivery drone lanes + drop-off/pick-up zones replacing of parking spots

- removal of traffic barriers in the middle of all multi-lane roads (so middle lanes can automatically be shifted to the opposite direction depending on traffic load) ala SF Golden Gate Bridge https://i.ytimg.com/vi/hMkLcAstxgA/maxresdefault.jpg

..but with software, and not just for rush hour

- redesigns of car interiors to support leisure time and work activity (swivel chairs, tables, screens, etc)

- eventual elimination of traffic signs/lights

Both electric/AI requires completely rethinking how traffic, infrastructure, and cars are designed, and both often compliment each other...so why not do it simultaneously?

Nobody talks about the infrastructure though. Roads, signs etc can't stay the same, they have to adapt to facilitate autonomous vehicles

I guess you didn't watch it before commenting...this video series mentions infrastructure multiple times throughout.

Almost a whole video in this series talks about infrastructure financing and taxation. As gas taxes, driver's licenses, traffic tickets, parking fines, etc will be eliminated. And 30% of court time currently dedicated to traffic will be eliminated.

It mentions taxing each Uber/Lyft/Waymo fleet ride or automated toll roads with surge pricing instead...

It's going to be a GREAT time to be a city traffic engineer in the next decade....

>It's going to be a GREAT time to be a city traffic engineer in the next decade....

No. It's gonna suck. The people who drive Teslas are the people the politicans will be leaning on you to please. Reality will still have 1995 Camrys and E-series.

The ecosystems that will be disrupted by self driving cars are vast https://semiengineering.com/giant-auto-industry-disruption-a...

Roads...where we're going we don't need... roads https://youtu.be/flge_rw6RG0

#crashfunction.com kubernetes generators of virtual autonomous drones?

The final video [1] has a pretty compelling argument for the conversion to self-driving cars happening quicker than expected:

Chen shows a picture of Easter morning in NYC in 1900 and asks the viewer to play "Where is the car?". After several seconds he highlights the one car in the frame surrounded entirely by house-and-buggy carts.

He then shows a picture of the Easter morning in NYC in 1913 and asks the reverse question: "Where is the horse?", as there is only 1 horse visible in the picture.

Obviously this is a carefully selected anecdote, but this is a really compelling argument that forward progress will be faster than expected.

[1] https://youtu.be/YSdspVJgSFE

It's remarkably hard to find 40 year old used horse though.

Great analysis from A16Z as always. I love seeing the myriad form factors. How they map to different use cases. And how pipelines of these vehicles and drones could be used to transport to any place on the globe!

A good place to get started with the tech details of EV design is with the Tesla, Inc patents themselves:

US Patent Application for Integrated Electric Motor Assembly Patent Application (Application #20170291482)

https://patents.justia.com/patent/20170291482

Reply on news.ycombinator.com