I know we'll have driverless cars within ten-years, and I know there will be benefits, but has any government yet started planning for the relatively near-instant transformation of the Trucker and Taxi industry once this "works"?
The fall-out is going to be intense:
* Fuel-stops are going to change completely, what's the point of half of the motels on long-haul drives when your car can drive all night and likely recharge automatically at a stop-point.
* Reduced downtime on goods movement will impact uptime in every other industry.
* Autonomous-vehicle-only lanes that line up with traffic timings creating a two-tier driving experience
* Huge influx of unemployed drivers who I guess might get "chauffeur" style jobs.
* Security issues, complex legal issues when there's accidents.
* Fake taxi's that drive a customer into a bad experience for "lols".
I know all this is extreme, but if history is studying the past to understand the present, science-fiction is studying the "future" to predict the problems of tomorrow.
There's no need. This transition will be slow. A few cities first with a limited number of cars. And then it grows organically and probably initially quite slowly. It would take many years to upgrade all taxis world wide. The production capacity to produce the vehicles simply does not exist yet.
So, the grid won't collapse. Drivers won't starve overnight. Legal issues will sort themselves out as early adopters encounter them, etc. Plenty of time for everybody to adapt.
I think when it really works it will be more like the transition from feature phones to smart phones. very very quick. If they actually work it's just too compelling.
Further, I expect the number of cars on the road to go way up because suddenly so many more people can use a car. For example, every one without a license. Kids, elderly, disabled. I also suspect it means there will be lots more long distance drives as suddenly it becomes "hey, wanna go to this place 3 hours away, we can play games and drink while the car drives us"
I'd certainly consider visiting my family 9hrs drive away more often if Friday at say 12am I could just climb in the backseat and go to sleep and arrive at 8am vs having to drive myself (usually 7am to 4pm)
> I think when it really works it will be more like the transition from feature phones to smart phones. very very quick. If they actually work it's just too compelling.
The definition of "very very quick" is important here. The smart phone adoption took over a decade. to get to 80+%. [1]
Cars are much more expensive than phones, and last longer. The average age for a car in the US is 11.5 years [2] and people replace their cars between 5 and 7 years. [3] And my guess is self-driving technology will be used at the top-end of the market, and slowly trickle down. This is similar to pretty much any consumer technology: the technology starts off as a premium and becomes a commodity over a longer period if time.
So, if we use cell phones as a benchmark, it will take 10 years at a minimum. It likely will take 20 years given the duty cycle of vehicles. (Assuming we all still own a vehicle. If we all move to subscriptions, that might be a different story.)
Fleets will be replaced sooner, because the commercial ROI will be compelling. But it still takes a ton of capital to replace a fleet, and they need to get the ROI out of their _existing_ fleet before they invest in new ones. So even that won't be immediate.
But it might be uneven. I 100% agree with your example of sleeping through the night. If I can do that with a rental, and that's cheaper than flying, I might consider it.
The bureaucracy is probably the larger hurdle. Snartphones don't really risk lives the same way cars do. And there are still a lot of people wary of the safety and capability of self driving cars. It could be hard to convince a city to agree to having them, and convincing every city in the US may take a long while.
When they are proven to be considerably safer (a question of "when" and not "if" IMO), I highly doubt many politicians will want to stand in the way of allowing them on their streets... nevermind if Apple ever comes out with a car, it could quickly become a highly sought after experience.
Not if existing fleets can be retrofitted to be self driving. At $15/hr, if you can keep a car or truck driving 100 hours per week that's $6,000 month of labor costs saved. It's reasonable to expect insurance costs will be lower as well. Fuel economy would be better too, since a self driving car would be happy to drive at 60mph or less to get more MPG.
Even if the system + retrofit costs $50,000 that's an ROI of ~8 months. These numbers are made up but even with the most pessimistic of assumptions I believe widespread adoption will happen much faster than 10 years.
Even retrofits still run into many issues. Delivering gasoline to gas stations for example requires hooking up hoses etc. Training people on site to do it is possible, but it’s a huge shift for an industry and unlikely to finish very quickly. The same applies to delivering cement or wide loads etc which require different behavior while driving.
On top of this retrofits for every existing model is a non trivial problem on it’s own.
Is there a simpler way to make long-haul more efficient by investing in trains along the major arteries of trucking traffic (which would include ports, big production and consumption cities/regions)?
And then more intense and automated load transfer between ports to trains and trains to last-mile-trucking.
This could be developed along-side trucking automation. Thus developing 2 networks meaning more resilience.
Infrastructure for trains is unfortunately very expensive. It's easier to spread the costs over thousands of trucks than to muster the political will to spend a couple of billions for a rail track.
Roads are just asphalt. Train tracks need all kinds of supporting equipment installed for safety. If you want to electrify the track I'd guess that just installing that is more expensive than building a road. I'd be happy to be proven wrong though, do you have any sources about the average cost per kilometer of roads vs tracks?
Technically, yeah.
In Europe we finally started to test fully-automatic couplings[0] for freight trains last year (as in, run a production train on the system, making sure to collaborate with freight customers that won't complain).
These couplings are a massive upgrade over the common hook&chain freight coupling, which requires substantial manual labor for (un-)hooking. The new ones theoretically allow for fully automatic freight shunting/train re-assembly.
Now we "just" need to up the speed limit for freight trains to at least the 140/160 km/h most commuter rail uses for max cruise speed, and retrofit the digital part as needed to the already-used mechanical coupling to some electric multiple units that operate on congested links with demand for intra-day freight.
A few (1-4) box/container/"swap body (truck/rail only deriviates of ISO 20ft containers)" cars should typically be no problem for these EMUs (especially the longer ones), and would allow to realistically compete with trucks on interstates (Autobahn). Just needs some very small shunting engines, probably ideally remote-controlled (from the rail yard's tower), to add/remove freight cars from the tail while the passenger doors are open. Timings is tight, but some automation for fairly aggressive acceleration/deceleration profiles during that shunting should suffice. Maybe later even add a low tier electric car's drive train hardware-wise to some such container carrier wagons, to let them do the shunting and possibly last-mile trip simply on battery power, as that'd greatly speed up the shunting (low speed acceleration is friction limited, and a separate locomotive can't use the freight weight on the freight axles for traction. They're only outfitted with brakes on typical freight cars.)
For example, these "swap bodies" which are like ISO containers with fold-out legs, should IIUC be able to get put on their legs coming from a freight car (just by folding them out and maybe using some small lifting jack style hydraulics to get an inch or two clearance), followed by that freight car (if it has a battery pack and some small motors) driving away under remote control, and the truck trailer driving under it. This would only need a switch and non-electrified streetcar-style concrete pavement integrated track long enough for one or multiple of these, on the side of an existing railway station. Some $50-100k plus whatever the concrete truck yard pavement itself costs. Adding the battery drive system thing should probably be around $20-30k on top of what the freight car (with the digital automatic coupling) would cost normally, at least if it's only manually operated with sight contact. (It'd recharge via the train's electric heating bus that was originally designed to replace the steam heating for passenger carriages. The coupling that was/is used with hook&chain is 600A@1000V 16.7 Hz AC (in Germany), with one coupling either side of the coupling. This should leave ample recharging power for a small battery pack on the attached freight.)
This obviously only works where trains and existing road infrastructure are level, but on cursory inspection, many train stations here in Germany already have that. On many commuter links there's hourly or even twice-hourly service here, so even fairly short trips could realistically make use of this shifting of traffic from street to rail.
If the ROI was in 8 months, I’d think the seller of the tech would be charging too little. Why give it away for that price, when an ROI of 24 months would still be compelling, would give far greater margins and a more balanced split between the consumer and producer of the tech?
Waymo is solving this for a handful of specific models, they would need to do significant work to adapt their system to say a 2010 Corolla let alone more exotic stuff.
It’s not just physical instillation their models would have to be adapted for different camera placement, vehicle weight, etc.
But not to the point where they are selling kits. They retrofit a specific vehicle for use in their service.
I think CommaAI is the closest to a kit, but it probably will never quite reach "no-human-required" autonomous driving; I don't think they're even aiming for that.
I don’t think retrofit will work at scale, as maintaining a retrofit for so many models for software as well as hardware will work, building a new model in keeping self driving in mind will be much more cheaper in long run.
> The Court agrees with Gray that it is incredible that Officer Hiser—who self-admittedly does not have a heightened olfactory system—could smell the scent of two resealable sandwich sized plastic baggies of unburnt marijuana coming from a moving vehicle when patrolling in his cruiser. This occurrence is not only contrary to any common experiences, but is “implausible” and seemingly “contrary to the laws of nature.“
The replacement cycle for taxi ie commercial vehicles is lot shorter. Taxi fleets will move to self driving cars a lot faster. 2 biggest running costs for commercial vehicle are driver and fuel. So driver being replaced by ai and ice cars being replaced by electric happening almost at the same time means the transitions will be a lot faster.
The phone -> smartphone transition stretched out over the course of a decade. It’s fast relative to the adoption of the automobile for example but nowhere near as quickly as you make it sound.
I didn't get a smartphone until 2014. It was frustrating how so many people in the tech bubble acted like everyone had one long before most people did and just didn't bother with web sites anymore, leaving me out of so many new things. The failure to peek outside one's own experience is a continuing problem among the people deciding what the world will look like tomorrow.
See also: the internet. People still discussed seriously whether or not it had staying power as recently as 2008. It didn't really take off until the 2008 election pushed so many people to participate in new mediums. 2004 was kind of a prelude with Howard Dean's campaign, but he mostly reached the already online set. I was already on in the 1990s, and it still surprises me when I meet people my own age who didn't get on until the 2000s or 2010s even though it's the more common experience.
> It didn't really take off until the 2008 election pushed so many people to participate in new mediums. 2004 was kind of a prelude with Howard Dean's campaign, but he mostly reached the already online set.
I find the idea that the Internet’s adoption was related to US presidential election cycles a little bizarre. I know they dominate the mainstream media for a few months every four years but surely the utility of the internet vastly exceeds that?
What about online shopping or even email displacing paper communications?
Thinking about it more...I think texting, especially group texting, was much bigger during the 2008 campaigns than it was during the 2004 campaigns, and for some, rudimentary web browsing. And they did it on the pre-iPhone concept of a smartphone, that we've since thrown out.
What they used to call feature phones back then (which I think are practically discontinued now - most current feature phones would be considered smartphones in 2006) supported texting but it was used less than on the 2006 concept of smartphones.
If you double check context, you'll see the US was the subject of comment I replied to and it's reasonable to see an implicit constraint even without me qualifying it. HN's guideline on charitable reads applies to everyone regardless of geography as far as I know. It's only bizarre if you read it in a way that would be bizarre if it was a fair reading.
> I find the idea that the Internet’s adoption was related to US presidential election cycles a little bizarre. I know they dominate the mainstream media for a few months every four years but surely the utility of the internet vastly exceeds that?
Not to mention the rest of the planet. I'd like to take this opportunity to remind everyone that the majority of the world is not-America, and views American presidential elections as a newsworthy diversion for a few weeks every four years, and not much more than that. Those elections don't really drive any changes in behaviour. Hell, my own country's elections don't really drive any changes in behaviour.
I put the very late 1990s as hen the Internet took off. It went from being a weird techie thing to something a good proportion of the population was using.
The main graph on the top of this page shows Internet Users as proportion of population in Developed countries went from 11% in 1997 to 31% in 2000.
Honestly, I doubt it. If you're gonna make predictions, you kinda gotta look at what already exists. People that buy cars to ferry their kids around will look at things like safety ratings. Safety reports for robocars are super hush hush, which just doesn't inspire confidence. The stats aren't there either. People that over-rely on Tesla FSD already get a bad rep.
Price is another really big factor. Car buyers are generally not willing to spend even $3k on optional bells and whistles. My understanding is a lidar system costs like $20k. You can buy an entire car for that price tag.
Sleeper buses and the like also already exist and demand isn't really that high. Greyhound shut down a bunch of lines during the pandemic. I.e. if demand drops, an operator can go out of business. What makes a SDV fleet operator any less prone to the same sort of risk profile?
Another factor that seems suspiciously absent from Waymo discussions is Google's abysmal track record wrt customer support. Comms from Waymo always come out as carefully written press releases, but people actually using the service have to sign NDAs. Seems... kinda fishy?
> Further, I expect the number of cars on the road to go way up because suddenly so many more people can use a car.
That's just assuming rides will be dirt cheap. I don't expect that. With billions invested in R&D, costly hardware for self-driving car and tax starved governments to make money, these rides will be nowhere cheap where anyone could take ride anything between 10 min and 10 hours without second thought.
With all said and done I believe pricing will be in range of cab rides not bus rides.
Smart phones when launched cost less than $1200 which a majority of people could afford. Cars that have this technology onboard would cost upwards of $65000 which almost no one can afford comparatively.
The economics of a robotaxi are insane. Main cost now are the driver, and then fuel, and then vehicle wear/tear/depreciation. An autonomous EV taxi makes all three of those comparatively negligible. Profit per year per vehicle could be $25k+, meaning it is in their best interest to produce as many cars as possible.
In addition, each taxi will be replacing at least 10 cars, if it is sufficiently cheap and convenient.
I am convinced people who believe "wear and tear" on electric cars is "negligible" have never owned a vehicle. I worked as a mechanic at a bunch of small mum and pop businesses. Other than oil changes and brake pads (which will be less of a concern with electrics) nothing changes. The bulk of work in shops like that is suspension and steering (bushings, shocks, hydraulics), running gear (wheel bearings, hubs), AC problems (pumps and clutches mainly) and failing electrical parts (bulbs, starter motors, coil packs, wiring gremlins).
Electric cars are heavy, so expect more suspension failures. Running gear doesn't really change. AC is identical but with electric motors (prone to failure). Electrics will be worse if anything.
Suspension failures are not particularly common, EVs aren't that much heavier, and maintenance can still be done. While EVs are somewhat heavier, there are plenty of crazy heavy ICE vehicles that work without issue.
Tire wear may increase, but a robotaxi can also optimize driving style to reduce it.
None of those costs matter, compared to ICE. EVs just have less things that can break, and even less things that result in EOL or catastrophic failure. The more miles a taxi runs for, the more money it makes. Even if costs increased, the profit would easily make up for it.
Only because we regularly service them by replacing bushings, CV joints and ball joints regularly. And shock absorber failures are still commonplace.
> EVs aren't that much heavier
They're much heavier. It's obviously hard to make a direct comparison because there's no such thing as "equivalent but with an ICE engine" but a Golf E weighs about 250kg more than a Golf GTI.
> there are plenty of crazy heavy ICE vehicles that work without issue
And they have heavier duty components which require more servicing and are more expensive to replace.
> None of those costs matter, compared to ICE.
Can you show me some evidence for this? I am not making the claim that EVs are more expensive to maintain than ICE cars, but you made the claim that wear and tear costs (aka maintenance costs) on EVs are "negligible". You will need to show some receipts for a claim like that.
There are undoubtedly other changes involved, but BMW makes an ICE and electric version of their current 3-series. The 330i is 3,582 lbs vs the electric 330e's 4,039 lbs. Roughly 10%'s difference; quite a bit heavier!
For an apples to oranges comparison, a Tesla Model 3 is 3,648 to 4,250 lbs; a Honda Accord is 3,150 to 3,430 lbs. Totally difference cars, but same-ish external size class. Those battery packs are heavy!
330e is PVEV, not BEV. We're talking for battery's weight so not good example. Tesla 3 is relatively lightweight compared to other BEVs. I wonder what's the difference.
According to RepairPal, the average Tesla maintenance cost is $832 per year. That compares to an average of $652 per year for all car models sold in the United States. Depending on which services your Tesla needs, you may end up spending much more than the average car owner on yearly maintenance needs.
Also, Tesla's aren't known for their reliability:
Though not much information exists yet about the overall reliability of Tesla models, early results aren’t encouraging. In the J.D. Power 2021 U.S. Vehicle Dependability StudySM, Tesla ranked 30th of 33 car brands for overall reliability. That’s better than only Jaguar, Alfa Romeo, and Land Rover among all automakers across the country.
J.D. Power reported an average of 176 mechanical issues per 100 Tesla vehicles, compared to an industry average of 121 issues. The organization notes, though, that it doesn’t rank the electric car brand with other major automakers because it doesn’t meet the study’s criteria.
What can be known is that Tesla models break down relatively regularly and that Tesla maintenance costs are pretty steep.
Unreliable cars and a pretty steep per year cost to maintain them, along with steep repair costs (a broken door handle will run you about 1K plus labor)? This doesn't sound like the pipedream you're selling me here.
> Tesla maintenance cost is $832 per year. That compares to an average of $652 per year
I would be very careful before relying on those maintenance cost figures - they are very dependant upon the age of the vehicles and neither figure sounds realistic, if you look at the following information.
The link shows the cheapest average cost to maintain a vehicle over 10 years was $550/annum for a Toyota in 2016, and the most expensive was $1780/annum for BMW.
More importantly, further down, the article shows the cost increase over time for maintenance is about +$150/annum. At age 10 years the average maintenance is $1500, and it plateaus at about $2000/annum (because people retire vehicles when they get too expensive to maintain).
A Tesla doesn’t have oil changes or combustion engine problems, so some maintenance costs are reduced. However the battery eventually needs replacing, which is a large expense, so you must know whether the battery replacement is included in average yearly maintenance costs or not, before the figure is useable over the long term.
The numbers align with my anecdata. I'd say $832/year is probably around what people I know with a newish (0-3 years) Tesla pay. And people I know with an ICE car probably average $600 or less.
Your source seems to argue in favor of the numbers as well. If we just look at your Toyota and BMW numbers, the average holds true if there are 11 Toyotas to every BWM. That sounds plausible.
>The link shows the cheapest average cost to maintain a vehicle over 10 years was $550/annum for a Toyota in 2016, and the most expensive was $1780/annum for BMW.
Gotta love how these numbers are both a decent sized integer multiple away from what you would think they are if you got your advice from the internet.
Keep in mind that BMW number is probably including a yearly maintenance/tune up at the dealer that includes an oil change, new wipers, cabin air filter, and nitrogen in the tires all for one low price of $599.
Their reliability score does not take into account severity. Most things reported are super minor, and all are within the first year. Luxury cars are the 'worst' in these surveys, because owners are pickier.
True reliability metrics are average years on the road and miles driven.
Tesla ranks behind Lexus, Acura, BMW, Cadillac, Lincoln, and many other luxury cars in the same reliability studies. Even if most things are minor, they still have significantly more problems than the established automakers.
I wonder whether the average amount spent on non-warranty repairs would be a good metric to add to "average years on the road and miles driven". That would hopefully remove initial defects from the first few years and also remove defects that are so minor that most people won't pay to have them fixed.
"each taxi will be replacing at least 10 cars"
Two cars maybe, but not ten cars. Ten of us can't commute to work in the morning, and return back with the same taxi. Two round trips might fit during rush hour.
One of the easy things that we could do to improve quality of life and reduce waste would be to increase WFH and also to simply stagger work place start times somthat were not all hammering infrastructure at the same time twice a day.
Not all people commute. Rush hour isn't all at once, but distributed over several hours. Not all cars will be replaced right at the start, but those who use their cars the least will probably be the first to go taxi only.
You appear to be ignoring development costs. Waymo has to be one of the most fantastically expensive projects in corporate history. Even if they ramped up massively tomorrow, they'd still be charging taxi-like prices and it will take a long, long time to break even on that initial investment. But of course, they aren't ramping up tomorrow.
> In addition, each taxi will be replacing at least 10 cars, if it is sufficiently cheap and convenient.
Maybe the demand would be even higher than that - I'm thinking of all people that cannot drive (e.g. because of age in a few years for my parents & other relatives), the ones that don't own a car (too costly, don't want it, whatever), introverted people that don't like driving in a taxi/uber (that's me! hehe), etc... .
Of course alternatives exist already now but often they lack something (e.g. bus to remote location X drives only once every few hours and/or only up to a certain time, taxi is very expensive and/or availability is unreliable) => a cheap robotaxi would be a game changer, there would be many indirect consequences (and in my opinion many would be positive, like less parking spaces needed, more mobility for old people, more flexibility about where-to-go-for-what, room in the trunk, etc...).
> Main cost now are the driver, and then fuel, and then vehicle wear/tear/depreciation
Especially thinking about an idle taxi vs. robotaxi (waiting for somebody wanting a ride): unbeatable, as the driver's cost not generating any revernue would not exist anymore (only vehicle deprecation based on time and cost of capital used to buy the vehicle, which are both very low) => the difference related to fixed costs would be incredible (but I assume that some additional/new cost will show up, e.g. some data-exchange-service to keep maps/routes extremely up-to-date, coverage for remote-assisted driver in the case that the car would get stuck, etc..., but that would probably still not increase costs a lot).
> Maybe the demand would be even higher than that - I'm thinking of all people that cannot drive (e.g. because of age in a few years for my parents & other relatives), the ones that don't own a car (too costly, don't want it, whatever), introverted people that don't like driving in a taxi/uber (that's me! hehe), etc... .
Another thing that would increase demand would be the cheaper cost of AI driven miles compared to human driven ones. Today, it's somewhat expensive to get anything delivered from local shops, due to how expensive it is to do local point to point delivery. But if the local bakery or restaurant could deliver for a dollar or two, that would change a lot of behaviors.
> Main cost now are the driver, and then fuel, and then vehicle wear/tear/depreciation. An autonomous EV taxi makes all three of those comparatively negligible.
How does autonomy reduce fuel consumption and wear?
EV reduces fuel consumption compared to ICE. I don’t think it necessarily reduces depreciation, especially if the price is much higher than a regular taxi to start with. Notwithstanding the recent year or two where cars have actually appreciated due to inflation, in general more expensive cars with more tech depreciate faster than more basic cars.
businesses can afford it though. 65K to replace a very large, yearly, expense of a driver? you bet your ass they are going to switch and fast. That is what is going to cause most of the fallout. the ROI is basically 1 year, 2 tops.
i disagree with the 65,000 figure. A baseline tesla model 3 is 38000 dollars, which has a computer that can handle full self driving beta. I can see the price going down TBH
Tesla's FSD is really far from actually being full self-driving. There's a lot of computer required to get from 85% perfect to 99%. Uber and Cruise also both have lidar which is spendy.
I think price is an important difference between phones and cars. When they started coming out smart phones were expensive, but within reach for a big slice of people. There's a lot of people who don't buy brand new cars because they're expensive and poor value so I suspect there'll be a natural transition where the tech takes ~10years or so to trickle down in to the second hand market.
The upfront cost could remain the same while externalities like road deaths, insurance, and even pollution (via e.g. more efficient braking / speeds) goes down.
It sucks, but most consumers are rational actors that don't care about those externalized costs (hence why we have tragedy of the commons). It's especially so for the price-sensitive demographics that GP was describing above, who care about their direct cost, which likely won't change for a while.
It’s not just cost though. I’m a bit over an hour drive to the nearest major city. I find I’m even less inclined than I used to be to head in for evening social event. If I could be driven, I’d be far more inclined to do that sort of thing and a (reasonable) cost wouldn’t be a huge factor.
The cost for me to get into the city with a human-driven car is close to $100 each way which is too high for pretty much anything other than going to the airport. Cut that in half and eliminate parking and it’s more thinkable for an evening event.
I don’t necessarily disagree. A lot of people seem to assume that self-driving will be super-cheap, hence my use of the term treasonable. In fact, some trike today’s IRS deductions are probably the floor especially for anyone who isn’t trying to use a clunker of some sort.
I actually agree that today’s Uber/taxi rates are probably a good estimate for many years even once self-driving is practical.
> I expect the number of cars on the road to go way up
This could also play out totally opposite to that - think cheaper on demand rides, no need to park near your work if your car can just go do something else, etc. This might play out as actually less people owning a car, or families owning just one car instead of two, etc. I remember a prediction from a few years ago that it may reduce the number of cars by 80%, which I doubt, but I also don't think it will necessarily mean more people buying cars.
First, it's unlikely, as most trips happen in small windows of time, and in the same general directions in a city, leading to very little chance to reuse: Peak cars on the road at 8 am on a Tuesday is not really going to drop, and then very few trips will be happening at 2am. It's the same thing that happens in many US cities for commuter trains: A large fleet spends the night on depots far from the center, and then makes a single trip downtown, where they stay put all day, just to make a lone trip out later in the day: The demand is just too low in the other direction to not have most trains idle most of the day.
But let's ignore that fact, and instead look at a different definition of number of cars on the road: Instead of how many cars actually exist, count how many cars are on the road at any given time. It's not just that there might be more cars in the end, but even with fewer cars, each car will be doing more miles, because you don't need a driver. Children visiting each other while being less of a nuisance for parents mean more trips. Old people that are uncomfortable driving. People going further away in geneeral, because being driven would be less stressful when the technology is doing well. More visits to the bar, now that you don't have to call an uber, or risk a DUI on the way. And let's not forget letting the car go back home to park, instead of paying for expensive downtown parking: Same number of miles with a passenger, but bonus congestion for everyone else.
When driving is cheaper, the number of miles driven goes up, and therefore the number of cars that you actually find on the road goes up, even with fewer cars.
The problem with the US transportation-wise is its low population density. One possible solution could be to have robotaxis take care of last mile transportation, and have reliable and punctual mass transport take care of the bulk of the commutes.
> I'd certainly consider visiting my family 9hrs drive away more often if Friday at say 12am I could just climb in the backseat and go to sleep and arrive at 8am vs having to drive myself
Have you ever done it via the bus. Because that gets you exactly what you want and overnight buses exist almost everywhere.
Except with buses you're limited to their schedule, still have to drive to and then wait around in a terminal with all of your stuff, deal with tickets, and then you're crammed into a bus with a bunch of random strangers with all their noises and smells. It's a much more involved and stressful situation than climbing into your own private vehicle at a time of your own choosing and not seeing another soul until you're at your destination.
Firstly, the adoption of smartphones is ongoing there are still people with "feature" phones (for whatever reason). As technologists/early-adopters we have an observation bias that everyone adopted them overnight, it's simply not the case.
Secondly, the adoption of tech in entrenched industries is much slower because the decision makers are older and more risk-averse. Next is the cost-benefit of the adoption. If it costs 5-10x to adopt a "driverless" truck that still requires a skilled driver for 2% of cases there is still very much a driver on-board, only now they have to be retrained in the new skill.
> Further, I expect the number of cars on the road to go way up
> I also suspect it means there will be lots more long distance drives
This is the wrong direction for the world to move in. We need to make our transportation efficient both to reduce emissions and not pillage the earth for rare minerals for these self driving electric cars.
MOST (not all) 3 hour drives can and should be done through train trips. Trains are efficient, they offer more equitable access to transportation, and they can be very fast. The American obsession with the automobile, and the way that obsession is pushed internationally through exported media, will ruin us all.
> I expect the number of cars on the road to go way up because suddenly so many more people can use a car.
With autonomous vehicle adoption, we may see the concept of blockchain "gas" (fees paid with rates based on real-time demand) come full circle as a real-time congestion tax.
> Further, I expect the number of cars on the road to go way up because suddenly so many more people can use a car. For example, every one without a license. Kids, elderly, disabled.
And in areas with already bad traffic, it will only get worse. Much worse. You thought LA or SF pre-COVID were bad? Those will be the "good old days." And that's not just traffic, that's also hunting for parking everywhere you go (those cars have to be stored somewhere, after all).
A comprehensive public transportation system means people unable to drive can still get around. People act like self-driving cars are the solution when we already had a solution but (at least in America) proceeded to tear it apart as soon as we heard "cars are the future, get with the plan" and derided public transit as "welfare".
Fundamentally, we cannot "fix traffic", self-driving cars or no, because a 15-foot car holding 1-6 people cannot hope to compete with a 30-foot bus holding up to 30 passengers (on top of the driver), and that still doesn't hold a candle to a single train car (an 85-foot Bombardier MultiLevel railcar, for example, holds 132 people...and has a bathroom).
Compounding this is the American obsession with single-family housing and large minimum lot sizes: it is a fundamentally inefficient and unsustainable use of land. There is not enough space to build in this manner in the areas most in need of housing (which are, not so coincidentally, the areas where most of the jobs are), and it is this same inefficient use of land which makes public transportation difficult to run frequently enough that people stop using their cars for most trips.
When you build densely, you can run comprehensive public transportation that alleviates the need to have a car every time you need to go to work, the doctor, the grocery store, the park, etc. This also means people who cannot drive are afforded far more freedom than if they were to live in your average American suburb where sidewalks and pedestrian infrastructure more generally are an afterthought.
Far from a utopia, the future self-driving car advocates proclaim would be a nightmare. A nightmare where people are more isolated [0], more municipalities are struggling with debt [1], and even worse climate change [2]. This is not the future we should be excited for, it should be the one we are attempting to prevent at all costs.
Self-driving cars are fine (not having to deal with the TSA would be wonderful), but the way people think they will reshape society is nothing but a net-negative - we shot ourselves in the foot, but the proposed treatment is just putting on a band-aid and calling it a day. A high-tech band-aid, but a band-aid nonetheless.
Edit: Another thing to consider - if self-driving cars enable way more people to get around, this isn't good for the environment because only 40% of US electricity is from renewable sources [3] (20% of that being nuclear, which we desperately need more of but which many environmentalists have an existential panic over due to envisioning them all as impending Chernobyls). Burning gas might be bad for the planet, but electric cars don't have a perfect footprint, either. In isolation, replacing all gas vehicles with electric (minus stuff like trucks) would likely reduce emissions, but adding a massive increase in trips would be far worse than if we had comprehensive public transportation in more cities and towns.
The cities that do make the transition quicker will have economic advantages, so I'm not so sure that the transition will be slow once a few cities make it. And we aren't just talking about the USA, we have to contend with large cities in Asia with worse traffic and more incentives to adopt the technology before American ones do.
Labor is cheaper in Asia so I'm not sure why the transition would happen faster there. Lots of jobs exists there for the sole purpose that the price of a robot/automation is way higher than paying someone to do it.
Traffic jams and limited road infrastructure (with no space for more, so use needs to be optimized) are a PITA in Asia. It isn't really the cost of the taxi drivers (who are increasingly becoming scarce as young people choose better careers). Coupled with authoritarian governments that can push changes more readily than western governments, and Beijing, Shanghai, Singapore, Tokyo are easily on their way in this direction.
why would congestion change if all the drivers were fully autonomous? It's one thing to have a fully autonomous driver, but having them coordinate perfectly to have good throughput even in dense traffic conditions is another altogether. I think this second goal is much farther off and if the articles I've read on autonomous vehicles is any indication I expect they will drive more cautiously and slowly than most human drivers when they finally "go live" in a widespread fashion
> if the articles I've read on autonomous vehicles is any indication I expect they will drive more cautiously and slowly than most human drivers when they finally "go live" in a widespread fashion
They won't get stuck at left turns, on-ramps and rotarys due to an inability to use the skinny pedal to its full capacity.
That won't solve the sheer numbers problem but it's probably a solid sized constant factors improvement.
> As long as there are still some vehicles with human drivers on the roadways
You understand. If the government can tell you that you can only drive even or odd days, tell you that you can't drive at all (and must use self driving cars inside the 4th or even 5th ring road) is a very distinct possibility.
I fail to see how driverless cars make any of that better. Congestion might even get worse as people might be willing to wait longer if that means they can be “productive” doing something else in the car.
This argument didn't stop cab driving from transitioning from a serious employment opportunity to employment as a service contractors - so I don't know if that's enough to hold back the tide for the taxi side of things... on the trucking side I think there will probably be a big reckoning at some point though. Teamsters unions mean business.
Yes, but, it unfortunately only leads to social progress if everything else keeps up - you'd be surprised how many social programs currently have work requirements ¯\_(ツ)_/¯. Unemployment benefits, some type of UBI, education/job training affordability, social security, etc. are all things that need to be reevaluated if a whole job category is going to be removed.
Job automation is great, but there are externalities that should be accounted for ahead of time. You can ask, "who pays for this?" Well, the obvious answer is the companies that will be enriching themselves due to society's sacrifice. Although, that has happened during the pandemic, so it's clear corporations aren't going to do so out of some sense of civic duty.
As automation gets better and better, the difficulty of the remaining jobs for humans goes up. Not everyone is cut out to be a software engineer, doctor, or lawyer. We have to deal with the fact that people of average (and below average) skills and intelligence still need to support their families.
I think hampering technological progress to protect jobs is the wrong approach though. We need to contend with the fact that traditional supply-side economics will start to break down as we enter an era where most or all human needs can be met with automation.
Imo, as automation takes hold, one of two things will need to happen:
1) We'll all need to collectively let go of our societal ideal that work is virtuous. We will need to accept that a significant portion of people have little to no productive work available to them, and that there is nothing wrong with that. It's just the natural culmination of increasing automation and societal advancement. UBI would be a necessity in this scenario.
2) We'll see an increasing number of "make work jobs" that exist solely to keep people busy, and to generate an illusion of productivity. These jobs may ultimately be government funded, although governments may obscure the source of the funding to keep the "productivity illusion" going. For example, governments may provide business grants to keep a certain number of low-wage workers employed. I suppose this essentially would be an indirect UBI program, except that we'd make people do some nominal amount of unproductive work to "prove" that they're worthy of the UBI.
I honestly can summarize it this way. Savers keep saving even when there is nothing left to save because there is no downside to having too much money but lots of upside to having more even if each single dollar is incrementally less useful. To maintain full employment all savings must be invested. The reason for that is quite simple. When you spend 50% of your income you are working more than you really need to. The extra hours that you are putting in mean someone else does not need to put in those work hours (50% less employment for them). So, you use the money which was previously sitting idle on a bank account by lending it out. Of course, the modern system has no direct connection between savings and actual money being lent out, rather it is the money hole (the money you have saved is not circulating) in the economy which causes unemployment that is driving further borrowing to restore full employment.
One could say that if you are working twice as much as you need, every second hour you are working is actually charity, it's not productive in the sense that you need it to maintain your lifestyle and it is not productive in the sense that anyone else has a need for it. There is no way you are getting a return off that charity when the economy is already saturated and everyone has everything they need. Yet money does not represent this act of charity, instead we pretend that you are still owed the same amount of products and services for all eternity, meaning what you are doing is not for the benefit of society, it's for your own benefit at the expense of someone else.
Probably the worst thing you can do is do what the Germany did before WW2. Do massive austerity programs during a depression hoping it makes things better. Homeless and starving people don't care that all the jobs are done by two thirds of the population if that means they don't get to eat. As I have said, if you are working more than you need to work for yourself, then you ought to work for the sake of someone else, anything else is selfish. If two thirds of the population is doing all the work and only doing it for itself then the rest is going to get angry very quickly.
> The extra hours that you are putting in mean someone else does not need to put in those work hours (50% less employment for them)
that's only true if you set the amount of necessary goods/services to a constant. I dont think that's true in the real economy - people have unlimited desires and demands, constrained only by the amount of money they have. The economy cannot be saturated, at least not until the world reaches star-trek level of tech where there's zero effort duplicators and materializers, and where no human labour is ever needed for anything.
Therefore, a person choosing a particular life-style that is less costly than they make in income is doing so because they expect to stop working in the future (and thus consume what they "save up" now, plus any improvements via productivity increases since). This savings takes a long and complicated path to turn into investment, but ultimately it would increase productivity, which is where the returns on investment comes from.
> if you are working more than you need to work for yourself, then you ought to work for the sake of someone else, anything else is selfish
well, the case study of communism ideal - that of https://en.wikipedia.org/wiki/From_each_according_to_his_abi... , vs what actually happened, tells me that this idealistic viewpoint doesn't work in reality. People will work more if it benefits themselves, and only if it benefit themselves.
If you shatter the "lie" that excess work is not for yourself, but for "charity", you also stop people from working more than bare sustenance, and it would stop the savings->investment->higher productivity cycle. Society would be poorer for it despite the equality that would result (everyone is equally poor).
True, rising productivity does force human beings to be more productive.
> Not everyone is cut out to be a software engineer, doctor, or lawyer.
Society used to say the same thing about reading and writing (not everyone is cut out to do that), and society was (mostly) wrong. Pushing people higher and higher up the value chain has been happening forever. Eventually we will hit singularity, but even the software engineers, doctors, and lawyers will be obsolete then.
> As automation gets better and better, the difficulty of the remaining jobs for humans goes up.
That doesn't follow, it just means that the demand structure changes. One thing that happens is the comparative advantage for being human increases, but you are human so that makes it easier. eg YouTube is a new kind of automation that produced a job called YouTuber.
Actually a lot of them didn't, hence the protests. Innovation can take away your job in a span of ten years while not creating new jobs for generations to come.
Luddites to the weaving loom, Blacksmiths to the steel mill, farm hands to the combustion engine, secretaries to the email, travel agencies to the search engine, etc.
Technology has a robust history of displacing jobs, and humans have a robust history of learning new skills. Change is never comfortable, but predictable.
> Innovation can take away your job in a span of ten years while not creating new jobs for generations to come.
That case isn't true. The new generations had jobs well enough (increased productivity grew the economy much faster than before), they weren't committed to weaving clothes by hand. The Luddites themselves didn't fair as well in the short term, but many of them would have aged out of the work force in a few years anyways (manual weaving is pretty unforgiving on the body).
Truck drivers are rapidly aging as well, if self driving vehicles move retirement up by 5 or so years, it isn't going that huge of a deal (we can rely even on America's social safety net), and their kids didn't want to be truck drivers anyways.
>Truck drivers are rapidly aging as well, if self driving vehicles move retirement up by 5 or so years, it isn't going that huge of a deal
That's quite optimistic. When did politicians ever talk about moving up the retirement age in the last decades.
The internet brought insane growth through productivity gains in all industries. Even more than I'd expect from self-driving cars. And we haven't talked about moving up the retirement age.
Also the question is what these people do now? In the next twenty years of unemployment until they reach the current retirement age?
I agree, it will be slow. These Waymo cars cost hundreds of thousands of dollars. A local pizza place in the middle of a small community will not be able to support the cost of a driverless car as compared to the cost of a local high school kid with a $500 Civic. Same with semis, companies invest in their truck fleets expecting millions of miles out of each truck. It will take time to change all that over, and the cost/benefit will take time to tip towards self driving. It will happen, but like seat belts, lead free gas and air bags it will take decades before they are the majority of vehicles on the road.
If an autonomous car costs multiples of what a human driven car costs, then adoption by consumers will, as you say, be slow. There is some upside to an autonomous personal vehicle, but it's not profoundly compelling. For starters, if it's your car, taking you someplace, you have to be in it anyway. You only get a benefit when you could be in it and doing something more beneficial than getting yourself where you're headed, while in transit. Sounds wonderful for commuting, but when push comes to shove, I doubt it'll shake loose the dollars for the majority of commuters. But taxi and like delivery fleets are whole 'nother matter. Here the cost of the driver is substantial, and being able to eliminate that cost compelling.
For that same reason, I think you're likely wrong about trucks. I suspect that autonomous - at least for the over the road portion of tips - will be highly compelling. American industry loves to substitute capital for labor costs, and that exactly what autonomous trucks do
I agree, it's interesting, I can't believe I didn't see it this way before. In the same way a company can easily justify investing in automation in a factory to improve the efficiency or remove hand labor in production line its the same for trucking and other situations where the vehicle is a tool, not a family car. It seems that if the tech existed, if a company could buy an autonomous 18 wheeler, then they could easily justify and get the capital to buy said trucks, thereby effectively replacing future labor costs with capital expenditures that they can depreciate. So in this situation the limiting factor isn't money, it's ability to deliver trucks. The current entire world production capacity of 18 wheelers is driven by current market demands, its not capable of replacing the entire world fleet of trucks in a year, or even 5. So we will be autonomous truck rate of manufacturing limited.
Just some napkin calcs. In the US alone there are a total of 12.5 million heavy trucks registered (1). There were 288k heavy trucks manufactured in the US last year, up 20% from the last year (2). That works out to 43 years to replace the current fleet at the current production rate. It also passes the gut check because you'd expect to see trucks manufactured in the 80s to still be running, not so much trucks from the 70s. So even if we double production rates, it would be 20 years before we can replace all heavy trucks. That won't meet demand.
My takeaway is that there is a lot of money to be made in an add on solution that can somehow add the value of autonomous driving without manufacturing a new truck. This is low hanging fruit. It's a big deal to build and tool up an automotive plant. It's a different story when your bolting on a few parts and adding complex software controls. You don't need million dollar presses, welding robots, etc. You just need a few key components sourced from existing suppliers (motor drives for steering, brakes) a computer, likely with some nvidia edge computing stuff, and some fancy software that is hard to make initially, but easy to install on millions of trucks with the above hardware. This is a much more easily scalable business than making whole trucks, and therefore has potential to capture more market share faster, and ultimately could meet the demand faster. In a rate limited situation anything to increase that rate adds value, and I argue that this business model has a lower cost of entry, is more easily scalable, and avoids many of the pitfalls of starting an automotive/truck company.
It'll be a combination of both new autonomous heavy trucks and retrofits I imagine.
Autonomous trucks don't need stationary rest so you can halve your fleet number just from higher fleet utilization.
Battery powered trucks are also expected to have lower downtime for maintenance and overhaul, so you can shave off another 5%. Rebuilding an engine and gearbox is an expensive maintenance activity that is completely eliminated, along with brake servicing due to regenerative braking.
Once the charger situation is figured out, probably an overhead catenary and dedicated charging lanes to keep the trucks in motion while charging, the savings will be compelling.
Many long distance truckers operate in teams and stagger their driving hours, so i think this will not be such a factor.
It's not true that brake services are eliminated, these trucks will still need conventional brakes. Yes there are some servicing savings, but these trucks have already been refined to be very reliable and have long periods of time between maintenance activities. The time savings due to less maintenance will be in the noise, maybe a few percent. The cost savings will be more, less fluids, etc.
The big factor here is the labor savings, all other efficiencies are there, but insignificant in comparison and will not be the driving factors for adoption - the labor savings will.
Governments are incapable of "planning" for something like this. Even if it was likely to happen quickly. It would cost a vast amount of money and almost certainly be a bad solution, whatever they came up with. You really don't want the "expert" bureaucrats spending billions of dollars coming up with some hare brained initiative by committee that probably involves the usual mega corps who own their politicians being given vast amounts of money to do something that probably just gets in people's way.
Even when self driving gets to a certain level, it won't be able to compete with the safety of a street that ONLY allows self driving cars.
As soon as a tipping point is hit, major areas will open roads to self driving cars only.
Add to that that the latest generation of cars are electric (quiet, cleaner, more efficient, longer lasting), and as more people move to that there will be fewer gas stations, more ICE hate, etc.
The result will be a very sudden shift to electric self-driving cars. Which will create a loop, and perhaps a shift much more faster than one might think. My guess is two years from when Waymo is convinced their vehicles can go anywhere to where such cars are mainstream, and another two years till ICE are outlawed on all major roads.
Maybe living in lala land with that, but even if it takes 8 years - that will still throw a lot of people out of a job.
This is going to be singularly the biggest money making vector for any car/delivery business.
Whoever is first will be able to go to ford or any other car manufacture and buy/rent their business to 100% switch to their full-auto cars at whatever cost.
There will be no way to compete with autonomous vehicles, unless you are a train.
Even given transition would be slower 5-10years, you are still talking about massive support sector of the drivers' related economy wiped.
How do you recon all of those businesses and people adopt?
The transition will not be slow. As soon as there's an addon to an existing truck or a taxi, they will be switched over very quickly. Businesses aren't stupid, they understand human drivers are expensive and get tired and cause accidents.
Millions of jobs will be lost within a few years.
And your comment didn't address of where these drivers will go in any way. "Adapt" is not an answer. Your average 46 year old truck driver will not be learning Javascript.
How fast has warehouse automation occurred? What has happened to all the former warehouse workers? Fully autonomous warehouses has been a reasonable thing way before autonomous driving in the wild. It carries all the cost-saving and safety arguments. It might affect more people's jobs. What can we learn from that?
> Fully autonomous warehouses has been a reasonable thing way before autonomous driving in the wild.
The issue in warehouses is that even before talking about control software behind automating, we simply don't have hardware for manipulators that are as versatile as human hands. That's why top of the line automated warehouses have sort of converged on automating everything except stations where humans can act like task reconfigurable pick and place machines. Better manipulators than biological hands is still an open research problem.
Automated cars don't have the hardware barrier, and as soon as the control software meets the required error rates, we should expect a rapid switchover to automated fleets.
And driverless cars will not be able to drive everywhere at once. Initially, it will be a few predefined routes and certain times. It will take much longer than 10 years to have a driverless car that can go anywhere, at anytime.
When people hear that term they think of the robots that amazon installed that move product around the building for you and massive conveyor systems like UPS has for sorting packages. While those kind of machines are getting better and are slowly making their way into "smaller" companies (i.e. not the size of amazon) the biggest changes are on the ERP and data management side, tbh.
Cheap/plentiful handheld devices and robust/cheap wifi tech has been a game changer for us compared to the physical processes of using paper to direct staff on the floor. The processes we used 10-15 years ago are night and day to what we use in 2022, at least in my mid-sized fulfillment company.
You're probably right but I feel that the incentives to create such kits would be quite big once true autonomous driving has shown to work on large scale.
As in the technology is available to make many 'off the shelf' tractors self-driving. I have no info on actual uptake just that there are components available to make 'normal' driving in to self-driving using third party modules
10 years is a pipe dream. It will be much slower. And you seem focused only on the easy cases, when it's the hard ones that will make it such a slow transition.
Long haul drives are the edge case, last-mile trucking won't be anywhere near the first thing to get automated, people aren't suddenly going to want to sleep in their car to do overnight trips, etc. The biggest future change to fuel stops is electrification, not automated cars.
Consider for a moment that Waymo is very early into the "hey, we are allowing driverless rides on very well mapped roads!" situation. How many years have they been almost at that point already? Now put that together with the typical R&D timeline for a new car (perhaps 6 years plus or minus), and the average lifespan of a car (11 years and climbing, last I checked). No manufacturer has legitimate driverless cars in the production pipeline. We're quite a lot farther than 10 years from seeing any consumer driverless cars on the road, much less a significant enough number of them to worry about 'two-tier' driving experiences. The EV transition is going lightning fast by comparison.
If you are saying that people will still be driving cars in 10 years, then yes you are right. I wouldn't be surprised if 100 years from now there were still roads where it was legal for humans to drive cars.
But we should expect that this transition will be very similar to the transition from horses to cars. The first cars were unreliable and unsafe and didn't work well on roads designed for horses. But as cars improved the advantages quickly became huge and demand for horses dropped quickly.
The same situation is happening with Tesla. FSD is unreliable and unsafe today but once it hits an inflection point the advantages are so big more and more people become willing to accept the downsides.
I remain skeptical. The differences between horse and car were dramatic. An automated car is still a car. I suspect it may have as many downsides as upsides, to be honest (similar to things like Uber making traffic worse, not better, counterintuitively).
Early cars were designed as [carriages with the horses removed](https://en.wikipedia.org/wiki/Horseless_carriage). It would not be until many years later until we saw car designs that leveraged the unique technological advantages of horseless cars to deliver designs that looked and functioned nothing like horse-drawn carriages. We're still at this stage with automated driving.
As automated vehicles become more commonplace, we'll see designs that leverage the inherent benefits of automaton. For example, we might see designs that essentially function as moving living rooms, bedrooms or meeting rooms on wheels. We might see other designs that resemble small busses, where groups of people travelling to geographically proximate locations can carpool into a single automated vehicle. Heck, we might even see something like moving diners on wheels (potentially serving several clients at once), because why wouldn't you want to eat a hot breakfast in the "car" while on the way to the airport or work.
Once these new designs and use cases become commonplace, people in the future will view our car designs as increasingly archaic and restrictive (just like how we view horse-drawn carriages today).
None of the things you're describing actually require automation. If people actually wanted "diners on wheels" then we'd all be taking diner buses to the airport. A driverless car just frees up a bit of space; it doesn't fundamentally change the physics of how vehicles work (like cars did for horses)
Autonomous cars are going to reshape the motel market and commuting.
If you can sleep comfortably as your car drives you from SF to LA, it is a game-changer. If you can WFH from your car as you commute to the office for mid-morning meetings, you can live farther from town.
Removing the labor costs of having a driver allows everyone to experience the transformative benefits of having a chauffeur. Rich people like chauffeurs because they buy time, the most-precious commodity.
If this were true, rich people would already have cars with chaffers and beds in them. They don't because the form factor isn't practical without compromising all the other things that you'd want to use your car for. The second scenario basically already exists in the form of bay area tech shuttlesbuses; most people who use them find them a (barely) tolerable tradeoff for living in San Francisco rather than something to be relished.
No, a horse and car were like night and day. You have to feed a horse, a lot. Horses poop, a lot. And horses get stressed, they die in intensive use fairly regularly. The poop and the dying were the public health crises of their day.
It’s annoying to drive a car, but a self driving car isn’t going to massively change things. The bottleneck in most road systems are where they interface with pedestrians and cyclists (one can only cross the street so fast); self driving cars might improve highway capacities but the bottleneck is usually a backup from too many people coming or going on the same ramp into a traffic light.
True, the first order effects can be worse in some ways. Horses can easily go off road and step over obstructions and cars cannot.
The big deal is the second order effects. Cars can driver 50 MPH so we build highways and now we can have suburbs.
For autonomy it means that logistics systems can be centrally planned and programmed like never before with no need to factor in the human costs or restraints around sleeping, stopping for food, etc. Then the third order effects of that are very hard to predict
Minor point:"If you are saying that people will still be driving cars in 10 years"
Motorbikes are a thing. Apart from weight/balance relationships to control, many riders are "take control from my dead hands" types so I don't think the market will easily convert to self-driving.
It'll be interesting to see what happens. Maybe motorcycles will fade out or be banned. I will enjoy mine while I can.
seems like waymo is ahead of tesla now, just gleaning from the articles I read on HN. and because tesla only uses cameras and not lidar I don't know if this is a temporary situation or something more fundamental
Also, cars allow cities to scale further than was possible without them. If cars are banned, inner cities will start hollowing out due to labor / supply shortages, etc.
I don't share your pessimism, I think having a geofenced self-driving experience, on select roads, with the option of occasional manual override by a remote operator, is definitely feasible in 10 years.
>>I know we'll have driverless cars within ten-years
That's sarcastic, right? Like "we'll have fusion power in 50 years"? Except that we might actually have fusion power in 50 years and I don't actually think we'll have truly driverless cars within 50 years.
Just the other day my Tesla model 3 drove me home from a restaurant 10 miles away, at night, with plenty of other cars on the road, through side streets and on a highway, without intervention. Did it do weird things? Sure, but nothing that made me nervous enough to intervene. Compared to just a year ago when it was still phantom breaking on the highway and couldn’t do side streets at all. All the people saying it’s still decades away don’t realize how quickly things are progressing.
10 years is a long time and progress is not always linear. Think about what happened to the internet between 1995 and 2005.
Progress takes many shapes. Sometimes it goes quicker than expected. Sometimes the “last 5%” ends up being completely intractable. Lots of technologies got “nearly there” (supersonic flight is a good example) but never made it across the finish line.
We haven’t even begun to address emergent behaviors. You say your Tesla does “weird things.” What happens when all the cars on the road are Teslas and they do the same “weird things” at the same time (because they have the same programming)? This is an important aspect of systems engineering that we haven’t even scratched the surface of when it comes to self driving cars.
1. People "hacking" cars for fun/malice. This could be anywhere from "painting on the road/sign" to wireless comms penetration attempts or nation state attacks.
2. People who OWN the cars replacing firmware/hardware for perceived benefits. IE put back the "rolling stop" feature. Or "improve" behaviour/performance. (Car modders changing headlights in a Tesla might mess with the optical only approach) Or change the default decision in "trolley problem" events.
It's even more interesting to consider the ethical engines built into self driving. If a car has no choice but to hit a man or a woman, which one should it choose? Adult or child? Two adults or one child? Should the car save a child pedestrian but risk the driver's life? This must be built into the driving self driving system or I don't see how it can be legally approved.
Okay, so you've got a private company explicitly writing down the value of human life in a big list, top to bottom. Let's pretend we fast forward the several years to decades for the public to accept this new reality. Now we realise that different locations operate under different value systems. In Saudi Arabia, one man is worth three women. You best believe they will require the car to run down the woman. So each country/location is going to require an ethics update, and it's all going to have to be public. Imagine the heat Tesla is going to take by complying with Saudi rules and explicitly programming their cars to mow down women instead of men.
I think that given that choice the car will choose "which is less likely to hurt my passenger" since the passenger is the (paying) client and not the pedestrian. Pedestrian infrastructure & defense will have to adapt to respond instead :).
I think this is the correct answer. It will be simpler and more straightforward from a legal perspective too, since it can be argued that any driver would do the same - they would try to protect themselves first, if for no other reason than the fact that humans simply don't have the ability to evaluate situations that quickly and play through a trolley problem in their head while the car is heading for a collision - you instinctively do whatever action feels right to protect yourself. So I think self driving systems will be coded with that imperative as well - protect occupants first and foremost, and they won't be calculating the "worth" of whatever or whoever they are about to hit. The extent of damage calculation will end at "there is one occupant in the driver seat, so the damage to them will be lower if I hit the obstacle with the passanger side of the car", but it won't actually know or care what the "obstacle" is.
Modern cars are already hackable - there has been several demonstrations where people connected to the car stereo, and through a chain of exploits could get access to the streering/breaking ICUs, essentially taking the car over.
Yet there has been no report of this exploit used maliciously in the wild.
If you can do 95%, that might be good enough. Obviously if you could automate the entirety of the process, that would be ideal, but consumers may simply change their behavior to accommodate the 5% difference. Uber/Lyft already changed consumer behavior in a marked way: they allowed more people to move closer to the city, not pay for a car, not pay for a parking spot, etc. In Seattle before the pandemic, it was common for young professionals that didn't own a car to take Uber/Lyft most days. They did this so frequently that yuppies were blamed for the increased traffic, even though many of them were served by Seattle's great mass transit options.
The cost, simplicity, and availability of self-driving cars will undoubtedly change the consumer.
Self-driving cars have been at 95% since the 90's. They've been at 98% since 2012. Humans are higher than 5 nines.
It turns out that human drivers, even poor ones, have a much better driving record than self-driving cars. With this latest Waymo experiment we may finally see some significant progress.
> 10 years is a long time and progress is not always linear. Think about what happened to the internet between 1995 and 2005.
And then compare it to what happened to the internet between 2005 and 2022...
Progress isn't always linear but it tends towards diminishing returns more often than exponential progress.
And we're not talking simple things like "can billions of people and Moore's law result in something a bit more exciting than Facebook with a newsfeed and ads and YouTube without Flash", we're talking about progress in the form of monotonic improvement in complex intractable software to solve edge cases the dev team haven't even thought of where the failure mode involves corpses, and do it to the satisfaction of regulators across many jurisdictions.
The gap between "drive 10 miles without serious incident this time" and "consistently drives a billion miles between incidents serious enough to be fatal without any possibility of intervention" is enormous.
Do you live in a place with weather? I think we'll get there eventually, but my timeline is closer to 20 years. I just don't see AVs nailing that last 1% for true peace of mind in places like the Eastern USA with torrential downpours in South Florida or snow and icy conditions in New England.
I hope I'm wrong though, I would prefer it if my kids and other road users didn't have to drive.
Apart from weather, a big issue outside the USA is also car infrastructure. The US is extremely car centric with large, straight, easy-to-maneuver roads everywhere. Here in Germany, it's different. We have rather narrow, chaotic roads, unclear signage, "right car has right of way" traffic rules which sometimes get resolved via hand signs, no jaywalking laws, etc.
I am still hoping for fewer cars on the road overall. The car itself is inefficient and hopefully on its way out.
I think about all the times during winter when my car's traction control straight up gives up as I try to accelerate through an intersection or when I need to start slowing down an extra two blocks away because the ABS goes on every time I try to stop. I have no idea how AVs will nail that stuff, it's so much harder than keeping the car between the lanes and sussing out road signs.
I think about all of the lanes in my city where line paint disappears intermittently, or where old lines/construction-era lines mess with my ability to comprehend where the lanes are, let alone an autonomous vehicle's.
And then I think about all the potholes that I dodge when I drive around town. And how 4-ton electric vehicles are going to make road wear worse, not better.
The physics and control of cars are quite well understood, even in extreme conditions. Further, sensors allow a car to detect changing road conditions much more quickly and adeptly than a typical human.
The hard part of self driving seems to be understanding and predicting the humans in its environment.
That example is still a car on a known good road conditions. Do you have an example where half the parking lot is covered in ice, or trying to break on a road with unpredictable traction so that they don't end up halfway through the intersection?
> Do you have an example where half the parking lot is covered in ice, or trying to break on a road with unpredictable traction so that they don't end up halfway through the intersection?
If the car is "smart," maybe it will simply refuse to travel (because the calculated risk of traveling over an icy parking lot is above some maximum threshold risk).
Humans have no trouble in these conditions (after a brief training period).
What happens the first time a steering wheel free car strands a family in a blizzard due to a few feet of icy pavement, ultimately freezing them to death?
I don't think cars will ever remove steering wheels.
That was just Elon's fantasy for a limited use (driving through Boring tunnels). Similarly how you don't have a steering wheel in a driverless metro.
Or perhaps it will be retractable when not in use, like a tow hook.
Isn't that like saying that motorcycles will happen eventually but not anytime soon because: weather?
Traditional cars with drivers haven't solved the weather problem 100%. But they are here now. Sometimes we ban them in some conditions, but they are a game changer the rest if the time.
Motorcycles are only on road for maybe 5 or so months out of the year where I live. If that is the best we can ask for out of autonomous driving, it'll never be more than a luxury toy.
One thing that is going to be awkward to navigate in the transition to self-driving cars is how they handle driving in bad weather. Obviously that's a technical problem but I'm not sure it's actually solvable because it's not actually solvable for human drivers. We just kind of do it anyway. But when a self-driving car has some quantified sense that "driving in these conditions is unsafe" will we get to the point where we just say f it and let them keep going? Or, maybe more likely, we remain in the gray zone where the car is fully self-driving but the human driver has to be ready to take control at any time which I think limits some of the larger impacts.
I mean if you live anywhere with weather your motorcycle lives in the garage for 3/4 of the year. Not useless by any stretch but also not “taking over the world.”
It’s not that traditional cars have solved for weather, it’s that self driving cars haven’t solved the problem of “being as good as human drivers in weather commonly encountered.”
And computers are better at reacting to unexpected things like icy road. And cameras are better at seeing through rain and fog. What's barely perceptible to you, computer sees clearly.
>>You're wrong. It will happen way before 20 years
Other than random people saying this on the internet, I have seen zero indication that this will be the case. You cannot will something into existence just by strongly believing in it.
Are you just saying things that you think are true? Have you ever tried to do even rudimentary image processing in foggy / rainy / nighttime conditions? Cameras are absolutely not better at seeing through rain and fog.
Humans aren't that good at seeing in rain and fog either (which is why many people avoid driving or drive much more slowly in extreme rain and fog conditions), so saying robotic vehicles aren't good at driving in inclement weather conditions isn't very interesting. In addition to cameras, Waymo and Cruise vehicles have multiple lidar and radar units. These are all affected differently by inclement weather but in principle combining sensor data for perception gives the vehicle a lot more data to work off of than a human relying solely on vision.
Driving in truly bad weather is still an open research problem, but it's also one that most of these companies haven't dedicated a lot of resources to, since they all have initial rollout plans in places that don't experience much extreme weather. But assuming these vehicles can drive well in good weather, I don't think it's a stretch to assume that progress will be made on driving in more extreme conditions once more resources are dedicated to it.
My Tesla still phantom brakes. And Tesla has been trying to solve it for, what 10 years ? My rain sensor also barely works. I have seen enough FSD videos and all the scary things it does. There is no way it will be ready to solve the last 1% in the next 10 years
I rode with my uncle in his new Tesla car in rural Pennsylvania this winter. Clear weather, daylight, only a few other cars on the road. In a 10 mile trip the automatic emergency breaking threatened to kick in with blaring alarms 5-6 times: whenever an oncoming car approached us in the other lane, and over hills where it apparently thought we were driving off a cliff. I thought it would do better than that, given how much it's touted as so good.
Tesla fans remind me of Linux fans. "It can definitely 100% do this thing... if you don't mind x, y, and z. Also you need to do a. And b. Some sometimes c if a does j. Also it doesn't work on Tuesday. But that doesn't matter because you shouldn't need to use your car on Tuesday anyway."
I dig the commitment and passion, but as a consumer, I want the thing to work perfectly first time, every time. If it doesn't, I don't want it.
Did it do weird things? Sure, but nothing that made me nervous enough to intervene.
Doesn't sound like much of a data trail. No one disputes that it at least sort works (while also doing "weird things", as you put it) in some conditions. As long as, you know, nothing weird happens out on the road. Like for example rain.
It's the fat tail of "weird things" (and how these systems currently handle them) that leave the question of whether FSD will ever arrive -- very much up on the air.
Totally agree. People become SOOO embrace full self driving so quickly.
I know 5 people with the Tesla FSD Beta, and EVERY SINGLE ONE of them has admitted to letting Tesla drive them home when they drank too much and went past the legal limit. And not a single one of those 5 people has ever had a DUI. So FSD rapidly changed there drinking behavior.
Also I get that I have an atypical and shitty social circle. I don't think this represents all Tesla drivers.
Yeah consider me unconvinced that someone willing to do tesla fsd drunk wasn't also willing to drive drunk themselves. Lack of a DUI is hardly proof otherwise.
I'd say closer to 50 years myself. But my standard means that we not only have near-perfect self-driving, but we're confident enough in its quality that we're willing to manufacture cars without conventional physical driver controls, and people are willing to ride in them. And then to actually go through the process of replacing the entire national inventory of conventional cars with these new cars without controls.
No way. At least 50 years to get somewhere. But we need computing power 1000 times better and optics at least 16 times better. Before that it never can fix all scenarios
Here is a video of a customer vehicle driving on snow covered roads, now. It isn't perfect, but it is leagues better than a year ago. https://www.youtube.com/watch?v=oDK3dRHxOzo He has tons of other drives in various weather/road conditions.
Improved software, and maybe one hardware refresh should be sufficient to get us all the way there for 99.99% of drives.
It appears that if it doesn't know where the lanes are, it drives in the middle of the road. If someone turned onto the street from 0:35 to 0:42 there would have been a problem.
I'm still waiting to see how an AI would handle Figure 5-1 from that. And no, the snow plow's solution of burying magnets in the road isn't practical (Cost of infrastructure installation for the test sites is approximately $11,000 per kilometer ($17,000 per mile), including surveying, installation, and magnets).
I’ve been hearing “ten years” for ten years. I’ve learned not to trust the opinions of tech people. They’re really good at fixating on only the pieces of a whole problem that they’re familiar with or interested in.
Sure, and in much the same way robot chefs that can cook a whole meal exist. You just need to open the door put your food in and push the start button and 30 seconds later your food is ready.
You can, today, order a car and go to sleep in the back for 20 min while it drives you to your destination and get out with no intervention. Can you explain how that matches up with your analogy?
Just because something is semantically similar doesn't mean it is what people expect when talking about an object. I'm implying with that comparison that most peoples definitions of a self driving car is not a car that runs in such a limited set of circumstances, just like most peoples definition of a cooking robot isn't a microwave no matter if you start pointing out sensors and movement to ensure better cooking. It may be a no true scotsman fallacy but for me (and I'm assuming a chunk of the people in this thread who are still talking about how long into the future until self driving cars are ready, in the comment section of an article about exactly the "self driving cars" you're talking about no less) a self driving car that I can't buy or hire (at any cost) as a replacement for my current car and have it self drive my commute to work or an address on the other side of the country is not a self driving car.
That opinion doesn't diminish the progress that has already been made or the accomplishments of the various engineering teams involved. It's just saying that when people talk about self driving cars they're talking about it as a drop in replacement to the general population of car owners' cars not as a curated location/digital track limited taxi service.
Ten years doesn't sound sarcastic to me. With fully driverless operation starting in SF (and with several years in Phoenix), having it widely rolled out within ten years seems very plausible.
They just announced 500k driverless miles in Phoenix. Over several years that sounds laughably small to me. For comparison, motor vehicle fatalities in the US are around 1.5 per 100 million miles. We're several orders of magnitude away from even being able to convincingly demonstrate that these reduce (or at least do not increase) fatalities in the real world.
I'm glad that Cruise beat Waymo to driverless in SF because they are providing the swift kick in the butt that Waymo seems to require to actually make progress in a reasonable amount of time.
Fatalities are the metric Waymo uses every time they give a presentation. They always start off by telling us how many people die on the roads and how they can help. And it's true! I think it's a very important metric and they are right to focus on it.
Simulated miles are not convincing. Miles with a safety driver are better but still ultimately different. Also the type of miles matters too, and Waymo has been focusing on the easiest miles possible until Cruise forced them to up their game and try something a little more valuable.
>Fatalities are the metric Waymo uses every time they give a presentation.
Source? Because since 2020 Waymo has been primarily touting their "contact events" metric, which is what most people consider to be accidents.
Simulated miles are still better than no miles. Given the better accident rate it seems to contribute to a degree.
The point is that it's an argument from ignorance to narrowly focus on one metric and say we don't have enough information on whether autonomous vehicles are safe. We do have a range of metrics and information already that point to them being safer.
There are countless other examples. It's one of the first things they bring up in almost every presentation to the general public. It's how they present themselves to the world, as the answer to traffic fatalities.
You may think it "ignorant" but the fact is that fatalities are going to be the single most important metric that the public and journalists and governments are going to focus on, and that's why it matters. The first few driverless car fatalities are going to be a huge, huge deal and are likely to result in action from all three branches of government. There will be court cases, there will be regulatory action, and there will be a lot of noise from politicians about laws and maybe even some laws passed about it. The first one already forced Uber out of the game.
Provided your starting point is AV software with some ability to drive, this isn't necessarily true. Tuning on simulated miles can lead towards over-optimization for the conditions of the simulation.
Your link says Autopilot, not FSD. Autopilot is highway autonomy, which is vastly easier and you can rack up many more miles very quickly. This is not at all a relevant comparison.
The mere fact that it won't panic on black ice is already a huge advantage. And it will have very detailed feedback to and from the traction control systems.
Oddly enough from what I've seen when it snows I'd imagine a computer's ability to simulate a multi-car situation in snow _vastly_ outmatches your normal humans.
> Oddly enough from what I've seen when it snows I'd imagine a computer's ability to simulate a multi-car situation in snow _vastly_ outmatches your normal humans.
What's the basis for your imagining? Whenever I read some strong "computer > human" statement like that, I can't help but think it's substantially based on science fiction (which can make technology work fantastically well because it's not real, e.g. like Six Million Dollar Man vs actual prosthetic limbs).
Also, have you heard about the time I was driving on the highway in winter weather, and my adaptive cruise control would stop working every few dozen miles because I had to get out and scrape ice off the sensor?
I think you might have missed the implication that the humans are doing a terrible job, so a computer doesn't have to be very good to beat them in this situation.
> I think you might have missed the implication that the humans are doing a terrible job, so a computer doesn't have to be very good to beat them in this situation.
That doesn't really matter, because I was responding to the the sci-fi sentiment that "of course the computer must be better because computer." It doesn't matter if the humans are doing a poor job if the computer is still worse.
Especially because the claim is only about snowy situations. Even in a reply about snow, someone that thinks computers are automatically better would normally just say that directly.
Nah, your understanding is incorrect. I've used it in Phoenix and it goes places a bus wouldn't like residential culdesacs and parking lots (whithin the geofence).
Lower stakes because of the speed, but a lot of special cases. Higher chance of pedestrians and parked cars in the road for both. Parking lots will have a lot of cars moving in atypical ways, and markings aren't as standardized as on roads.
Self-driving is not even on the same scale of complexity as a board game. The latter is an inherently digital problem, easy to parse and not played in “real time”
People's inability to accurately assess how easy an inherently easy problem is has no bearing on people's inability to accurately assess how hard an inherently hard problem is.
Said differently, I start from the assumption that some problems are obviously easy, some problems are obviously hard, and every other problem is somewhere in the middle. Self-driving is in the "obviously hard" bucket. Feel free to argue for where you think Go sits in the spectrum, but I would argue no board game is in the "obviously hard" bucket
Hard and easy defined as the ability to solve in a reasonable amount of time with the algorithms and computational models we have today or in the near future -- regardless of ~marginally increasing processing power.
> People's inability to accurately assess how easy an inherently easy problem is has no bearing on people's inability to accurately assess how hard an inherently hard problem is.
Really? It seems to me that if people over-estimated the difficulty of an "easy" problem like mastering go, then they're even more likely to over-estimate the difficulty of a hard problem like self-driving. In fact, the over-estimation could scale up faster than linearly, if estimating two problems of size X is easier than estimating one problem of size 2X.
> with the algorithms and computational models we have today or in the near future
That's the thing. When predictions were being made about the difficulty of mastering go, people didn't have the algorithms and computational models that we have today. Similarly, predictions made today about the progress of self-driving cars may be lacking critical information about the algorithms and computational models that will be available in the near future.
I saw my first self-driving car parked at Google in roughly 2008 or 2009, and watched them toodling around campus shortly afterward. Many people (like Larry) thought it was a 5 year project.
There is this famous story about research into image recognition back in 1960s - everyone involved thought it was literally a matter of few months, maybe a year at most, to have a system that has 100% accurate recognition capabilities.
Yet we're in 2022 and the best of the best image recognition systems spew out absolute nonsense no matter how much money is poured into them.
It's just the same sort of issue. It can work amazingly well 90% of the time, but the remaining 10% is basically impossible without some kind of actual general AI that can understand the context of what it's looking at. I honestly believe that current approaches to this tech will not get us to general self driving tech that is safe enough for use on our roads.
What do you mean spout absolute non sense? The progress in image recognition has been insane in the past decade so your comment might have been accurate 10 years ago. If there's one thing DL/ML are very good at it's computer vision
Sorry I thought it was clear from the rest of my comment. Image recognition is amazing in 90% of the cases, and then in 10% it's worse than a blind human. Another famous example of that is the(now solved) Google image recognition saying with 99% certainty that a sofa in a zebra print is in fact, a zebra(4 legs, zebra pattern = must be a zebra). Look at any demonstrations from Tesla at what comes out of their clasiffier as the car drives around, sure, most of it is correct but some of it is insane, the car thinking there's a ship or a cow in the middle of the road when there's nothing there. Sure the car doesn't act on it, but the fact that the recognition gets it so catastrophically wrong is indicative of the quality of the whole thing.
Another story in a similar vein. There are many, many stories about how the advent of DNA sequencing would finally solve the issues of creating the tree of life for all species on Earth, and solve the problem of defining species boundaries, this was some 20 years ago or more. We're still nowhere close thanks largely to the unimaginable quantity of biodiversity out there. We have the low-hanging fruit genomes (specimens readily available, known to be commercially important, etc), but that's about it. Even with those "genomes" (most are not curated to any real level of completion) we still haven't solved compute at scale, even small analyses (10k terminals, 2k genes) take months if not years to comprehensively run (i.e. explore parameter spaces), albeit thanks in part to limits on resources available to scientists.
The Earth is big and complex, I can't see the singularity on the horizon yet, but my vision is aging.
I solved the computational problem for genomics already, since it's a mostly-embarassingly-parallel problem (in paticular, I made an idle cycle harvester at Google that applied 1+M xeon cores to science problems, one of which was genomic assembly).
I don't know that creating super-accurate species trees is really the most important problem to solve in biology, though. Certainly DNA sequencing for health has been a really mixed bag.
Nice work with that solution! I did similar! I solved the earth's hunger issues, funny thing is we produce enough food, we just can't deploy it to the people who need it. Problem is I can't figure out the logistics of how to make it work... I think maybe you have the same problem? We should compare notes.
I don't have a complaint about sequencing that leads to vaccines. That's what sequencing is for.
What i'm troubled by is the projects that sequence a million people, using $1B to do so, and then just dump a bunch of genomes with some hand-wavy claims about how it's going to cure cancer. The costs here are rathre significant, but the health outcomes are not.
Another example. An airplane broke the sound barrier in 1948. Today, we are as far removed from Yeager breaking the sound barrier as Yeager was from a civil war hero being President (Grant). But we still have no supersonic airplanes in commercial service.
tbf, lack of supersonic airplanes is a commercial problem rather than a technical one.
On the other hand, most of the commercial jets we do fly are basically 1950s designs with bigger engines and some computers in the cockpit, and HN thinks it was crazily irresponsible for Boeing to have decided to use sensor-triggered autonomous systems to do something as simple as temporarily adjust the trim system...
Likewise, a lot of computer scientists didn't take Engelbart's 1969 "mother of all demos" seriously because they thought general artificial intelligence was right around the corner. Why bother with a mouse and bitmapped display when you can just tell the computer what you want?
the mark i perceptron did image recognition with a single layer neural network and was surprisingly good at it. Some days I think Rosenblatt was on to something, but about 40 years ahead of the time.
That was wrong, clearly. But it was a prediction made with far less evidence than we have today. Considering progress so far (which we had far less of in 2012) I don't think ten years is wrong today.
And even if it is wrong, it's obviously not stupid to think we'll have driverless cars in 10 years. Waymo are literally down to solving edge cases. There are more than people might have expected, but there isn't anything fundamental to solve.
But: there are very large numbers of these edge cases and they are of the kind where new ones keep popping up all the time. This is a very easy problem for the first 99% or so and then the last 1% is super hard.
Sure - but they've been working on the last 1% for years now. Nuclear Power plants are almost always over budget and over time, big misses. And they've built lots of them. Every piece of software ever built suffers from this issue to some degree. Elon Musk misses every deadline he ever sets. Something taking longer than initially thought shouldn't instantly set off alarm bells suggesting it won't ever get done.
> Sure - but they've been working on the last 1% for years now.
Indeed, and we aren't any closer compared to say 5 years ago. That is more or less my point: you can solve for some of this and then you are still left with a mountain more. And then at some point there is an even harder problem to deal with: how to ensure that fixing one problem won't regress one or more others.
I'm sorry, I evaluate tech for a living and I don't see us being closer in a way that moves the needle. Objectively, however long it will take to get to self driving we are 5 years closer, but that doesn't really matter. What matters is whether or not the state of technology is at a level where it will enable self driving to the point where a car will no longer have a steering wheel. Because less than that isn't going to cut it.
And that may well require another leap of capabilities on the AI front, it might even require GAI. But you are entirely welcome to your own opinion, as I am to mine, but after watching the self driving industry since the 80's my timeframe is something like 25 years before fusion will arrive.
There may be some intermediary 'artificial artificial intelligence' solution where when stuck a remote driver can take over. But that has its own set of problems and I don't want to set this up as a strawman.
Ok, let's see one drive through Amsterdam tomorrow morning at 9 am between two random zipcodes and house numbers. I'll bet you it can't do it. "Edge cases" indeed.
I'd take that bet at even money all day long. But, maybe it can't. It doesn't change the fact that they are materially closer than they were 5 years ago.San Francisco streets aren't a snack. There are hilly, narrow streets with cable cars, constant construction, homeless people wandering onto the street.
It's real progress. Anyone around Mountain View, or SF will have seen the progress with their own eyes. It's not a theoretical debate.
But that's the whole point: if this were a reality there wouldn't be a maybe there. So thanks for making my point.
"Materially closer" isn't the bar to cross the bar to cross is "It just works" which is for me what I'm getting at when I say that we are not 'closer', because it's a binary thing, arrival means 'it works' not that we have made some incremental change towards an ill defined goal.
The 'it works some of the time or even most of the time' bar was crossed years ago, but it needs to work all of the time.
Let me give you an analogy. Say we are on Earth looking up and we see the moon. We decide we want to go visit. Every day we add a couple of stones to a large pile on top of a mountain (those are the edge cases). We are provably getting closer to the moon every day. But we will not get there because we need better technology, we don't need piles of stones, we need a rocket.
In your view we are within striking distance of the moon, another couple of piles of stones and we are there. In my view, I'm not sure that it will just take more stones and not a better technology, and I'm beginning to lean more and more towards the latter because the problems that we were seeing 10 years ago we are still seeing today: more and more edge cases. And edge cases can be a symptom of a problem that is almost solved or it can be a symptom of a problem that is being solved using the wrong toolset, especially when there are lots of them and new ones keep popping up all the time.
Clear now?
And by the way: none of this is to imply that what has been done isn't impressive, the MB statement that they will accept liability if their software is in control of the vehicle is a step in the right direction and shows a lot of confidence. But it's incremental progress towards a goal that might not be reachable in that way.
This whole line of reasoning sounds like a philosopher debating the number of teeth in a horse's mouth.
Just come visit San Francisco and see the evidence. Look in the horse's mouth, so to speak. The cars are driving around a complex city, without a driver, and doing an amazing job. The progress is real.
>How would materially closer not meet the bar for closer?
I don't have a horse in this race, but I think what the other commenter was getting at is that: if there were 10 problems that needed solving 5 years ago, and in the course of solving those 10 problems you discovered 100 more than needed solving that you weren't initially aware of, then sure you got closer to the goal, but the progress was so tiny that it's almost not relevant.
I'm personally not arguing one way or another, but that's how I read the other person's comment.
I guess that would be a reasonable take if Waymo announced they were "close" or had solved "x technical problems". But here progress = cars on the road with no driver in more and more difficult situations. So, even if more problems keep coming up they seem to be overcoming them at a good clip.
fwiw, I agree that progress has been astounding in the past 10 years... but on the other hand I really do wonder what the limits of the current approach is. After all humans don't drive by analysing 1 frame at a time, we have a model of how the world works, we anticipate, we have intuition etc...
I know the whole human flight is different to bird flight argument, so I guess we'll just have to wait and see how it all plays out. Exciting times :)
> but on the other hand I really do wonder what the limits of the current approach is.
If the technology stagnates, the hardware stops improving, and the software stops getting better, then we would hit the limit pretty quickly.
> After all humans don't drive by analysing 1 frame at a time, we have a model of how the world works, we anticipate, we have intuition etc...
All self driving cars also maintain those models as well, and use frame by frame input to correct those models in real time (but they have already guessed where something is going to be 100+ frames from now).
Isn't the kind of driving that waymo and others have been doing so far the edge case? Normal/ideal driving conditions is not what I spend the majority of my time driving in even if the different kinds of edge cases I drive in are individually less of my time.
Do we? Where? The best version currently in existance could be (generously) described as an automated bus that only drives on a pre-mapped grid and still needs human supervision because you can't quite trust it yet. The most well known commercialy available "autopilot" (from Tesla) is so hilariously bad that I honestly don't understand how it's even legal to use, much less to sell for real money.
Neither waymo nor cruise currently require human supervision. That's the whole point of this announcement: waymo is expanding their driverless program to SF.
Whether the grid is pre-mapped or not isn't particularly relevant. It's just how the technology works. The vehicles aren't driving on rails, they just have more semantic information about the world around them.
Even wide distribution of long-haul trucking exclusively on the interstate, or an automated bus-like commuter service (but actually good, unlike the regular bus), would be a pretty meaningful innovation, wouldn't it?
Of course, and it would be awesome, and I believe that sort of thing is actually possible within a relatively short timescale. General self driving like 90% of people do every day? Nope, don't see that.
Please explain how fossil fuel companies can keep auto manufacturers from selling electric cars? The reality is that the biggest obstacle continues to be the high cost of battery raw materials.
Depending on what you mean by widely rolled out, I agree. Driverless trucks driving on the highway and in well-marked industrial areas seems extremely plausible within 10 years. Driverless taxis at slow speeds as a shuttle service on campuses and fair grounds are more or less already a thing and will only grow.
Driverless cars in a European city center, during snow, ice and rain seems unlikely within 10 years. But we don't need that to see transformational change.
> Driverless cars in a European city center, during snow, ice and rain seems unlikely within 10 years.
The fun thing about European cities is that many of them are extremely willing to make (to Americans) unthinkable changes to traffic patterns under their jurisdiction. I would not be the least bit surprised to see those downtown centers closed to all but driverless traffic (if they are not closed already).
Maybe, but I don't see driverless cars ready for bad weather.
BTW, humans are not ready for bad weather either, even though we drive in it all the time. However with driverless cars we will collect statistics proving it isn't ready and never noticing that it is better than humans.
It's borderline obvious we will have a taxi/Uber replacement for most of lower Manhattan, east San Francisco and southeast Phoenix before 2030. (Technologically, at least.) Less confident in long-distance interstate-only driving by 2030, but that's better than 50/50.
Beyond that, the edge cases are innumerable, so it's difficult to estimate. Nevertheless, we have clear line of sight to domains where, for all intents and purposes, a human-driven vehicle will be obsolete (if not necessarily yet unquestionably uneconomical) in well under ten years. That's far more than we have ever been able to say for fusion.
If you asked me 20 years ago if cars would drive as well as they do now autonomously, I would have made the fusion power joke. Today, no way. The technology hasn't gotten "stuck" like fusion or (before DNNs) speech recognition did.
I agree with you. I used to be in the "driverless car in 10 years" camp. But after driving a Tesla for the last 5 years, I know it is close to 50 years. My Tesla still does phantom braking. So forget about dealing with weather, construction, people, traffic, weird turns, weird signs etc. There is zero chance I will turn Tesla FSD on in a place like Market street in San Francisco
For a lot of the things mentioned such as highway driving and sleeping in the car, auto recharges, etc. I don’t think it’s far fetched at all. For a driver assisted version of the others, meaning autonomous driving lanes with precision timing I don’t think it’s far fetched either given it’s basically building the infrastructure of highways in cities, but even if not, I can completely imagine same but a driver is required to be behind the wheel and aware.
I have a mostly driverless car right now parked in my garage. My Tesla does about 95% of driving for me.
I rode in a fullY autonomous car, with nobody in the drivers seat, about a year ago (we have had fully autonomous waymos in Phoenix for a few years now).
Yes, a do you not find yourself anxious about the remaining 5%? I've seen FSD do some batshit insane stuff on the road that even a geriatric blind driver wouldn't do. And it's not like you have any time to react either, like, you have to take over the steering NOW because the car decided to just give up in the middle of an intersection.
Not at all, because I know that if the driving conditions change and I have a feeling the car won't do well, I take over. In other instances, where I know the car is going to do just fine, like driving on the freeway, I enable it.
FSD/Autonomous driving isn't an all or nothing scenario for me. I use it when I deem it safe to be used, and drive manually every other time.
The remaining 5% are almost all situations where I take over because I’m underestimating what the car is capable of.
For instance: I don’t like how close it drives to the curb. I don’t drive close to the curb (and bias towards the center of a road) because I only have two eyes, and can’t really see the curb as I’m moving. I have an intuitive sense of how close I am, but that’s it.
The car knows exactly how close I am. So it driving in the center of a lame, while it makes me nervous because it’s now how I drive, is perfectly acceptable.
What I wish Tesla would do: set up some closed courses for people to come and experience what FSD is actually capable of. It would help the current FSD users gain confidence in the system, and it would show it off to people who haven’t ever experienced it.
I seriously don’t believe that most of the people who are critical of this thing have spent much time with it. I’ll admit after the first day or so I was skeptical too, in the same way I was skeptical of lane keeping and proximity aware cruise control.
But at this point: I hate driving without these things. They are an almost indescribable upgrade to the driving experience for me.
I was really disappointed with the service area though. I was hoping it would be available as you say "in Phoenix", but it's only available in Chandler which is just one suburb south-east of Phoenix. Service has been this way since it was introduced a few years ago. I really expected them to expand by now.
Seems like a strange take. Companies are consistently making measurable progress in driverless cars but Fusion is barely inching forward over 5 year increments.
Given that by some reasonable definition we already have driverless cars, albeit limited to two cities, why does that sound implausible to you?
Cars that can drive everywhere, in every condition, without humans intervening at least remotely and occasionally in some limited way? Probably not.
Robo-taxis in most major cities and trucks that can reliably handle the full interstate network, to be picked up by a human driver for the last mile? Why not?
Well, it depends on your viewpoint. Just today somebody celebrated a fusion Q of 0.005, while each year since many now Tesla‘s Autopilot pulls off more accident-less miles than humans.
Sure, it‘s apples to oranges, but for self-driving, we‘re solving the last (and obviously hardest) percent, while for fusion, we‘re still stuck solving the first.
We'll probably have a lot more partially automated vehicles in 20yr.
Look at advanced ports and inter-modal yards. That sort of human to vehicle ratio will probably move up the spectrum of complexity in which it can be deployed and down the spectrum of cost.
It does not look far fetched at all. Current issues stem more from legal wrangling over who is liable than whether cars can drive in 'driverless' mode. I don't know what the future holds, but it is nothing like fusion.
>>Current issues stem more from legal wrangling over who is liable than whether cars can drive in 'driverless' mode.
Honestly don't believe that at all. The tech is immature and works only in a very narrow set of circumstances with a lot of constraints - it's so far away from being functional that the whole legal discussion around it is almost academic at this point.
Seeing is believing I suppose[1]. I dislike both Tesla and CNN, but it is hard for me argue from this video that the technology is immature. I stand by my original comment. The tech is here. At this time, the main question is who picks up the death toll responsibility.
I'd say 30. War and climate slowing the money and motivation. But it's really close to get done. It need one more leap and I don't consider it impossible at all.
There's this great compilation video on youtube of Elon Musk promising fully autonomous cars "next year" since 2015. https://www.youtube.com/watch?v=o7oZ-AQszEI Sorry for the dumb clickbait title.
Those are true/eventually did happen up until 2018
2014 - 90% capable of autopilot, for sure highway travel
2015 - autonomous driving for highways and relatively simple roads
2016 - S/X can drive autonomously with greater safety than a person - generally true, it still messes up for human-correctable situations but in general helps more than it hurts in terms of "crashes where autopilot was in control during impact or within 5 seconds of impact"[0]
2018 - incorrectly predicted - self driving will be 100-200% safer than a person by next year (this talking about their full self driving thing that will do turns and stop signs and whatnot)
People, for some reason, still confuse AutoPilot (literally just ACC) and Full Self Driving. I guess the creator of the video doesn't know the difference either. The naming is really bad though.
A blueprint for handling this can be cribbed off of "just transition" plans for folks in fossil electrical generation (coal) [1] [2] who are going to be out of a job in the next 5-10 years. Looking back at history, the productivity gains from automated vehicles can be split with the firms operating these fleets with those being transitioned out (as happened with the invention of the cargo container and Longshoremen's unions [3]). Another resource that covers this topic (in part) is The Box: How the Shipping Container Made the World Smaller and the World Economy Bigger [4].
> solar panels require massive amounts of coal to produce
I'm would prefer to not derail this thread (as I really wanted to focus on the existing economic solutions for industry transitions), but this is patently false and can be determined as such with 5 minutes of using a search engine with terms around solar EROEI (energy return on energy invested) and the electrical generation mix of grids where solar panels are produced.
He's not disputing that, he's disputing your "shell game" comment. It's not a zero sum, and solar panel creation is better in the long run compared to burning coal.
That's not what he quoted so I misunderstood. I was responding originally to the person insinuating that we need to find new jobs for people in the coal industry. Regardless of the energy returns, solar requires coal which would require people to work in the coal industry.
2. Approximately 90% of most PV modules are made up of glass, so let's assume (incorrectly) that the other 10% (i.e. 2 kg) is PVC
3. It takes three pounds of coal to create one pound of polyvinyl chloride (PVC), so 6 kg of coal
4. 127 GW of new solar capacity was added in 2020, which would require 127 GW / (300 W / 6 kg) = 2,540,000 tonnes of coal
5. Global coal production in 2019 is estimated at 7.9 billion tonnes, which is 3110 times more than is needed
6. The coal mining industry employed 42,117 people in the United States in 2020.
Therefore if that number of people could be scaled down linearly with the amount of coal, to cover just the production needed for solar panels, less than 14 people would be needed to work in the US coal industry.
>>but has any government yet started planning for the relatively near-instant transformation of the Trucker and Taxi industry
Let me put it this way.... How often do any governments have a plan for changes *beforehand?."
Almost every western country saw "deindustrialization" at some point, predictably cutting off whole cohorts from wages, pensions and lifestyles. None had a plan.
They watched Walmart behead high streets and then watched Amazon behead the survivors. No plan. War is, arguably, the thing we plan and provision for the most. Wartime refugees always seem to take us by surprise.
I'm sure someone somewhere, in theory, had a plan for a pandemic. But when it happened, a whole lot seemed to be off the cuff.
> * Fake taxi's that drive a customer into a bad experience for "lols".
Some of the other points seem legitimate, but this one seems silly.
It's trivially easy _now_ for someone to sign up for uber and drive people to random locations, and that doesn't happen.
Presumably, with driverless taxis, we'll still have the same level of assurance we have now (use the app, app tells us a license plate for a valid car, that's the one you get into). Right now, you can get a driver who misbehaves, or uber could get hacked and send incorrect dropoff locations. In a driverless world, only the second attack matters since the chance of a malicious driver is gone.
I, personally, am waiting for the first self-driving car flash mob where everyone on 4chan decides to order a taxi to pick them up at mission & embarcadero (as an example) to swarm and DDoS intersections - or try to route a lot of requests to an intersection cut off by a parade. If self-driving cars were a thing last year do you think antifa folks would order a bunch of cab pickups for key intersections in the middle of the Jan 6th march?
Humans would error correct - if there are too many uber drivers in the area already new drivers would drop the ride request. I think if you managed to remove all humans from that process you'd be able to get some truly bizarre scenarios.
It will be interesting, but I'd guess the "near instant" part is wrong. Goods need to be moved around - trucks are in constant use. It will take a long time to retro-fit trucks or manufacture new ones. It will still be positive ROI to have a human driving an existing truck for years. Taxis is harder and could happen faster, but I'd still bet it's a 5-10 year process.
I hope there are new limits to trucking and a concentration of those that are licensed to off hours. I-80, for instance, already feels like a truck route more than a human commute route. That said, so many of the accidents (jack knifing) there are trucks in wintery weather, I don't see that being included v1.0 of the new trucking reality.
I-80 always comes to mind when these autonomous vehicle discussions occur. I'll believe in autonomous vehicles when I see them driving across Pennsylvania during a winter storm. There are days where I seriously doubt an autonomous vehicle would be able to navigate out of my driveway.
Your comment makes me think about the interplay of driving and loading. A properly loaded truck can probably be driven by an AI in a way that greatly reduces jack knifing compared to human drivers. But what about improper loading - and if done on purpose to sabotage the endeavour.
I am a bit petty. After my last speeding ticket in a small drive through town I thought to myself "In a decade or two, these police departments will need to find a new way to fund themselves"
More like they will actually have to fight crime or find new jobs. I can only imagine how many we will need significantly less police if they no longer are required to patrol for speeders. I can not wait.
I know we'll have driverless cars within ten-years
You certainly don't know that - at least on roads generally. You might have a reasonable hope that in some highly constrained environments they might be an option, but that's very different.
On top of all these , this particular announcement is nothing close to a commercial launch. It is like this forever beta phase where only Waymo engineers are getting picked up and nothing different than a thousand other startups that are running autonomous cars in SFO along various spectrum of autonomy.
Andrew Yang has been hammering this issue home for years. There are millions and millions of jobs on the line here, and "just learn to code" is not going to cut it. I suspect the transition won't happen overnight, but nor do I believe it's going to happen gradually over 50 year. It will be a catastrophic disruption to multiple industries, including insurance and finance, the entire oil industry, the entire delivery and transport industries, automotive repair/paint, and everything related to these industries. Which is basically everything.
In typical human fashion, we're not going to do anything until everything blows up.
Here in Spain we just ended this week a 14-day illegal transport strike. Truckers all around the country stopped work completely without any minimal service, that meant no produce, no milk, no gas, no medicine, no nothing. Supply chain completely broken.
They effectively blackmailed the government.
I have ZERO pity or remorse to replace these people with self-driving trucks.
This is (hopefully) overly cynical, but having worked in and closely with politicians in the past, I highly doubt it.
There are surely interest groups and think thanks that are working on the problem, and they will (if they don't already), have some legislation either ready to go or close to it.
The politicians won't even think about it until it becomes an actual problem that threatens votes, and when they do start caring they will just take whatever they can get from friendly/trusted think tanks and try to ram it through. The bills will be thousands of pages long and practically nobody will read them before voting for them.
Good. Law should always lag social consensus. The power of the state should not be thrown around at the behest of unaccountable experts or a razor's edge majority. It should require the support and consent of the bulk of the people.
Oh fuck no. Social consensus is too slow. Look at the social consensus that tracking was bad followed expert consensus by like a decade, and really only started appearing after EU regulation.
The "experts" also thought if Vietnam went communist the whole continent would. The "experts" gave us the war on drugs. The fatherless black household and the lazy black man weren't stereotypes until the "experts" tried to help minorities.
How many lives are you willing to ruin and how much blood are you willing to spill getting things wrong? The power of the state shouldn't be thrown around on a whim or subject to subject specific industry fads and circle jerks.
So if you don't touch the lever on the trolley you aren't responsible for the millions of people it plows through on the default track? Uh-huh. Keep telling yourself that.
Look, I invest in companies that stand to benefit in this transition and I want my stonks to go up -- but I'm horrified by the apathetic excuses people toss out to rationalize ignoring the externalities. Strengthening the safety net is the least we can do. "What if we get it wrong?" sounds an awful lot like "it's gonna cost money and I don't want to pay" to my ears.
History is littered with the stories of people who thought the end justified the means. The 20th century eugenicists thought they were improving humanity. The people who tried to civilize the natives with Christianity thought they were saving them from eternal suffering. The "if the .gov doesn't get involved things will go to shit and you'll be sorry" stuff you are saying is basically what people who were ardent proponents of things like restrictive zoning were saying and we all know how well that worked.
I'm not saying don't do anything. I'm saying it's more than ill-advised to get the vast resources of the modern state involved in things on the whims of "experts" because hubris abounds.
I think you and GP are both right. The Trolly Problem is legit and real. OTOH humans (especially those exercising any sort of power) are grossly incompetent, arrogant, and good at leaving trails of suffering behind them.
I also think you're not really disagreeing. GP is talking about specific implementations while you are talking about high level philosophy.
I don't believe that there's a right answer that will fit all cases here - for issues like same sex marriage, you can find all permutations of social acceptance and lawful acceptance all over earth, yet I think it's hard to justify any position besides lawful and social acceptance.
I think you have experts and politicians confused.
> The "experts" also thought if Vietnam went communist the whole continent would.
The domino theory was advocated by politicians, not experts.
> The "experts" gave us the war on drugs.
Nixon gave us the war on drugs as an excuse to be racist towards black people and let cops beat up hippies.
Experts have derided the war on drugs (as opposed to treatment for addicts) and the classification of weed as schedule 1.
> The fatherless black household [wasn't] stereotypes until the "experts" tried to help minorities.
The stereotype preceded the experts, but yeah, Moynihan fucked up there.
> How many lives are you willing to ruin and how much blood are you willing to spill getting things wrong?
I mean, a not insignificant number. The US has 350 million people. Any decision at that scale is likely to fuck over a lot of people, whether that's allowing any company to put out a self-driving car or changing fuel efficiency levels.
> The power of the state shouldn't be thrown around on a whim or subject to subject specific industry fads and circle jerks.
On the contrary, I'm advocating the power of the state to take action against industry fads faster than the public notices the issues. Such as using governmental power to limit invasive tracking of internet activity.
I will believe self-driving car is near once I see truck drivers automated. If truck-driving cannot be profitable (with the advantage of no breaks, unlike humans, and a semi-fixed route) how could a personal vehicle or taxi be profitable? It seems it could be one of the first places to optimize. Trucks are so expensive, so they could be early adopters.
I still think your prediction is too optimist. IMO (I admit, based on gut feeling) either we won't have it or the goals will have changed and it will only work in special roads (almost like rails for trains). See you in 10 years (if hn still exists!)
what do you mean 10 years? i remember Elon saying to buy a Telsa in like 2018 because in a year it would make you 60,000 a year by being a robo taxi? this has not happened?
I still get a chuckle out of the Elon fans who thought for some reason a huge fleet of robotaxis would somehow make them rich instead of become an instant race to the bottom on prices.
and why did they think a company who is there to make money would give you a device that will print you money... and not just keep it for them self's...
On the other hand, if I were creating a fleet of self-driving taxis, and I could choose between 50mpg cars or 20mpg trucks/SUVs, I'd use 90% the former.
Most of the self driving cars should be fleet, so I doubt any of these specifically will occur. It will be more like having cheap reliable taxis around, and anyone who has lived in a cheaper country (like China) might understand that.
Coupled with electrification going on at the same time (many countries and states have set 2030 or 2035 deadlines), the CO2 impact won't be as bad, although we should still focus on public mass transit.
There are a lot of countervailing trends, some directly the opposite of what is predicted here: Delivery, not pickup, for example. Right-sized shared vehicles with surge capacity instead of "orbiting" private cars. Smaller-lighter vehicles as the need for crashworthy vehicles declines. More use of bicycles as roads become safer for cyclists.
Wouldn't it be the other way, through fleet network effects?
For example, your fleet of low- or zero-emission vehicles would greatly offset the average vehicles that people would have been using instead.
Also, zero occupancy vehicles never have to drive around looking for parking near their destination. They are immediately reusable or could go to an out-of-the-way pool.
> Isn't this backwards? Why should kids have to go earlier than parents if the parents don't need to drop them off before work?
Because in most places school starts before work starts, this has always been the case, even in places where people don't have cars, and not everybody has a car in places where most people do.
> Won't most automated vehicles be electric?
Yes, but that doesn't automatically mean all that power is made with renewables.
>Yes, but that doesn't automatically mean all that power is made with renewables.
Electric vehicles are consistently more carbon efficient than individual combustion engines, except for specific outlier instances in developing countries which are 100% dirty coal power grids.
This means a fleet of electric vehicles will have less of a carbon impact per mile than the average human driven vehicle.
This is true in the case where you compare a fleet of electric vehicles with the same size fleet of combustion based vehicles.
But that's an entirely different subject. We are simply looking at the effect on the total number of miles driven here, which will absolutely explode once driverless vehicles are a reality. And those KWh will have to come from somewhere and they will have a carbon footprint even when they are entirely made with renewables.
The solution to this is to drive substantially less not to drive more with cleaner vehicles.
There is no reason to believe that driverless cars will be driven more than regular vehicles, and even if they were, you'd be able to regulate fleet pricing to stem overusage.
There is still no comparison between electric vehicles and combustion vehicles in terms of carbon footprint.
- Mass transit is full of big heavy vehicles running mostly empty on off-peak hours. It's more efficient only at rush hour. I'd expect net efficiency over the full usage is worse, compared to electric cars.
- As they start making custom robocar hardware, it'll become a mix probably including smaller lighter single-occupancy cars.
- Mass transit ought to become much more useful and attractive given on-demand robocar rides to and from your station.
- (Added:) Congestion pricing becomes more doable for cars that are already networked for other reasons. Congestion pricing is the obvious approach to really solving congestion.
> Mass transit is full of big heavy vehicles running mostly empty on off-peak hours. It's more efficient only at rush hour. I'd expect net efficiency over the full usage is worse, compared to electric cars.
That is really only true in the states right? In other countries with decent mass transit systems, it seems like they are really full during rush hour but only mostly full at other times, utilization is very good. American systems seem to be dysfunctional...probably because cars are more convenient for off peak travel. I can see how door-to-door automated vans or even taxis could fill in off peak travel here, while the big vehicles serve mostly peak demand (in the states, elsewhere it isn't as needed).
> cars doing the minimum to stay 'in traffic' to avoid paying high fees for parking
Parking will get cheaper, if anything:
- Electric AVs can park themselves 5x further away with no UX downside and minimal energy cost.
- Nobody needs to use the doors while parked, and AVs can cooperatively rearrange themselves to allow ingress or egress, meaning they can park with almost no clearance in all directions.
- Parking garages can be designed with the assumption that vehicles are driven with perfect precision, allowing wasted aisle space to be significantly reduced.
I think the cars picking up things/people without drivers could be the worst part. Imagine being driven in a taxi, and about half of the vehicles around you have nobody in them. I suppose it's not inherently wasteful since you never really needed to be in the car to go pick something up. However, if the price of delivering something by car decreases a lot, the volume of deliveries will increase a lot too.
A lot of places have looked at taxation by miles travelled, and that would address a lot of those. For places with high fuel taxes today this is particularly attractive with the transition towards electric cars.
Sending your kid by themselves in a taxi isn't an uncommon thing in China. I can't imagine that changing much with self driving cars, but I understand the American perspective isn't already used to such things.
On the trucking side of things, I think it'll actually be a positive for the industry in the short term (a decade or two), before the jobs really start to dry up.
Rail freight will probably be impacted by automated trucks sooner than road freight, weirdly enough.
My guess about the industry's progression:
* Automated trucks take over long haul routes (e.g. highway driving) between major distribution hubs, and previous long haul drivers move to middle and last mile routes.
* Shippers begin to shift freight away from trains and towards trucks (because price of trucks goes down, speed of delivery goes up)
* Automated truck centric freeways (or lanes) begin to show up, allowing for higher density and speeds of truck traffic. Price further falls, volume increases. More jobs created in short and medium haul trucking to handle increased volume of things being shipped.
* Automated trucks begin to be able to handle medium haul (inner city between warehouses), jobs start declining.
* Automated trucks begin to be able to handle short haul (last mile), truckers are a truly dying breed at this point.
Wouldn't it be faster for the autonomous trucks to go on autonomous trains in-between major hubs then drive off to do the last mile?
Might require a built up of rail infrastructure and I don't know the limits of freight trains in general, but seeing passenger trains, going 2-3x faster than road speed limits I reckon while maybe not by as much, freight trains could still go quite a bit faster than driving the trucks long distances.
Once the driving is automated, what would prevent us from having a "fully" automated package delivery infrastructure? The roadblocks that come to mind (Box & labeling standardization, warehouse automations like pickers/packers/unpackers) all seem like much simpler problems than automated driving.
Conversely, a lot of people in smaller, denser countries suffer from anti-car-brain, and don't appreciate the difficulty implementing comprehensive public infrastructure on an America scale.
And those difficulties are imaginary. If I wanted a high speed public rail line throughout Florida, a dense, narrow, flat state that's the size of a small country. Where is the scale challenge?
Europe and America are about the same size. Why aren't you concerned about European scale? Why does it matter how your brain pictures administrative boundaries?
I think it's the 'relatively near-instant' that's not happening; it's slowly happening in a few cities in limited routes; I'm sure it'll get better/wider but as long as it's an incremental thing then a lot of these problems aren't suddenly going to happen.
A person's professional life lasts for around 30 years (probably much more for the people born on some interval from the 1980's to 2010's, depending on your country).
30 years seems kind of short. I get that in some places (France comes to mind) it is common to retire pretty early. With the associated destruction of cognitive skills, unfortunately. But in the US I'd wager most people have a professional life more along the lines of 45-50 years.
If industry leads regulation on this, I predict a Challenger-like disaster that sets the self-driving industry back by decades. A fatal school bus collision or something like that.
I hope the companies making these things are calculating margins for safety, tripling them, adding triple failsafes for everything, and not compromising on any safety standard, ever.
But we all know they aren't. There's far too much money at stake.
If there ever was a case for the government to proactively apply standards, this is it. Yes, there would be a delay, but it would be worth it, measured against the psychological and legal damage that will occur when (not if) the unthinkable occurs. Given probability p, and sufficiently large N and t, it is inevitable.
There's already been an industry-shaking accident that led to vast reforms in regulation and program design: the Elaine Herzberg crash.
I can't speak to every company in the space (and I can think of a couple that don't seem to have much in the way of safety), but all of organizations I've worked for are very diligent about safety and looking to improve. N > 1 here, so they exist.
Neither did the challenger disaster. Both are continuously (and somewhat tediously) referenced in the introduction to every safety-related discussion, alongside therac-25 for software talks and the martian orbiter for units library talks.
I don't think a single fatal school bus collision will do much. After all, there are fatal collisions with human drivers all the time and yet there is very little public support for banning human drivers.
We can't live without human drivers, though. We can absolutely live without automated cars. And when it becomes clear that corporate greed drove some cost-cutting choice which made the automated car less safe by design, it will be the kiss of death for the technology.
* Inner city car transport will become more popular, since parking won't be a huge limit
* Self-driving rideshare cars can outcompete there consumer counterparts on price - there's no need for a powerful engine, and a large battery since it will be used for small distances at low speeds.
* Car ownership will plummet, which will free a ton of expensive parking lot real-estate to be developed
* Ridesharing will reduce the amount of cars by 2x-4x, eliminating traffic jams
* Owning a house with an 1 hour commute outside city limits will be much more feasible, as you can just sit in the car, and commute into the city, generally the value of inner city lots will go down.
What will happen:
- Truckers will go out of work in very large quantities.
- They will go into other trades.
- Many won't. Crime may go up if we don't figure out how to create more jobs for the type of people who chose trucking.
It's not really clear what you mean by this - we have driverless cars now, but I assume you're talking about them being ubiquitous.
My assumption is that we haven't really gotten any closer to generalizing autonomous car behaviour outside of very "clean" situations (like exhaustively tested routes in a city with known intersections) in the past 10 years, and I'm not sure we're going to make much progress on that in the next 10 years either, without a significant shift in the ways we're solving these problems.
To echo another commenter - people were saying this 10 years ago too.
* reduced income for cities from speeding and parking tickets, and the reduction in people who give the tickets, all the processing that goes along with that.
* availability of large lots of urban land due to lower parking needs (though there are many good uses for this like urban farming)
* significant decrease in the need for auto repair shops. Partly because electric cars have less moving parts and are more reliable than ICE, but also because autonomous cars (possibly/likely) have fewer accidents. This may not happen all at once, but as they improve..
I used to have a much larger list, but I forget what many of those are.
has any government yet started planning for the relatively near-instant transformation of the Trucker and Taxi industry
I can't say for taxis, but according to the newspapers, Texas has been running automated tractor trailers between Dallas and Houston (about 230 miles) for at least a year, and at least one other state has been doing it for at least two years. (Possibly Nevada, between Vegas and either Reno or Phoenix. But my memory is fuzzy on that, and I have zero confidence in Google.)
Texas has also been thinking about building special lanes or highways just for automated tractor trailers.
> special lanes or highways just for automated tractor trailers.
I know that there's last-mile advantages to this compared to what already exists, but god damn I just cannot get over the fact that we've just reinvented trains with a much larger environmental impact.
the way to think of these is as instantly reconfigurable trains. in an ideal market, an enterprising entreprenuer would be able to install tracks on the most used routes and lower costs (and emissions). but we've long succumbed to corruption and financialization in markets (note that if risk is priced properly, capital wouldn't need to evaluate opportunities deeply, but would rather take every NPV-positive project available, in order of expected returns, in a large and diverse enough portfolio to mitigate systemic risk).
That is in my opinion the way to go with this, to either adapt the infrastructure to make it work or to at least separate the self driving traffic and the 'normal' traffic so the two won't interfere.
The existing government doesn't plan for that. Candidates for office and thinktanks do, and then when the threat materializes, they run for power, take over, and make a change. On the one wing, there are gonna be AOC types who say Green News Jobs is the play. In the center, you'll have the Yang Gang types who call for a UBI program that isn't paid for with carbon taxes. On the other wing, you'll have folks who say it's time to annex Greenland and mobilize paramilitaries. So we'll see
My personal wish is for the cars to go hide themselves inside underground parking lots when dismissed, and for us to recover hot asphalt parking lots into walkable towns and green areas.
Hopefully governments do what they do best; get out of the way.
There will still be plenty of jobs for drivers for the next 10-20 years. We'll see a mix of people retiring early, retraining, and switching into support and operational roles.
Out legal system is extremely flexible and will easily accommodate self driving cars. It already handles more complex ownership issues related to product liability.
There won't be a problem with fake taxis just like there isn't problem with fake mail delivery (delivering bombs for the "lols").
Well I didn't say they should always do nothing. They are just better at getting out of the way than solving problems in most cases. Sometimes the issue is that nobody else has the incentives or means to solve a problem (climate change, food safety)
Reorganizing the economy around trucker job displacement is not something the government needs to be involved in.
Providing worker retraining & job placement assistance isn't the government getting out of the way. They are humans and they will need assistance so we should provide it.
I don’t think we will have FSD in 10 years on most roads and in lots of countries, and I also don’t think cars with FSD will be allowed to drive everywhere. I think it will take years of trust after a real FSD comes to market. And only then will infra be changed for it.
My dream is to call a car to a place near my home, get in, and automatically drive me to one or more locations, before returning me home. I’m 45, and I doubt this will happen in my lifetime.
Lowering the cost of driving doesn't mean there will be less jobs in the transportation space, if anything it means there will be more jobs.
Truck drivers jobs are the monitor and secure the shipment just as much as anything else. Taxi drivers have to keep their cars clean.
And a lower cost of shipping will just mean that more stuff gets shipped in what would otherwise be a less efficient way, no different than how faster computers often just lead to less efficient software design.
There is a labor shortage. Shifting labor from essentially low productivity things like overseeing gas stations to other industries is better for the economy.
There is a labor shortage. Shifting labor from essentially low productivity things like overseeing gas stations to other industries is better for the economy.
Only if you believe that every person is equally able and qualified to perform every other job that exists in an economy.
People aren't cogs, and they're not universally interchangeable, no matter what the economic textbooks would like people on the internet to believe.
Exactly. They are extremely adaptive, capable of learning and doing many things. Sure, not everyone can do everything, but there are also no purpose-built humans capable of only one job.
> People aren't cogs, and they're not universally interchangeable
People aren't, but humanity is...meaning those drivers might not adapt to new work, but at least their children will. The Luddites had huge problems with obsolescence, but their kids got by OK. A social safety net could take the edge of the transition rather than just delaying the transition indefinitely.
Sure, but the labor demand is also large and spans a wide gamut of skills. Someone who used to be a gas station cashier can certainly find work elsewhere.
“Anybody who can go down 3,000 feet in a mine can sure as hell learn to program as well... Anybody who can throw coal into a furnace can learn how to program, for God’s sake!”
Autonomous driving is going to be an absolutely huge driver (lol) of consolidating wealth into fewer states, imo.
It will be in a states best interest to create a tax structure around autonomous vehicles from other states performing services within its borders. Possibly a defensively high tax rate in order to promote in-state ownership of in-state autonomous vehicles.
The unemployment part really concerns me. We are going to be automating so many people out of jobs and there is going to be no work or only minimum wage jobs out there for them. Whatever field the majority of them train to will be slammed with pay cuts as a glut of new workers gets added to the pool.
We really need to start implementing universal basic income
The cost of UBI isn't low: so let me ask - what's the cost of retraining those displaced by automation to another job? Is it more, or less than the cost of UBI?
Why is it assumed, a priori, that those displaced people cannot be retrained for some other useful job?
Because we will be destroying way more jobs than will be making and more than will be available. It will be highly disruptive to other labor markets and will cause a race to the bottom for workers' salaries.
In cities with good weather I can see this (it basically sunny clear weather here in phoenix all the time)
However, I could not see this working well in places like Chicago. I have a hard time believing a self driving car could safely navigate a road mostly covered in snow.
Either way it will be interesting to see what happens in the next 10 years.
> Fuel-stops are going to change completely, what's the point of half of the motels on long-haul drives when your car can drive all night and likely recharge automatically at a stop-point.
Impact will be minimal. Truckers don't pay much, and over the years, truckers business has gone down significantly.
Can you clarify your comment? There is a shortage of truckers in the US (and many other countries from what I understand). What do you mean by "Truckers don't pay much"? Truckers don't pay much to whom?
At the same time, motels aren't likely the big losers here as I believe most long-haul truckers sleep in the beds in the truck
The changes you list here would take longer than 10 years. But yes the fallout will be intense.
It is also hard to imagine anyone in government is incentivized to go after such issues. The huge lobby machinery that will influence the planning when the time comes.
This looks exactly like a comment I saw around 2008. In all seriousness, I don’t think it makes sense to waste resources on planning for autonomous vehicles as it is not guaranteed (probably not even likely) to have them in the near future
Right now there is a vast undersupply of truck drivers and it's only getting much worse within the next few years. There's so many truck drivers just about to retire. One presentation called it a truck driving apocalypse.
I think you overestimate government planning and the amount of infrastructure that will actually be built. For example, assuming autonomous vehicles will get signal preemption, when not even city buses get even signal priority today.
> I know we'll have driverless cars within ten-years
If we get driverless cars (meaning there literally isn't a steering wheel), it may happen within 10-years BUT only for last mile stuff (or vice versa for trucking). It won't be an overnight event. We won't see truly driverless vehicles across the board for decades, if at all.
Autonomous cars that can "tattle" to the authorities on vehicles that break speed limits or violate traffic laws. Maybe the owner of the autonomous vehicle would even get a cut of any traffic fines.
I also keep thinking about how much of an impact driverless taxis would have on living in the city. If they are available 24/7 and cheap enough you could more easily live outside the city.
Also high-jackings. Not sure if automation would encourage it or discourage it. It’s not a big problem presently but who knows… trains in LA were getting looted in their yards.
Are there examples of a government seeing a large shift coming and preparing well for it? I guess there’s be fewer because they might not mention it after the fact, but…
Yet another reason why Universal Basic Income is going to become a necessity. As each industry automates itself more and more, it'll be UBI or total riots and chaos.
And thank God. People are not robots. They shouldn't have to spend their valuable lives doing repetitive, monotonous tasks for 8+ hours a day to barely survive anyway. So long cashiers, so long truck drivers, and hello future.
I disagree with this. There’s a labor shortage in just about every industry right now, we aren’t in a high unemployment world by any stretch. UBI is not needed when wages are high and unemployment is low.
There are huge lifts to get to mass adoption. Most of the current tests aren't even general autonomous driving in the sense that they are being conducted in specific areas only. Autonomous trucks are being tested on routes between specific destinations, cars in specific cities/neighborhoods.
For example, it's going to take years (I'd put my money on 15-20 years at least) to get general availability in NYC. The roads are terrible, inconsistent, and often lacking in clear markings; the density of cyclists and pedestrians is high; and the regulators hate any kind of change, let alone one that challenges employment.
Then there's all the infrastructure that needs building - charging stations, road improvements (I assume, for example, many highways will get some kind of infrastructure dedicated to making autonomous easier), maybe some last mile handoff to human drivers in certain jurisdictions, upskilled maintenance centers, warehouse facilities that can handle the presumably increased loads, communications infrastructure, etc, etc.
TLDR the build out is going to be a much bigger lift than you might think. Just look at how slow the rollout of 5G is going.
No, I don't think anyone with the possible means is preparing meaningfully.
Yang did a good job of grokking the implications, but he wasn't taken very seriously and his messaging wasn't great I think. But the crux of it was solid - "XX thousand mid 30's truckers w/ iffy education all out of work, and that work was one of the last solid, fairly high paying long term jobs that demographic could land so how happy will that demographic be?"
Buttigieg is a smart guy and he might be thinking about it per his DoT seat, but I haven't followed his takes on it if they exist. Watching some of the cybersec-focused hearings and comments from Ben Sasse also seem to indicate technical awareness.
Overall, I do think some of the more tuned-in politicians have read some Jaron Lanier though based on the policy ideas that nod in this direction I've heard about and how they mirror his writing. He doesn't paint a great mid-term picture (tldr a lot of unemployment) but long term he offers some interesting ideas on how there might be a positive landing - data-aggregators collapse under the weight of their own data management costs vs. ability to really parse it well for profit ("siren servers"), automation/internet taxes which open up a UBI based on the robot labor, that sort of thing.
This sounds really cynical and ageist but I do think we're in a standstill until Buttigieg and his generation are the senior leaders. Even breaching the surface of the pretty good certainty of what you describe (to include the sci fi -> real life mapping trend you mention, which I agree with) when watching these tech hearings makes me think we have some time to wait.
The same arguments are made about EVs and not enough power on the grid, but those are similarly false for the same reason, the change will be gradual. Also people love to hate on capitalism but it does a very good job of adjusting for this type of thing, especially when changes are slow.
My feeling is that the trucker protests in Canada and D.C. were ostensibly about masks, vaccines and personal freedom, but a lot of people pointed out that didn't make sense since mandates were lifted, so they had anxiety about something else.
I think the actual anxiety is about being put out of work by self-driving trucks. Truckers can see the future of trucking and don't see themselves in it. Nobody seems to be working on a plan for the millions of workers whose skills will no longer be needed.
How Many Truckers Are There? Approximately 3.5 million truck drivers are employed in the United States, out of which 1 in 9 are independent, and most are owner-operators.
I was biking in SF the other day, pulling my kid in a bike trailer. A waymo vehicle pulled up behind me while I was waiting for the traffic light. I was pleasantly surprised that it waited for me when the light turned green and didn't run me over. Robot anxiety is real though.
Edit: anxiety also because I wasn't sure it would recognize the bike trailer as a real thing. How many bike trailers end up in the training set?
Waymo has taken a responsible approach to self-driving (level 5 only, Lidar) over Tesla's collateral reduction approach. Lidar gives you exact distance to obstacles around it. Given those 2 factors, I expect Waymo to have an inbuilt hard-stop for when it is on collision course inside of braking distance.
I expect it to be friendlier to bikers and pedestrians than human drivers.
As a biker, most of close calls are due to impatient drivers who try to overtake me through close calls and when they enter turns without looking first. AFAIK, Self-driving cars do not initiate same-lane overtakes and can be hardcoded to be paragons of patience.
With how hostile American drivers are towards bikers, self-driving cars would need Skynet-esque malice towards us to match those injury rates.
Thats a refreshingly honest statement, especially contrasted with claims by Musk that Tesla Lvl 5 will be easy to do and available to purchase by 2020.
Does exhaust smoke consistently reflect lasers ? In my experience, a dense lidar will see a subset of the lasers go through such occlusions and get a 2nd return from the truly opaque object behind the smoke.
LIDAR can usually distinguish translucent smoke from solid objects like bikes/pedestrians even before the ML algorithm kicks in.
Paper bags are tricky, because a bag on the ground could just as likely be a solid object that the car should be stopping for. Stopping or slowing down for bags may be a good idea.
I would like some kind of visual feedback (maybe countdown timer for how long it has stopped for?) from autonomous car.
Something similar to checking if driver is paying attention and is aware that I'm going to cross in front of their car by looking at direction where driver is looking or a hand gesture.
How do you achieve this? I feel like every time I'm crossing the street windshield glare is so severe that I couldn't even be convinced the driver was in their seat.
This is one of the reasons I wish drivers would stop trying to let me cross the street when they could just drive past. Most of the time I can't see them clearly enough to make eye contact and be sure that they are not just lost/confused (I live in a tourist town with diabolical roads so that's pretty common).
I remember Ford doing some of this work in a research phase a few years ago (by putting real people in hilarious empty seat costumes). I don't recall anything more recent from the major players and it still feels like a missing component to to having vulnerable road users feel comfortable around autonomous cars.
https://www.youtube.com/watch?v=KqQyYOPPn7w
Good idea. Maybe the car can turn on an indicator when it has made the decision it is going to move, something that even gives an idea of where it is intending to move.
Tesla FSD beta has something like three orders of magnitude more miles driven under full autonomy now, almost all of it by non-corporate users free to post whatever mistakes or mishaps they can find.
Waymo has one (now two) heavily geofenced service areas with a few dozen cars. It won't take highways. It won't enter certain intersections or transit certain neighborhoods.
Just recognize that there's some level of cherry-picking error in this kind of analysis. FSD beta and Waymo both make mistakes (mine curbed a tire once, there's the now-famous video of the car hitting the plastic bollard in San Jose, etc...), but you're going to see far more of those from one system than the other.
> Waymo has one (now two) heavily geofenced service areas with a few dozen cars. It won't take highways. It won't enter certain intersections or transit certain neighborhoods.
Literally this is why I trust it more, the people running the company are willing to realistically evaluate the limitations of their systems and act responsibly
They sometimes post videos of situations they deal with, especially around bikers and pedestrians. It's quite impressive how well it handles such cases.
It's notable that most of the failures in that video are human process and communication failures, not failures of the self driving system itself.
It's the kind of thing that could all be fixed in a few days of rehearsals of what to do when cars get into this state. I suspect that the main failure was that rider support had a button to 'pull over', but that the route to pull over was blocked by a cone, so the car failed to pull over, and then rider support seeing the failure probably tried to cancel the pullover (to try again), and it ended up keeping driving.
A simple logic change or even just better training/documentation could have fixed this.
But at 16:46 it continues on and moves into traffic before once again failing, this time with people having to go into the oncoming turn lane to go around the Waymo vehicle.
The video contained almost no information. Just a clip of a cyclist cutting in front of a Waymo car and then a cut back to the Waymo employee. It doesn't even show how the car responded.
That still doesn't say anything other than "our data includes interactions with cyclists." As far as I can tell, they don't treat cyclists any different than they do other obstacles on the road.
They do, the model specifically marks cyclist, pedestrian and other obstacles as different, and they discuss how they're modeled differently. Furthermore, they have videos talking about their simulation system, and how they create thousands of "variations" of each scenario, with the "obstacle" doing different things.
Scania did a neat affordance for this in their (otherwise rather incremental) autonomous AXL concept. A band of LEDs around the vehicle that light up "towards" a pedestrian once the vehicle takes them into account: https://youtu.be/0WN9xvAvEls?t=499
Waymo and cruise vehicles are much safer for bikes than other drivers in my experience. I've been commuting by bike in SF for years, have a near miss every few days but never even close with a waymo vehicle. Every single time I've been hit has been a door, another bike, or a taxi. Taxi drivers in SF seem to see themselves as natural predators for bicycles.
I wonder what the stats are on those bike trailers. I see people using them in traffic and they look plain terrifying: a child dangling out in a flimsy canvas pod behind the cyclists's back, too low for drivers to see when close, surrounded by vehicle wheels... yikes.
There shouldn't be anything inherently dangerous about them, a car shouldn't be getting that close to the back of a bicycle really. I've seen some that have a tall flag-like thing, I guess that's to help show that there's something behind the bicycle
Shouldn't doesn't mean doesn't. Seems to me that common sense suggests they're pretty dangerous to use in traffic, but I'm not claiming that is actually the case. I'm just wondering what the stats are - the only real way to know.
For context, I live in a major city. I've seen people pulling these things round busy multi-lane roundabouts and the like.
Many people seem really careless with their children on the road. Another example is walking out into a crosswalk with a stroller without looking both ways. Yes, it's my responsibility to slow down. But for god's sake don't assume I'm going to!
I was thinking about this other day with my 4 month old. We were looking at four or so houses to purchase with a real estate agent. We were having to load her in and out of her car seat in the car repeatedly which is not fun (she cries when you set her in the thing). One of the houses was 1/2 a mile or so away and it crossed my mind that I could just hold her in the back seat while we coast through a few intersections. I then pictured the outcome of a minor accident and decided that it was worth the additional effort to ensure that my kid wouldn't be ejected.
Everything we do is a risk calculation and some people have decided that the risk is worth the convenience or cost.
I suppose my thought was illegal while toting your child around in a bike trailer isn't so there is that factor too...
Those kid trailers seem very dangerous, I ride my bike in cities all the time (NYC and SF) and get very close to reckless drivers. I have to imagine being a passenger in a bike trailer, in a city, is quite a risk
The most common accident from biking is falling. Much of the danger of falls are removed for a child in a proper bike trailer. With a flag and reflectors helping visibility my intuition tells me that it is much safer for children in trailers. Safer than riding their own bike or with a child seat on the adults bike.
It must feel very gratifying to be able to say, "hey, you shouldn't have done that" after a driver crushes your child into a paste. Never rely on what drivers "should" or "shouldn't" do.
As a motorcyclist, I'm well versed in not relying on drivers. SMIDSY all day. My point is that it should be safe; if it isn't, something is wrong with the system.
Most people I know who regularly bike in SF (for commuting at least), have been hit by a car in some capacity.
I have no idea why parents use those bike with the kids in the back. Why not just use a car? It's not even like those bikes are particularly cheap. The ones I see sell for thousands of dollars.
In Germany many bike trailers have a small brightly colored flag on top, so there's an object in your field of view even if it's very close to your car.
There's a subset of cyclists that don't seem to have fear. Not me though. When I take by kiddo of out I'm very careful with route planning. They're harder to see because they're lower down, but the kid literally can't fall out of it if I take a fall. That, and it effectively has a roll cage.
The stats are that probably hardly anything happens with them. Dying on a bike is extraordinarily rare in the first place. For example in my city of some 4 million, 30 people on bikes die a year which seems like such small odds.
I decided to go ahead and do some ballpark math. about 2% of my city bikes to work:
36/80,000 = 0.00045
80% drive a car and about 250 die a year:
250/3200000 = 0.000078125
So its not quite a full order of magnitude more dangerous, but either way I'd say your odds are pretty low in either case. Even lower if you choose to bike safely, such as taking less busy residential streets vs commercial arteries or even opting for the sidewalk on those busy arteries, wearing helmets and using lights, using defensive bike strategies like avoiding lefts in favor of utilizing a perpendicular street, not being drunk, etc. I bet out of those 36 deaths in my city, none took place on like a 25mph residential street.
I bike around the city often, mostly at night. In the last week I almost got hit by one turning in front of me, and I watched another blow through a red on 9th and Judah. I've also observed them slowing drivers down with erratic behavior. Now I understand why the sudden uptick in volume, but am predicting it gets locked back down soon bc from what I've seen they have much more to learn. I'm very cautious around them. Good for the company and team behind it tho, it's a good fight to wage. Fingers crossed.
Surely you don't solve this problem by recognizing specific objects, you just identify "is there /anything/ in front of me except for road or vehicle travelling at a similar speed to me/within a certain distance of me" and stop if so.
It's been about a year now, but I used to skate around the city all the time -- even then I found the self-driving test cars to be way less treacherous than the human-driven ones!
Pulling your kid in a bike trailer in SF is awesome and admirable, but also a little bit crazy :-)
Yeah it's funny how human psychology works! We know the rational truth is that the AI drivers are millions of times less likely to run us over than human drivers, but we still can't get rid of the image of evil robots going bezerk from Hollywood movies and the (extremely unlikely) possibility of a malfunction in the back of our minds!
Uh... people have already been killed by autonomous cars. And they're very new. I don't think it's funny/weird at all that someone would be nervous about being in front of one, without an airbag.
The fact that someone was killed by a car with an Uber computer on board does not mean that people have been killed by autonomous cars. It's easy to draw a bright line between realistic, responsible self-driving tech companies like Cruise and Waymo and the move-fast-kill-people shenanigans of Uber and Tesla. They are really two different categories.
Maybe for someone who is paying close attention, or even works in the industry, this distinction can be made.
And that's valid for now. That a company has not had a major accident so far does not mean that I'm not going to be their first victim.
I love the idea of self driving cars.
But they are also really scary.
They're being brought to us by the same industry that gave us
* BSOD
* "why the heck can't I share my screen right now? can you share your screen? yes? well why can't I share my screen?"
* Your account has been terminated for breaking the terms and conditions. You cannot appeal this decision. We are now deleting 4TB of your life's work.
* An attacker has gained access to your email account. Your life is now ruined.
`But they're safer than human drivers!`
So what. They're new, they're capable of killing you, they're brought to you by tech.
They're f'n scary.
Self-driving advocates will do well to acknowledge this rather than being dismissive.
It's easy to dismiss you because your arguments are nonsense.
Uber's murder car was not autonomous. It was not capable of ever operating without a "safety driver". Therefore it was not an autonomous vehicle. Nor is any Tesla.
Waymo has operated cars with zero humans aboard at various points going back quite a few years. They are legitimate autonomous vehicles.
I think the problem is that (at least for now), we don't have good mental models of AI drivers and what their failure modes might be.
Dangerous human drivers are likely to be consistently erratic (e.g., drunk), or distracted and oblivious in certain situations (bikes, intersections, lane changes).
For all we know, an AI driver might drive perfectly for three hours, and then accelerate full speed onto the sidewalk.
I'm not concerned about evil robots. A coordinated cyber attack? People intentionally exploiting bugs in perception? Errors in general?
I think the risk of a malfunction is much higher with a robot than a sober human, and probably most legally intoxicated humans.
I don't base it on Hollywood, but in years of programming experience. Anyway, did you hear about the 737 that crashed due to a software bug last month?
Have to say, seeing my Auto Pilot graphics with cars "jumping" all over and appearing and disappearing.. (this is not the FSD Beta graphics mind you) compared with what seems like a stable and coherent visualization of the environment, it really makes the difference in my feeling of confidence that it's aware and able to react to the same things I see around me
That's part of it. Another part of it is that Tesla's computer vision is still a work in progress which introduces unnecessary jitter into the visualization. The good news is that they seem to be making improvements and the jitter is slowly going away.
yeah I am still confused because in the first part there's jumpy graphics and in the second part there are stable graphics, and I'm not sure if this represents a change over time, or different vehicles, or what
For some reason Tesla still uses old visualization software for non-FSD beta vehicles. This both means objects the system sees aren't on-screen and the interpolation algorithm for making the 3d models look smooth-moving is pretty bad.
I wonder if its possible that the FSD beta graphics use more power to render. Would be dumb to lower range on all Teslas for what is essentially eye candy. Even on FSD vehicles the visualization reverts when FSD is not engaged.
That looks like it's partly because it shows the raw LIDAR data (with some camera data mixed in I think). That's intrinsically not "jumpy" but it's also not what the car uses for driving - you have to look at the shapes on the ground.
It's kind of hard to tell, but they do seem a lot more reliable than Tesla's. Of course that's not surprising seeing as they use LIDAR.
IIRC correctly in their Waymo videos you can see inside the cabin. Each passenger gets a video feed of that video, so I think it is real-time. Let me find a link.
Glad to see they never exceed 20 mph in this video. If AVs can stick to city speeds that prioritize pedestrian safety over travel time, they'll have a huge head start over human drivers in safety and "human decency" metrics. Hopefully have a traffic calming effect on the rest of us, too.
As far as I know, they still don't/won't drive in the rain, and they operate safely and slowly. I want to believe this thing is moving forward, but I'm skeptical about building a marketplace where the service is unavailable exactly when demand is highest. It creates conflicts of interest, that at best, encourage actors to take risks that can undermine the entire industry.
Same. My shower thought has been, has this affected the implementation timelines of safety features in traditional cars? The general public seems to think we’re going to have self driving cars any day now, but anyone reasonably well informed can probably think of a dozen edge cases they haven’t even begun solving yet.
So I wonder, have regulators stalled or stop pushing for new safety features under the false belief that we’re going to have self driving cars any day now, so what’s the hurry? I feel like it has, although admittedly it’s just a feeling and I’ve done no research.
Collision avoidance systems are our generations seat belt. It’s still pretty shocking to me this tech isn’t mandated as a safety feature yet considering the very real potential to save lives from distracted driving.
Has the pr shit storm all these companies shamelessly capitalize on affected this technology implementation?
We generally think of these sorts of lies as harmless and a result of overly ambitious visionaries. But I think in a real way this is the type of thing that could cost lives, in the sense that if regulators understood they are 20-30 years away from mass market, they might actually push for more safety reforms which could save lives.
I'm still waiting for a video of a self driving car during snow five miles outside of Fargo, Minneapolis, or Madison when the road hasn't been plowed in the past hour.
They probably will never do that and there’s nothing wrong with it.
Consider that airlines are a well established and reliable business worth trillions around the globe. And yet they also don’t take off or land in adverse weather conditions. Sometimes there’s a snowstorm and a whole day of flights is canceled.
But none of that invalidates the existence of the industry as a whole.
A driverless taxi service in any city where it snows will never exist.
It will not be possible to replace my personal car with a driverless car.
Relatives who are in the southern states (with driverless cars) will not be able to visit for many winter holidays.
---
Self driving cars are very interesting - but can you get to the point where it is possible to replace them in Chicago or Minneapolis with the same availability for personal or services as a human driven one?
With airlines, yes, shutting canceling the flights out of O'Hare is certainly inconvenient. Is it reasonable to also make it so that people can't get to work via local commuting because of snow?
To get to the point where self driving is default, it needs to be able to handle the conditions that a human driver currently does with the same availability. This goes for taxis, trucks, and the daily commute and errands.
I think you are setting the bar in the wrong place.
For self driving cars to succeed, there just has to be a few good use cases where people prefer them to normal cars, and there's sufficient economic value to run a business. They do not have to murder the existing car industry in order to 'succeed'.
Look at electric vehicles - they've been around for more than a decade now. Many ppl have one and it adds value to their life. Still you can't do _everything_ a gas car can do like drive 1000 miles in a single day. But that's not important because there are places they do shine.
And even though there's still way more ICE cars than EVs on the road, you probably wouldn't argue that EVs "aren't a thing".
Also, you're assuming that the world/environment around the innovation is static, when in most cases the world adapts around the innovation. For EVs, we now see charging stations at grocery stores and office buildings to assist people which the battery capacity problem
Just 2 days ago I saw a waymo car stuck at 4th and Townsend in SF facing NE (toward the bay). The car was in the left of 2 lanes. The light for the left lane turned green. The car sat there. People behind started honking. Eventually it went. Not if there was a driver or not.
In the driver's defense (computer or human), SF's signals have gotten very complicated in the last ~10 yrs and even non-robot drivers fail to follow the rules about 20% of the time. 3 places I can see this happen every day are (1) 4th and Townsend on Townsend facing NE (the one above). The issue is the left lane has a green light but the right lane has a separate right turn signal because there is a bike lane further to the right that gets green first. Sit at the corner and I guarantee within 10 right turning cars someone will turn on red.
Similarly, 5th and Bryant on 5th going NW (into the city). This has a similar deal. There's an on ramp to the freeway but it has a separate right turn signal from the green forward signal. This one, for me, is around 100% violation by which I mean I've never NOT seen a violation at the corner. Not ever car, but ever signal at least one car will turn right on to the freeway even though there are 2 large no right turn lights and a green bike light.
The last is 4th and King, on 4th going SE facing the bay. I'm not sure these are technically violations. The left lane is painted as left turn only. The right as right turn only. The middle lane crosses King. Again, this is close to 100% for at least one car per signal ignoring those markings. I don't blame them as there's no way to see the markings until you're just a few car lengths from the intersection and if there are other cars there you can't see them at all. Further, looking up the law, solid white lines are just guidelines. It's legal to cross them. (double white are not). Still, people in the center lane get angry and honk when people in the left or right lanes cut them off since they weren't expecting it.
"Waymo and Cruise self-driving cars took over San Francisco streets at record levels in 2021 — so did collisions with other cars, scooters, and bikes" [1]
"...Many of the accidents, which the companies are required to report to the California Department of Motor Vehicles, occurred while the vehicles were operating in manual mode, with a safety driver in control.
...But according to an analysis by Insider, a majority of the 98 reported accidents in 2021 occurred while the vehicles were in autonomous mode, or within seconds after the autonomous mode technology had been switched off..."
I'm not sure what you are getting at? As that article notes, in 2021 accidents only doubled while the autonomously driven miles more than tripled?
Waymo has been scaling their level of testing with their level of safety. What has changed is that Waymo think they are safe enough to start doing more widespread testing in SF without a safety driver.
This seems to be a strong counter to all those who were claiming that level 4 self-driving cars would be limited to flat, dry climates for the forseable future.
"...Waymo sued the California DMV this week to prevent it from revealing certain information about its cars' performance, including Waymo's internal analyses of collisions. Waymo argued that the information, including any "potential technological remediations," represents a trade secret that would have "a chilling effect across the industry" if released, and would dissuade potential market participants from investing time and resources into developing self-driving cars..."
Now the question is, did the government officials who approved this, been given access to this data? And if not, why not?
What government officials are you referring to? These are disclosures made to the DMV, so obviously they have access. Maybe you mean the court? But they have access too as part of the lawsuit... so I'm not sure what you're talking about.
So do you mean a private company just decided to move to driverless rides with no required approval from any supervising government official? Is this like Boeing self-certification or even worst because they need no approvals at all?
I don't know if there's anywhere that goes through and categorizes who's at fault, but you can look at the collision reports from the CA DMV. E.g. the report from March 9th:
> The Waymo AV was stopped in traffic, eastbound on Irving Street in autonomous mode. A parked car to the right of the Waymo AV began reversing into the street and made contact with the rear right side of the Waymo AV.
Can't read the article because of the paywall, but just because an accident "occurred while the vehicles were in autonomous mode, or within seconds after the autonomous mode technology had been switched off" doesn't mean the collision was the car's fault. The real question is how many of those accidents would have been prevented if a human was driving.
I think waymo has advanced the tech pretty far but I am still worried about the legal and business aspects of self driving cars:
1. liability cross section scales with size of fleet if waymo is solely responsible for developing their algorithms
2. added operational cost of running a self driving car - is it worth it to average consumer? This is something everyone throws numbers out for but here is what i'm thinking:
- we have lots of examples of transportation systems that have some level of automation (eg. train, subway, planes) but even after decades of robust operations, nearly all of them still require operators in the loop. This is especially striking since rail systems/subways are as closed-loop and simple as you can get in terms of automation. As an outsider, it's unclear to me why they are needed but they are there. Is it because as fleets age, their parts wear down in unpredictable ways? Is it some regulatory requirement?
- for self driving cars, will the same expectation be there eventually? either a remote assistance driver or a chauffer safety driver? If that is the case, then it will be difficult for a self driving car to be cheaper esp with the added cost of all self driving tech stack and operational costs.
> added operational cost of running a self driving car - is it worth it to average consumer?
Early high-cost self-driving cars largely won't be owned by individual consumers where the cars spend the majority of their time parked. They will be owned by companies that sell access to them as taxis by the minute/mile and can justify the cost by minimizing downtime.
The clip at 0:42s impressed me https://youtu.be/O8TSA-X9UlU?t=42 because the behavior seems to be trying to "edge in" and assert fairness, which is what humans expect of each other in these situations.
There's a huge danger in trying to anthropomorphize these systems too much, unless you believe there really is a "fairness" model built in. If not, then we shouldn't try to understand what it's doing in these terms.
A problem I spotted is, shortly after that turn, it sees the person ahead trying to parallel park... and then it give them several car lengths of distance while it waits. It's waiting so far back that the vehicle behind it couldn't fully clear the intersection.
That's not particularly "fair" of it to do that.. nor is giving such a large space to the parallel parking particularly useful or expected. I do not have high expectations for a road filled with this level of barely autonomous agents operating smoothly en masse.
Google has been implementing "fairness" models for a probably going on a decade in their self driving cars. the first example I can remember is them talking about intentionally nosing into a 4 way stop to signal to other drivers that the vehicle understood that it was it's turn to go. Not to say its perfect but the car is "thinking" along those terms.
I’m curious on the HN community opinion on whether Cruise is ahead of Waymo in SF, or if Waymo is just more careful and has spent less time training models specific to SF, but will expand to a fully public program sooner
1. Cruise only operates between 10 PM and 6 AM and under 30 mph. Waymo operates all day without an artificial speed restriction.
2. Waymo's service area in SF is more than twice that of Cruise's. Unsurprisingly, both of them avoid downtown areas.
3. Waymo has hundreds of vehicles in San Francisco. Cruise only has a fraction of that — last I heard, only <10 of them were providing driverless rides.
4. According to CA DMV 2021 disengagement data, Waymo is driving almost 3x more miles than Cruise.
Waymo seems to love a slow and steady approach, but when they announce a service it's fully featured as we've seen in Chandler, AZ.
I don't think it's fair to say that either Cruise or Waymo avoid downtown areas. Cruise cars come pouring out of their depot at Van Ness and California nightly at 10pm, like bats or something. That's not downtown but it's in a very dense, very busy area. Cruise's offices are in SoMa where Dropbox used to be, and I see tons of Cruise cars in the vicinity of Folsom/Embarcadero. I also see Waymo working Embarcadero on both sides of Market, which is downtown.
Google tends to explicitly kill projects they don’t want to maintain instead of letting them languish. This strategy has obvious disadvantages, but it does mean that their products are usually reliable (until they get killed).
If Steve warns you 2 years in advance that he’s doing to stop working for you, with a gradual ramp down in service and comprehensive transition plan, would you really say he’s unreliable?
At this point, being let’s say a “transportation safety activist” in the Bay Area, they seem about the same.
That said, I’ve seen Waymo still being too cautious relative to Cruise. Like I was waiting on the crosswalk with my bike at a 45-degree angle, and Waymo seemed terrified that I was going to bike diagonally through on-coming traffic. It was kind of hilarious.
That said when I’ve tried to test the safety parameters on Cruise while on my bike or skateboard and the system has called my bluff. I don’t know if this means Cruise is less safe. Or they’ve found a way to give the AI a sense of confidence and assertiveness against people using micromobility to aggressively bypass urban traffic congestion.
> Or they’ve found a way to give the AI a sense of confidence and assertiveness against people using micromobility to aggressively bypass urban traffic congestion.
I'm looking forward to mass robotics.
I'm not looking forward to corporate dark patterns in mass robotics.
I live/work in SF and signed up for both Cruise and Waymo yet have never been able to get an actual ride. As best I can tell almost everyone getting rides from these services in SF are employees or specially selected. Neither is available the general public.
Waymo’s “Waymo One” app makes you fill out a pretty extensive questionnaire to sign up and suggests they will do some kind of checks about the other members of your household. Filled out rather invasive info out yet still no access.
Cruise makes you fill out a simpler form but puts you in a cue to get the actual app. Know several people who have filled it out but none have gotten the app.
This is not very scientific at all, but I see Waymo and Cruise cars driving around SF all the time. Waymo cars appear to drive much more naturally than Cruise to my eyes.
It’s hard to put my finger on any specific thing, and it may even come down to the fact that I see a lot more Waymos than Cruises, and have also seen Waymos around SF for much longer (or at least noticed them for longer).
I feel like Cruises are a bit more cautious, to the point of making me skeptical. But again, I’m struggling to think of any example in particular.
This might be silly but I also think the Cruise cars are much more ugly.
I wonder if it’s something like the “bozo bit”? Once you see one Waymo car doing something that looks unsafe, your entire perception of “Waymo car” becomes (nearly irreversibly) lower.
My take as someone in the industry, but at neither of these two companies:
Cruise is limited (by their own choice) to operating between 11 PM and 5 AM. Waymo is operating at all hours of the day. That alone indicates a much greater level of confidence in their capabilities. The complexity of driving in a busy street in SF at daytime is dramatically higher than at night, especially in SF, which tends to be pretty sleepy after hours. Note that both companies avoid driving in SOMA, FIDI, and other challenging parts of the city.
Cruise is running _driverlessly_ at nighttime. Like Waymo, they're driving autonomously 24/7 around San Francisco, just with a safety driver.
This new announcement doesn't specify their driverless hours. I suspect that they would have highlighted it if they were the first to be driving around SF in the daytime, just as they highlighted the first of being driverlessly deployed in multiple cities.
My guess (also in the industry) would be that Waymo combines Google's incredible technical skill with Google's incredible incompetence at focused execution. This is why they haven't had a milestone in a couple years since their deployment in Phoenix. I also think that this combination of strengths and weaknesses is a good fit for the AV industry.
I'd guess you answered your own question. It's an icon of a car next to some car-shaped blobs, so "double parked" or "stopped in a traffic lane" sounds like a good guess.
Considering Waymo has had almost a year of no-one-in-driverseat service in Phoenix, I would say they are overall ahead, even though that's not in SF specifically. Being first to driverless in SF also shows that. They are definitely taking things very very carefully, as they know a single major accident is all it takes to set them back years.
I think it’s “public” in air quotes - technically non-Cruise employees have taken rides, but it doesn’t seem like they’re pulling people organically from the waitlist. It’s mostly local influencers or Cruise investors in the few videos I’ve seen that aren’t Cruise employees.
They're pretty limited by the number of vehicles they have available and resources to support them. Even employees have trouble accessing the program due to the lack of vehicles and the waitlist is in the tens of thousands of people. I imagine they prioritized influencers or actual users since they can't possibly meet total demand.
People who have zero connection to Google or Waymo have gotten in the program and used the service for months. Waymo also removed any sort of NDA allowing people to post videos of the full experience, hence all the videos on Youtube. Lastly, I would still say "10 PM and 6 AM at a maximum speed of 30 MPH" is quite a major restriction compared to what Waymo can do.
When I see one of these cars in South Boston (narrow streets with street parking) during the winter (piles of snow everywhere), I'll believe they're a reality. Until then, they'll remain 90% of the way there.
Boston will certainly be one of the boss levels for AI drivers to try and beat, but even then, there's massive value in them working for even 50% of the roads out there. If my robot car said "I can take you from Rhode Island to Maine, but you're on your own while we're in Suffolk County", it would still be very useful, or it could just avoid the tricky areas and take a longer trip while I sleep!
The catch is how contiguous that 50% is: humans are terribly at switching contexts so it'd probably be okay if your car could be fine on, say, a long interstate trip but force you to handle the cities but it'd be a different story if you're zoning out on the highway but every couple miles there's some situation (road work, weather, other drivers, etc.) which it can't handle safely but you're distractedly looking up from Netflix.
A level 4 car would, by definition, never require you to take over while the vehicle is in motion. This level of automation operates safely with nobody in the driver's seat.
A level 3 car requires a driver to pay full attention and "zoning out" would be unsafe behavior on behalf of the driver. Level 3 is considered by many to be dangerous given the human propensity to zone out despite being responsible.
If you read more carefully note that this is precisely why I drew that distinction: until you reach the point of level 4 everywhere under all conditions, you are going to have build systems which safely handle that challenge of inconsistent attention. Nobody is anywhere near that consistently high level of performance so there's an important question about how you can safely implement these systems — for example, a system which could be level 4 on a well-maintained road might be suitable for a driver-less transit system with a well-defined route even if the fire department are driving their own trucks for decades to come.
One key question here is how you'd handle the cases where the system can't reliably perform at the top level. Consider the scenario I described earlier with a system which was level 4 in reasonable conditions: attempting to switch between full self-driving and human oversight frequently would be a safety hazard but you could probably safely release a system which is, say, L4 for long highway trips where a human driver could take a nap for an hour before having to take control to enter a complex city like Boston where the system would you to treat it as L3 or less for the entire trip during bad storm conditions or construction because it's safer to set the expected level of attention clearly at the beginning of the trip.
> the point of level 4 everywhere under all conditions
That isn't level 4, that is level 5.
> you are going to have build systems which safely handle that challenge of inconsistent attention
Level 4 is geofenced, but able to perform safely without any level of attention from any human driver. If a human is required to be in the driver's seat to take over on demand, then you are describing level 3, not level 4.
> attempting to switch between full self-driving and human oversight frequently would be a safety hazard
It shouldn't be. If a self-driving system can't safely reach a place to park in a region, with no human interaction, it is not level 4 in that region.
I'm certain my mother in law couldn't drive in South Boston during the winter. I'm 50% certain I can't.
Everyone keeps going on about the edge cases, about self driving cars not being ready until they can handle everything. Half the people on the road can't handle all the edge cases. They exist, they are drivers.
Let's not pretend the bar to be a "driver" is so high.
I guess there's a mismatch in some people's understanding of what a driverless car should be.
It sounds like you think "a vehicle that works in a limited set of circumstances" qualifies as a car. As a mode of transport in some areas, I get why that would seem useful. Busses, trams and trains work like that. And subways! I use all those: they're great.
I imagine the GP thinks of cars as "can go almost anywhere with limited immediate infrastructure, even some way off-road". That's how I think of them. I can imagine "driverless cars that operate like 1-5 person taxis in urban areas that have specialised roadways" could show up. But to me, these are not "cars". They don't get me to a party in the woods; they don't let me park up illegally when I need to sleep; they probably don't let me rush an injured person or someone giving birth to prepared to deal with the legal consequences later.. etc.
These things will not be "cars"; they'll be non-mass-public-transit.
> in some people's understanding of what a driverless car should be.
I don't understand at all what you mean by "should". The comment I was responding to said that [whatever this is] "wasn't a reality". This was a top-level comment to the announcement itself, not a response to a comment claiming "well, human-operated cars are officially over". Doing exactly what a current, human-operated car is an irrelevant, arbitrary point of comparison, introduced by the GP comment.
This makes exactly as much sense as saying that the first motor vehicles "weren't a reality" until, like a horse, you could take them on trails, jump over low obstacles, and live in the wilderness with them indefinitely.
That's what I'm pushing back against: it's classic ignorant Luddism to set imaginary bars for new technologies and then insist that the new tech isn't valuable at all on that basis. I know this is super-common, but it's an eternal pet peeve of mine, as I can't help but imagine how much better the world would be without people like this.
Many cars today have severely limited ability to drive on even rough roads (let alone off-road), that doesn't make them not "cars".
In general, I think the argument you are poorly phrasing the argument you are trying to make.
You seem to be trying to say that until self-driving technology can handle all use cases, you don't want to own a car without human driving controls.
I don't think such exclusively driverless cars will be marketed to consumers for a long time.
What we will see in the nearer is more standard consumer cars with varying ranges of level 4 self driving capabilities. These cars will match every one of your criteria.
What does whether or not a rural town can sustain a subway have to do with autonomous cars?
Roads currently exist that human beings can currently drive on. If you tell me you have a fully autonomous car, I would expect that to mean it can do everything a human can do on currently existing roads in all conditions that a human being can operate in. This seems entirely reasonable.
What does the behavior of current human-operated cars have to do with the value of domain-limited AVs?
Your comment was not "these are not 100% replacement for cars yet", which would obviously be uncontroversial. It was "these are not a reality yet". You've chosen an arbitrary bar for considering these "real", in exactly the same way that serving small rural communities is an arbitrary bar with which to dismiss the real value that subways create.
You won’t believe they’re a reality until AVs are perfect?
Technology doesn’t arrive perfectly in its final form. Progress for AVs is going to look painfully iterative from here. Waymo and others will roll out their AVs to increasingly more difficult areas while also reducing the number of disengagements. They will all likely have a human operator fallback for quite some time.
Edit: I'm wrong, misread the context of the quote. Please disregard the below. Thanks to the people who double-checked me!
Despite the claims of "no human driver behind the wheel", there'll still be what amounts to a driver, they're just calling the person a "specialist" instead of a "driver":
> Just as we’ve done before, we'll start with Waymo employees hailing trips with autonomous specialists behind the wheel, with the goal of opening it up to members of the public via our Trusted Tester program soon after.
I think you're misreading the post? I'm interpreting that sentence as being specifically about the Phoenix expansion:
"Our commitment to Phoenix and the community there remains strong, and we’ll soon be expanding to another area: Downtown Phoenix. Just as we’ve done before, we'll start with Waymo employees hailing trips with autonomous specialists behind the wheel, with the goal of opening it up to members of the public via our Trusted Tester program soon after."
(Disclosure: I work for Google, a sibling company to Waymo.)
Some of the rides going forward will include safety drivers, but no, this announcement is about how they took the “specialist”/driver/human safety operator out of the car for the first time in San Francisco.
Edit: looking at the full context of that quote, they’re talking about their plan for removing the safety driver in downtown Phoenix.
>Our commitment to Phoenix and the community there remains strong, and we’ll soon be expanding to another area: Downtown Phoenix. Just as we’ve done before, we'll start with Waymo employees hailing trips with autonomous specialists behind the wheel, with the goal of opening it up to members of the public via our Trusted Tester program soon after.
One thing I am glad for is the presence of Waymo in SF has upgraded the detail on Google Maps in that city. Every curb, sidewalk, marked crosswalk, and other feature of the streetscape now appears in Google Maps. It's really improving. Compare:
Google Maps has this level of detail in. S.F. and Phoenix, their two self-driving launch cities, and not in other places like say Denver or Charlotte, so I surmise this data is attributable to Waymo.
OpenStreetMap has all the curbs, sidewalks and crosswalks, too.
They are not rendered in OSM Carto [1], but show up in the data, when you click on the edit link, or load the region in JOSM [2].
Do you happen to know if this is because of Waymo? I've always wondered how much of a feedback loop there is between Waymo and Google Maps. They've never officially confirmed anything and I can't find any literature on it.
I got the impression that the extra data is not collected by Waymo, it's collected by ariel photography and streetview cards. However it is in a sense because of waymo - they need to collect this additional data to provide the best maps to the autonomous cars.
I'm in Bogotá. Traffic here is worse than anything in San Francisco. The main danger is probably the motorcycles, flying between cars, sometimes with inches to spare. I don't know the fatality stats, but they must be worse than the USA. It's hard to see how an AI could deal with this.
Is anyone thinking about "road neutrality" issues, such as:
- Tiered priority: can people pay up to give their autonomous ride priority over others?
- Differential access: can certain neighborhoods be made inaccessible to Avs originating in certain other neighborhoods? Or only accessible to friends and families of residents? (virtual gated community / virtual redlining)
This isn't really a set of "problems" that don't/couldn't exist already with human driven cars. You can already pay to use the HOV lane in some places, gated communities already exist, etc.
I'm thinking more like an algorithmic bias, which we have already seen in retail and banking ML/AI systems. Virtual version of how taxis won't serve certain neighborhoods or pick up certain people based on appearances. How do we ensure that autonomous vehicles don't surreptitiously implement biased behavior like this?
And I think it's actually a problem if it's possible to pay up for expedited routing. Implementing faster car travel for rich people is not the same as HOV lanes.
I hope that Boston Dynamic can do will make robotic horses/tauntauns that need to be tied to a charging post whilst I head into the daily shopping would be awesome.
And if I took one to the pub and later that night it could get my intoxicated body home safely without my guidance yeah that would be a huge plus.
I'm tired of cars and trucks and segways and scooters.
And I have this dream of owning and driving around this very particular supercar. I've been slowly saving up, and within a few years I'll probably be able to outright buy it. And I plan to keep it for about 15 years atleast.
This car is neither electric nor self-drive capable.
Should I abandon that dream because I might not be allowed to drive around for too long?
And yes, I do understand and agree that in the long run, autonomous cars are probably going to be safer overall. Like, I'm not opposed to it.
(Please don't lecture me on why buying a car outright (instead of leasing) is financially stupid, or why buying cars with internal combustion engines or high CO2 emissions is a crime, or how my conscience allows me to buy a fancy car when there are a million other problems in this world).
Definitely do not abandon your dream car. human drivers will be on the road for the next 50 years or more, and I suspect that even after most cars on the road are self driving, there will still be car shows and events for the classics. At least I hope so given my dream car is a Plymouth Superbird and I'm not close to attaining it quite yet.
"On March 24th, 2022, I was riding my bike near the beach in San Francisco down a slight slope. My hands were on my brakes to make sure the bike was in control. About 150 feet ahead of me, a vehicle suddenly stopped in the middle of the road. To avoid oncoming traffic, I veered towards the vehicle's right, which suddenly moved directly in front of me.
At that point, I could not avoid impact and crashed through the vehicle's back window with my bike. As I lay on the road, I saw that the car that hit me did not have a driver."
I can not wait until the day cars no longer break any rules and the police are forced to take a hard look at what they do and justify being employed. How will they civil forfeit if they have no reason to pull you over? I really hope I live long enough to see that day.
I have seen this news but have not seen any videos taken by a rider. On the other hand, members of the public have uploaded videos of their experiences in Waymo (Arizona, driverless).
In the case of the Uber autonomous car which killed a pedestrian in 2018, Uber reached a settlement with the victim's family and faced no criminal charges (I think the answer to your question is clearly yes). The "driver" behind the wheel of the autonomous car did get charged with negligent homicide in addition to being sued by the family afaik.
> Here's Steve, who joined us for a special drive on a carefully programmed route to experience being behind the wheel in a whole new way. We organized this test as a technical experiment, but we think it's also a promising look at what autonomous technology may one day deliver if rigorous technology and safety standards can be met.
Does anyone know what Waymo is doing in Phoenix? I saw a lot of there cars there but they were always being driven by someone in the driver seat (Jan-Feb 2022). They say they are driverless there but I never saw it and probably saw 500 Waymo cars/Waymo cars 500 times while there.
Waymo has a relatively small service area in Chandler, Mesa, and a very small section of Tempe. The first two are essentially Phoenix suburbs. Waymo doesn't reach some of the areas that would seem to be particularly useful: light rail stops (the Phoenix area now has about 20 miles of light rail: see the map tab at https://www.valleymetro.org/maps-schedules/rail), or Arizona State University in Tempe, or some of the party/bar districts (like Mill Avenue in Tempe).
I live in the Phoenix area and have wanted to try Waymo, but, sadly, it doesn't go anywhere useful for me.
Now we seem to have very clear evidence that both waymo and cruise consider LIDAR sensors as essential, while Tesla is still trying to do "full self driving" with visual spectrum cameras only.
I am mostly ready for my own self-driving car (with my stuff inside and my settings and my cleanliness) that can Drop me off at the front door of my hotel and ferry itself back to the parking garage; or one that I can drive myself to work, and then it ferries itself back home to pick up my wife and kids to get to school. I don’t need my car to drive me necessarily, I just want it to go on and perform the next task for me. It can perform these autonomous tasks slowly and safely.
"You've driven through the area where somebody used a stolen credit card with their Google account. Your Waymo account is now blocked, please leave the car."
I wonder what kind of bubble tech bros and the people running these companies live in.
Ten years to transition? Try twenty at least. Most people still buy second hand vehicles and those are still gonna be gas. Think about the motorcycles too, and the pedestrians, and the cyclists.
Adoption will be much slower due to costs, safety concerns and privacy.
It would be better to just spend the money on better infrastructure and public transport, but that’s a pipe dream in the States at least.
I moved to live in a more walkable, less car dependent city. I care about us improving transit infrastructure, but the idea that American cities will significantly divest from their existing car-centric infrastructure is just wrong. We’ve already invested in that infrastructure and it is incredibly expensive to replace.
I think the replacement of manually driven vehicles will be decently rapid - the bottom 1/3rd of drivers cause a large percentage of motor vehicle accidents, as soon as they’re better than that group I think the financial incentives (like insurance) will dramatically change and it will be cheaper to ride in someone else’s self driving car than to own your own older vehicle.
If legit self-driving capabilities became available, I feel pretty confident in saying that there's a sizeable consumer market for this that are not sensitive to the price. I would immediately drop an extra 30-40% on a car if that's what it cost to get that self-driving.
For those that are more price-sensitive, well, new cars become used cars in a relatively short amount of time.
I wonder if it'd just be cheaper and easier in the long term to create some sort of track system that the robotic cars could drive on? You could even have one car pull other "dumb" self driving cars that could hold cargo or even more passengers.
Maybe even have them powered by electricity straight from the grid instead of batteries.
A good illustration of why autonomous EVs will stop trains in their tracks.
Tracks are expensive, precise and single-purpose. A road is cheap, flexible and multi-purpose. Being tied to the grid means you can't go where you need to get to, only pre-set destinations. And since every train-car can now drive itself, there's no longer a need to tie them together, so instead of 5 cars bunched up all at once every 10 minutes, you can get a single car every 2 minutes.
That is, once this tech matures, no new train development will make sense.
Trucks that drive highways solely automatically, I'd expect would come way before cars navigating cities. I'd think trucks could just pull over on the shoulder in unusual conditions, e.g. bad weather, construction.
I suppose it's more a legal issue. Cars just need permission within a city, whereas trucks cross lots of legal boundaries?
There's also a perception challenge that doesn't exist in cities. At highway speeds and longer stopping distances, you need your perception stack to see much further. Still an active area of research.
I'm hoping that self-driving trucks result in separate highway lanes, or perhaps even dedicated roads. A road just for self-driving trucks wouldn't need a passing lane, or a breakdown lane (if a truck breaks down, someone can be dispatched to fix it, and it's okay if the whole roadway of autonomous trucks comes to a stop).
(Yes I know, I've just reinvented trains, but trains would likely coexist with this, and they serve different purposes (trucks are better at going over mountains, for instance))
With some minimal standardization, these could also drive with almost zero distance between each other, braking together. This would significantly reduce air resistance and road space usage.
That's been the confusing thing for me. I'm sure automated trucking is being worked on, but all the media/conversation seems to be around consumer vehicles.
Automated trucking is a much more impactful, and frankly easier problem to solve. You could simply automate highway part and have manual take over near cities to remove 90% of the challenge.
Wayve pushes very hard on what is called "end-to-end" approach. That means, you input sensing into the deep network and out comes the controls. No ifs and buts. When it came out on 2017, no one I knew believed that it can possibly work with reasonable reliability to actually put things in production. Deep learning is simply not there. You can converse with GPT-3 and find few great example to cite but within few tried you will encounter things which don't make sense. Deep learning is a great component to provide pattern recognition but there are still a lot many more pieces such as memory, abstractions, compositions etc that we have no idea how to do it. The pattern recognition can certainly emulate those pieces to certain extent but it's not the same thing. Consequently, E2E self-driving only using current state of deep learning is unlikely to produce something beyond video clip demos.
I am not a Tesla fanboy but could be highly misinformed, so honest question: Is it generally accepted that Tesla's approach is the worst in the industry?
How will this scale? As they release to more locations there will be a higher likelihood of a bad accident. Once there is any bad accident the whole fleet is shut down. On the other hand, if an Uber driver has a bad accident only the driver is fired.
The graphic showing the platforms from Austin to SF shows a disturbing trend of these autonomous platforms getting bigger, and therefore more dangerous for those of us on the outside of these vehicles. I would love to see regulations capping the size and speed of these autonomous platforms to minimize the danger to the public.
Sure, eventually they might be proven to be safer than having humans behind the wheel, and then we can consider relaxing those regulations, but even then perhaps the standards should be higher, since they are likely to increase the overall number of cars on the road significantly - traffic will not discourage people from driving as much, if they are not in fact driving.
(FWIW, I would love love love to never have to drive again. I just want us to take a measured approach to getting there for safety.)
Considering that cars already kill people en mass there's no real advantage to slowing the pace of SDV development other than to make people feel warm and fuzzy. I personally prefer Tesla's approach where they give customers permission to go ham with their latest beta software. The quicker we get to L5 the quicker we can put the automobile safety nightmare to rest permanently.
Given that cars already kill people en mass, there's no real advantage to slowing down something that will likely 10x the number of cars on the road, or making those cars smaller and safer?
Sure, let's get to L5 as quickly as we can...without killing lots of people unnecessarily.
Is this really "fully autonomous"? I get the feeling the driver (the one with the legal liability) is watching from a remote terminal back at waymo hq.
There’s no remote driver. There are people at a remote center that can add labels to scenes if the computer doesn’t have enough confidence to proceed. One the computer has enough information to start planning again, it resumes driving.
I took a fully autonomous Waymo in phoenix a year ago. Very cool. A little jerky and at one point it said "Figuring this out" but whatever it was got resolved very quickly.
Also, it was a pain to book it. I tried once and there were no vehicles available. I tried a couple days later and it told me I would have to wait a half hour for a ride.
>Just as we’ve done before, we'll start with Waymo employees hailing trips with autonomous specialists behind the wheel,
Ok. I tell you what, I've got an innovative idea. You see, these "autonomous specialists". They sound expensive, right? Got to be trained not just to drive, but to debug the computer, to have higher concentration levels to mitigate the inevitable higher risk of sudden interventions. So what you do - and stay with me now - you strip out the autonomous tech and just replace it with a steering wheel. All of a sudden you no longer need these useless "autonomous specialists" and anyone could drive it. It's a billion dollar idea, I guarantee it!
The fall-out is going to be intense:
* Fuel-stops are going to change completely, what's the point of half of the motels on long-haul drives when your car can drive all night and likely recharge automatically at a stop-point.
* Reduced downtime on goods movement will impact uptime in every other industry.
* Autonomous-vehicle-only lanes that line up with traffic timings creating a two-tier driving experience
* Huge influx of unemployed drivers who I guess might get "chauffeur" style jobs.
* Security issues, complex legal issues when there's accidents.
* Fake taxi's that drive a customer into a bad experience for "lols".
I know all this is extreme, but if history is studying the past to understand the present, science-fiction is studying the "future" to predict the problems of tomorrow.