Isotack
Active Member
EVs are going to eliminate a gas station that is a blight and attracting sketchy folks? How does a gas station do that exactly? Wow...how bout the injection/respite sites? Seems that they do the job pretty well.
|
|
|
The Coming Disruption of Transport
- Thread starter kEiThZ
- Start date
CityStay
Active Member
Gas stations are almost always a blight, not sure about the sketchy folk. Maybe they're site specific?
I could see many of them being turned into EV charging stations. I think they're already doing that in Norway.
(the gas stations, not the sketchy folk!)
mdrejhon
Senior Member
Referring to this chart:
These are revolutionary price drops of lore. We're familiar with tech price drops ($5000 laptops of 1980s becoming $500 this year) but this unusually rapid for a mostly-chemical assembly. I rate this similar to Thomas Edison's promise to make the (formerly luxurious) light bulb cheaper than a candle. Or like the how formerly luxurious cotton textiles was able to be mass-market-priced thanks to new automatic looms of the 1800s. Now even the poorest wear cotton T-shirts that are simply absolutely opulent by 1700s standards -- the fabric quality of your ratty shirt at bottom of your drawers was worth literally $1000 (inflation adjusted) not too many centuries ago when those tiny threads had to be hand-weaved! Today, lithium battery price drops this is one of those "price drops of lore" moments.
I still remember when the first lithium laptop batteries cost a giant fortune, the first batteries costing a significant percentage of a $3000-to-$5000 laptop -- with replacement batteries costing many hundreds of dollars. For just a couple dozen watt-hours in those early lithium laptop batteries. A single replacement laptop battery cost more than the cost of today's laptops!
Now, we are hurtling to $62 for a battery (at least in factory bulk quantities) that has enough juice to power an electric space heater for almost an hour.
And now Eglinton Crosstown LRT will be battery powered during power blackouts with the new Metrolinx lithium backup-battery farm connected to the LRT catenary they're building for the Crosstown -- It's an uninterruptible power supply (UPS) for Crosstown. This new Metrolinx battery is a 30-megawatt-hour battery, similar to about 8 million Duracell AA batteries (3.75 watts-hours).
To other forum readers, you didn't know that Eglinton Crosstown is going to be battery powered? Tech improved so fast, prices fell so fast, that Metrolinx cancelled the Crosstown gas peaker plant and switched to a battery bank as the power backup for the upcoming Crosstown LRT. Back at the time Crosstown LRT was designed, a formerly-expensive battery was a pipe dream -- but -- as your graph indicates, it quickly became cheaper to build a giant battery for Eglinton Crosstown.
Definitely won't be the weak link in autonomous vehicles. There's so many other weak links.
Last edited:
mdrejhon
Senior Member
Yes...
That said, there'll be way more accountability to worry about in some cities, and less accountability to worry in others. That's why it's much easier to deploy legally in Arizona than Ontario. Again, my spouse studied legal.
There's less legal worry when ultra-car-optimized rich cities in great climes are severely punished/penalized for minor infrastructure deficit such as simple flaked road paint (= impeccable car infrastructure) or when there's nearly no pedestrian rights in a specific juridisction. And so on, and on.
Many reasons why it's legally easier to let cars roam there. A novel's worth. And now, some cities now lets them 45 miles per hour without any safety drivers! (WayMo, circa 2019, citation here) The car industry made sure of that in some American cities; so it just makes self driving cars a bit legally simpler.
Yes, a Uber self driving car hit a cyclist, but that's a symptom of flawed software culture and go-go fever -- I'm surprised that it didn't happen earlier, but it's just a symptom not needing to worry about rare pedestrians -- due to pedestrians not even daring to jaywalk in cities where it's a occasionally a jailable offense (if done repeatedly). In some cities, it's a police state for pedestrians -- they only cross at intersections only when the WALK appears. Strong enforcement occuring. In some cities, they don't step even when the roads are empty. I've been there. Seen first hand. Pedestrians have almost no road rights in some locations, police-stated into obedience. So, not all cities are as legally easy to let empty cars roam. You don't see a police state, but if you go to their schools in their state system it's taught pretty aggressively. The moms/dads thump it into their kids about pedestrian obeyance. And you see tickets quickly slapped to jaywalking pedestrians in some cities. Compilance is so high -- that even at 2am pedestrians just stand at a DONT WALK sign in empty roads, where we Torontoians/Ontarians/Hamiltonians think nothing of crossing an empty road. In some cities that it's almost a free passport for self-driving cars not needing to have many optimizations for pedestrian avoidance (at first). That's what happened. Now they're scrambling to fix and strengthen thanks to the Uber accident. But the underlying anti-jaywalk culture is still there.
Toronto is roughly in between Amsterdam (where no concept of jaywalking exists) and an ultra-car American city (where jaywalking is almost jailable). Anyway, it's surprising to see how liberally permissions were granted to many self driving cars with weak pedestrian-avoidance / cyclist-avoidance -- and no disaster happend until that self-driving Uber accident killing the cyclist. I'm amazed it didn't happen earlier, but there were almost no pedestrians to worry about, with that police-state-league of enforcement of jaywalkers in some cities. Now AI is improving fast to fix this problem, to also train for this and minimize odds, even if you can never completely design it out (sudden kids popping between parked cars closer than the car's stopping distance ability), but it does legally help when a jurisdiction has far fewer pedestrian rights. Fewer checkboxes for insurance companies to worry about too.
Here in Ontario, we have a lot more complexity letting cars roam. We have more pedestrians crossing the roads on "Don't Walk" flashes or mid-street. We have more kids playing in the middle of roads. We have something called a "winter" that requires Ontario to dream up new legal rules. (For example, one of the >100 regulations will be mudane things like mandating what self driving cars must do when unexpectedly caught in winter away from a safe parking spot -- and whatever first rules they dream up will be contentious/imperfect as I explained). We have infrastructure deficits here that doesn't exist in other places. We have flaking road paint here that's treated as if it's illegal to exist in some rich American cities (= very prompt maintenance of road paint, nearly no infrastructure deficit). Etc. Etc. Etc. When potholes simply don't exist even in winter (full week vacations of car-rentals and zero potholes yearround, rarer than four-leaf clovers) because of a legislative mandate of zero infrastructure deficit (highly vocal public + quick reporting + repairs made promptly with faster haste than ambulance) -- you might as well not need to worry technologically about programming pothole advoidance if there's no legal mandate that self driving cars must not clog a road when damaged by a pothole. There's a lot of cascading-domino effects with thousands of rules with thousands of technological nice-to-have features for an AI training regimen fighting against corporate cultures, lawyers, governments, etc.
Have you been to enough of the most pro-car parts of America to see how ultra-car-legal-optimized it is?
It's a giant legal-ease chasm for Ontario versus most "pro-car-legal-frameworked" US cities. Some cities have coincidentially/accidentially made it easy (technologically AND legally) for truly driverless cars to begin in some cities far ahead of my predictions I made a few years ago, and now I understand a small taste of the unforseen complex interplaying elements why. And frosting on top, some of those cities roll the red carpet to drivereless fleets in go-go fever.
(P.S. When replying, remember, some of these are cherrypicked examples to give a remote taste of the legislative/legal difficulty chasm. There's 10x-100x more detail I haven't mentioned here)
(P.P.S. On the record, I don't like anti-jaywalking rules. I feel urban roads are for people, not just cars. Just, pointing out legendary car-culture-cities far beyond any Canada road culture, even Alberta's. And beyond well-known Los Angeles and their freeway love. Jay walking is somewhat more permitted in California once you're off the freeway ramps. Despite those $250 ticket in Los Angeles if you started walking with a flashing "DONT WALK" in downtown LA)
Legal is definitely one of the many weak link and will throttle the pace somewhat in Ontario. And interplay between legal + technological. All pushed to a more realistic timeline well beyond the 5-year predictions made in the video in the first post of this thread.
Last edited:
MisterF
Senior Member
Battery banks are going to become a big part of the electricity grid as costs continue to drop. Issues with wind and solar power being generated when it's not needed can be solved by storing that power in batteries for later use. The fact that those power sources have been getting cheaper too doesn't hurt. The Crosstown project is just the beginning.Referring to this chart:
These are revolutionary price drops of lore. We're familiar with tech price drops ($5000 laptops of 1980s becoming $500 this year) but this unusually rapid for a mostly-chemical assembly. I rate this similar to Thomas Edison's promise to make the (formerly luxurious) light bulb cheaper than a candle. Or like the how formerly luxurious cotton textiles was able to be mass-market-priced thanks to new automatic looms of the 1800s. Now even the poorest wear cotton T-shirts that are simply absolutely opulent by 1700s standards -- the fabric quality of your ratty shirt at bottom of your drawers was worth literally $1000 (inflation adjusted) not too many centuries ago when those tiny threads had to be hand-weaved! Today, lithium battery price drops this is one of those "price drops of lore" moments.
I still remember when the first lithium laptop batteries cost a giant fortune, the first batteries costing a significant percentage of a $3000-to-$5000 laptop -- with replacement batteries costing many hundreds of dollars. For just a couple dozen watt-hours in those early lithium laptop batteries. A single replacement laptop battery cost more than the cost of today's laptops!
Now, we are hurtling to $62 for a battery (at least in factory bulk quantities) that has enough juice to power an electric space heater for almost an hour.
And now Eglinton Crosstown LRT will be battery powered during power blackouts with the new Metrolinx lithium backup-battery farm connected to the LRT catenary they're building for the Crosstown -- It's an uninterruptible power supply (UPS) for Crosstown. This new Metrolinx battery is a 30-megawatt-hour battery, similar to about 8 million Duracell AA batteries (3.75 watts-hours).
To other forum readers, you didn't know that Eglinton Crosstown is going to be battery powered? Tech improved so fast, prices fell so fast, that Metrolinx cancelled the Crosstown gas peaker plant and switched to a battery bank as the power backup for the upcoming Crosstown LRT. Back at the time Crosstown LRT was designed, a formerly-expensive battery was a pipe dream -- but -- as your graph indicates, it quickly became cheaper to build a giant battery for Eglinton Crosstown.
Definitely won't be the weak link in autonomous vehicles. There's so many other weak links.
crs1026
Superstar
^ The train guy in me wonders .... if batteries are going to be that good and that cheap, how far off might a disruption to locomotive technology be?
If the barrier is price, one would hope to see proof-of-concept models hit the rails soon, at least as test beds.
If the battery technology has to improve just a little more, in terms of Kwh/unit weight, or charging time, etc.....it may take longer.
But what’s good for the Tesla ought to be good for the Acela.
- Paul
If the barrier is price, one would hope to see proof-of-concept models hit the rails soon, at least as test beds.
If the battery technology has to improve just a little more, in terms of Kwh/unit weight, or charging time, etc.....it may take longer.
But what’s good for the Tesla ought to be good for the Acela.
- Paul
TrickyRicky
Senior Member
44 North, good point on power of ICE. I still use gas powered lawnmowers but would love to convert. One of the issues with battery powered small engines is that the battery designs are all proprietary. If battery docking was universal it would help. I don’t need to fill my mower with proprietary gas blend with a proprietary gas nozzle.
Alvin, thanks for posting those articles. I don’t think however they either address my concerns nor give me greater confidence in the automated car design philosophy. The Tesla examples do the opposite actually. First Tesla arrogantly claims they have achieved full automation. Next they defensively blame externalities for instances where there technology fails causing death. Sorry dudes it’s not the world it’s you.
The more I look at it I’m not sure what the end goal is for full automation. Is it universality in which case I think you need an AI that makes more emotional decisions (emotions are not inferior decision making systems they are primary decision making systems). Or is it standardization where the world becomes like a very complex idealized “train track” and autonomous vehicles navigate personalized loops on that track with reduction input from the user
Alvin, thanks for posting those articles. I don’t think however they either address my concerns nor give me greater confidence in the automated car design philosophy. The Tesla examples do the opposite actually. First Tesla arrogantly claims they have achieved full automation. Next they defensively blame externalities for instances where there technology fails causing death. Sorry dudes it’s not the world it’s you.
The more I look at it I’m not sure what the end goal is for full automation. Is it universality in which case I think you need an AI that makes more emotional decisions (emotions are not inferior decision making systems they are primary decision making systems). Or is it standardization where the world becomes like a very complex idealized “train track” and autonomous vehicles navigate personalized loops on that track with reduction input from the user
MisterF
Senior Member
Simple, the end goal is safety. Automated cars are better and safer than any human driver in every conceivable way. Cars kill 1.5 million people every year IIRC, a number that with any other technology would be a global crisis. Autonomous cars will reduce that number by a huge amount. And the technology is only improving.
Isn't that the whole point of AI though, that it learns? It doesn't have to have every conceivable scenario coded into it. That's what autonomous cars that companies have been testing for years now have been doing - they've been learning. Any fully autonomous car will have the collective experience of all that learning built in.
AlvinofDiaspar
Moderator
Expecting AI to be infallible shouldn't be the goal here (and I think it is misguided to expect AI to prevent all deaths, or not cause any deaths inadvertently due to programming) - what should be the goal is overall reduction of accidents/fatalities from the current human driven system. Trees vs forest.
And just because "emotional" decision making is predominant does not make it superior - the current state of affairs in road safety is illustrative of that. When you can demonstrate that automated system is safer and more reliable than those involving human operators, any point against adopting it will be emotional and not logical in nature.
AoD
Last edited:
crs1026
Superstar
I think that's generally understood, actually. That doesn't mean the trees are irrelevant - we live in a world where a single lawsuit over a single event can set precedent for everything going forward. There's a saying - bad facts make bad law - ie a dispute that invovlesr an odd, perverse set of facts is a terrible place to establish a precedent for more straightforward issues. Plus, every person harmed in an AI car has a family and an existence. We may not stop the train, but we have to give appropriate consideration for those who will be harmed while we work the bugs out.
We are a raising a lot of granular "what if" scenarios that may only happen once in a billion cars. They may be trivial in the big picture, but they reflect a general unease about just how much has to be worked out before the cars hit the roads, versus how much ought to be left to resolve itself through (often painful) experience.
I do think that law and regulatory paradigms have to be changed before we roll out. The comment made about Tesla's corporate arrogance rings true, and Tesla isn't particularly different in that regard. Just wait until the cars hit the streets and we have disputes over whether a problem is a province's, or ottawa's, or maybe beyond. Our lack of faith in corporate and government ethical behaviour is well earned and proven. Let's fix some of those issues of accountability and transparency before we set the cars loose.
- Paul
TheTigerMaster
Superstar
It’s 2020. Where are our self-driving cars?
www.vox.com
Also check out the related Hacker News thread. Some interesting discussion regarding the difficulties, and potential impracticality, of self-driving: https://news.ycombinator.com/item?id=22328293
It’s 2020. Where are our self-driving cars?
In the age of AI advances, self-driving cars turned out to be harder than people expected.
www.vox.com
Also check out the related Hacker News thread. Some interesting discussion regarding the difficulties, and potential impracticality, of self-driving: https://news.ycombinator.com/item?id=22328293
Admiral Beez
Superstar
If there‘s one space that’s desperately needing disruption its the TTC’s monopoly on mass transit.
mdrejhon
Senior Member
Zero stopped charging time is needed.
Full charges (within common SoC range) are gentle and doable if approximately 50%+ of a route has catenary, for on-the-fly charging. No need to stop train to recharge.
It is possible by 2030s that future battery-backed trains may be what extends GO EMUs to 15min RER in Hamilton, Ontario. Bypassing the contiguous-catenary question.
Just electrify contiguous catenary to Burlington, and catenary at the Lewis Rail yard (or at Downtown GO) — and have enough batteries embedded in the EMU, and it should be quite doable to skip catenary on CN freight lines for electric trains.
160kph Battery Train Technology In GTHA (Electrification Solution for Kitchener Line)
As a longtime armchair GO analyst around here (including my well-known GO article from 2015), I have discovered lots of evidence that Metrolinx is now seriously considering battery trains for Kitchener Line. Summary (For Those Unfamiliar With Battery Trains): Battery trains have arrived in...
MisterF
Senior Member
I have a couple more thoughts on this. Tony Seba is predicting that with the speed of AI advancement, autonomous cars will have all the kinks worked out and be ubiquitous by 2030, and that transportation as a service will be 10x cheaper than owning a car. He's also predicting that because of this, the vehicle fleet will fall by 70% and the remaining cars will be continuously driving. And that the cost savings and increased productivity will be a tremendous boost to the economy, and that the need for parking lots will be drastically reduced.
This has a few implications that he hasn't addressed. First, rush hour will still be a thing. For all the hype of telecommuting, face to face contact is incredibly important in most sectors and the vast majority of us need to get to a physical workplace. So the minimum number of cars we'll need is how many are required to deal with rush hour. At non-peak periods and especially at night, a lot of cars are just going to be sitting somewhere, so we might need more parking lots than Seba thinks.
Second, if car transport is going to be the same annual cost as a Netflix membership, it makes car transportation essentially free. And since reducing the cost of something, drives up demand, people are going to take trips more often and farther than they do today, and that means more cars and more traffic.
I'm still bullish on the future being electric, automated and on demand, but I do think that it'll be more complicated than Seba thinks.
This has a few implications that he hasn't addressed. First, rush hour will still be a thing. For all the hype of telecommuting, face to face contact is incredibly important in most sectors and the vast majority of us need to get to a physical workplace. So the minimum number of cars we'll need is how many are required to deal with rush hour. At non-peak periods and especially at night, a lot of cars are just going to be sitting somewhere, so we might need more parking lots than Seba thinks.
Second, if car transport is going to be the same annual cost as a Netflix membership, it makes car transportation essentially free. And since reducing the cost of something, drives up demand, people are going to take trips more often and farther than they do today, and that means more cars and more traffic.
I'm still bullish on the future being electric, automated and on demand, but I do think that it'll be more complicated than Seba thinks.
That's fine in theory, but reality doesn't really work that way. Whether it's the internet or smartphones or cars themselves, governments are rarely if ever in front of new technologies. And it's not just governments, it's everyone. Disruptive technologies have a way of changing things in ways that nobody could have predicted and only make sense in hindsight. When the internet was emerging, nobody could have predicted that a humble search engine would power our smartphones and TVs, lead the way in autonomous car research, and allow Youtube stars to be a thing. When the first airplanes started taking paying passengers it would have made no sense to ask governments to regulate a global airline industry. It's the same with autonomous cars; they're going to change the world in ways we can't even imagine yet. The regulations and the technology develop in parallel.
crs1026
Superstar
Totally agree, most of the effects are things we haven’t anticipated and won’t see coming.
I do feel confident in predicting there will be a “regulatory lull” that delays the rollout. Government will be reactive, I agree, and they will not want to be seen as getting in the way by imposing strict controls. But insurance companies? Just wait til they start quoting rates for insurance. With no actuarial track record, they will be very conservative. I can see school boards for instance seeking injunctions that require self driving cars to select routes that divert away from school or playground zones. Perhaps ratepayer groups will do likewise to keep traffic out of their neighbourhood - a new form of Nimbyism.
Can’t wait to see the slogans that the liability lawyers put on billboards ;-) “Hurt by your car ? Call.......”
- Paul