I wish all the ink spilled on controversy over Tesla would instead be spilled on doing basic safety improvement to the road system that are cheap and proven to save many lives.
Lets be real here, automated driving or not, having actually save roads helps prevent death and harm in many cases.
The hyper focus on high tech software by all the agencies engaged in 'automotive security' is totally wrongly focused. What they should actually do is point out how insanely unsafe and broken the infrastructure is, specially for people outside of cars.
See: "Confessions of a Recovering Engineer: Transportation for a Strong Town"
If we're actually serious about this, we should get rid of roads altogether and replace them with rail tracks. That solves 95% of the automation problem anyway.
Yes I am absolutly for that in a lot of cases. But lets be real, you are not gone end cars anytime soon. Lots of cars exist, more will exits.
Even with the largest possible investment in rail, cars will exists in large numbers.
So yeah, rail and cargo tram ways in cities are great. But we can't just leave car infrastructure unchanged.
Specially because existing car infrastructure is already there and cheap to modify. Changing a 6 lane road into a 3 lane road with large space for bikes and people is pretty easy.
Replacing roads with railways means nearly complete reliance on a central authority for all transportation needs. This authority decides who gets to leave town and who doesn’t, which companies have priority when delivering goods, how or if the parallel railways should scale to meet more transportation demand, etc.
Transportation safety is important but it shouldn’t be considered in isolation when there are potentially catastrophic consequences to prioritizing safety at the cost of everything else.
with roads, you already have a central authority (the DMV) that decides who can drive and who can't. Try taking your home-built motorbike with no licence plate out for a spin on public roads, you'll likely get pulled over by police.
In the US trains have 17 times fewer deaths per passenger-mile than cars, and even then less than 1% of deaths from trains are passengers (the overwhelming majority are trespassers).
That it could be even better does not mean it is not a substantial improvement.
Just let Japan build US rail, and copy paste Dutch city planning, guarantee it’ll be much better and cheaper than anything anyone in US government would come up with and requires 0 brain power.
I disagree. If someone is a bad driver, and causing crashes, we don't say "we need to improve the road system". We suspend the drivers license until they can prove they are capable of being a safe driver. We should hold this software to the same standard. Until it can demonstrate safety at or above human level, it should be outlawed.
Road systems should always be worked on, but when a crash happens its usually the drivers fault, except in a minority of cases where bad road engineering is to blame. this self driving is fucking up enough that it cannot be blamed on the roads anymore, if it ever could.
> I disagree. If someone is a bad driver, and causing crashes, we don't say "we need to improve the road system".
Well yes, and that is literally exactly the problem. That is exactly why its so unsafe in the US. Because instead of building a safe system everything is blamed on people.
In countries that take road safety seriously, every crash is analyzed and often the road system is changed to make sure it does not happen again. That is why places like Finland, Netherlands and so on have been consistently improving in terms of death and harm caused by the road system.
Again, the book I linked goes into a lot of detail about road safety engineering.
> We suspend the drivers license until they can prove they are capable of being a safe driver.
An unsafe designed street often leads to situation where even good drivers intuitively do the wrong thing. Again, this is exactly the problem.
If you build a system where lots of avg. drivers make accidents, then you have a lot of accidents.
> We should hold this software to the same standard. Until it can demonstrate safety at or above human level, it should be outlawed.
Yes, but its a question of how much limited resources should be invested in analyzing and validating each piece of software by each manufacturer. In general software like Tesla AP would likely pass this test.
I am not against such tests but the reality is that resources are limited.
> Road systems should always be worked on, but when a crash happens its usually the drivers fault, except in a minority of cases where bad road engineering is to blame.
I strongly disagree with this statement. Its a totally false analysis. If a system is designed in a way known to be non-intuitive and leading to a very high rate accidents then its a bad system. Just calling everybody who makes a mistake a bad drive is a terrible, terrible approach to safety.
Once you have a safe road system, if somebody is an extremely bad driver, yes taking that person of the road is good. However in a country where so much of the population depends on a car, that punishment can literally destroy a whole family. So just applying it to anybody who makes a mistake isn't viable, specially in system that makes it incredibly easy to make mistakes.
The numbers don't even show the problem, the unsafe road system leads to less people walking in the US, and somehow still creating a high rate of deaths for people who walk.
America hates any solution to a problem that doesn't involve training individuals to act differently. I constantly hear people who want warning labels gone and periodic PSAs against mixing bleach and ammonia removed, people who don't want mandatory baby left in a hot car detectors, people who hate OSHA, etc.
We think the solution should always be human. That the way to solve climate change is homesteads, the solution to poverty is individuals working harder, etc.
When anything we don't want happens, we think "I'll try harder next time" not "We will eliminate this as a possibility".
People treat life as a sport rather than an engineering challenge, they want to "win fairly", not make it impossible to lose, they want everything they do to say something about their own ability rather than say something about a clever process that makes individual skill irrelevant.
I really don't exactly like that aspect of humanity, constantly making everything into a sport.
Positive sum games >>> zero sum games. The solipsistic individualistic mindset stems from the zero-sum game system IMO and that's a larger social and biological issue.
It's bigger than zero sum. The people I'm talking about have amazing amounts of altruism and community ethics.
They truly seem to want the whole world to succeed together, just... excluding the people who want to put you in jail for sneaking lead fuel in your truck, and anyone who doesn't lift weights and eat lots of meat, or anyone who complains about them going out while having a cough spreading germs.
Go look at Salt Lake City. Its is true, Salt Lake City despite horrible bad road design, like Huston. But because people drink less they have slightly less accidents. However international they are still terrible. And that is with there being very little walking.
If they had pedestrian and cyclist numbers like Amsterdam it would be 24/7 mass murder.
Tesla publishes their safety data quarterly no need for your many assumptions and speculations - Teslas are already much safer than the average driver, especially when autopilot is on
One interesting bit from a different 3rd party study:
> Cambridge Mobile Telematics also found that people driving Teslas were 21% less likely to engage in distracted driving with their phone in their Tesla compared to when they drove their other car.
Maybe the software integration helps avoid this? Lots of other cars have much more complicated interfaces to hook up calls and reading texts. My mom struggled to figure out her car even supported Android Auto.
It might just be it's a higher end car, but they didn't see it for an EV Porsche
> These findings include an analysis of Tesla drivers who also operate another vehicle. These drivers are nearly 50% less likely to crash while driving their Tesla than any other vehicle they operate. We conducted the same analysis on individuals who operate a Porsche and another vehicle. In this case, we observed the opposite effect. Porsche drivers are 55% more likely to crash while driving their Porsche compared to their other vehicle.
The reduction in speed is likely influenced by automated driving, especially considering how fast a Tesla car can accelerate vs normal cars:
> They were 9% less likely to drive above the speed limit.
> Maybe the software integration helps avoid this?
With auto pilot on the car is watching you. If you take eyes off the road it issues more pay attention nags. Failure to comply removes FSD Beta. So you have a feedback loop where paying attention becomes more important then your phone.
FSD is an extension of the autopilot capability and the autopilot capability has proven extremely safe. As soon as there is statistically enough data they will release FSD numbers too but that means tens of millions of miles logged on roads since crashes are generally uncommon
Is autopilot available for all roads? If not, that’s a significant and unstated statistical bias in this data. To have a true comparison, you will need to include only human accidents on roads where Teslas can use autopilot.
Something is changing and it’s very broken so people are paying attention to it.
If people really wanted to fix transportation that’s great, high speed rail and public transportation reducing the number of cars on the road seem to be the best solution.
But hey, Elon’s hyper loop was a publicity stunt to discourage investment in that. So I say, whether you want to shit on Tesla or public roads, shit on Elon.
> reducing the number of cars on the road seem to be the best solution.
No actually its actually not. Less concession in a system that depends on concession for safety will lead to more accident not less.
That is what was shown during Covid, less driving, but accidents per mile went up.
So yes, of course public transport, bikes are great, but if you don't fix the underlying problem in the road system, you are gone have a whole lot of accidents.
> But hey, Elon’s hyper loop was a publicity stunt to discourage investment in that.
This is a claim some guy has made, not the truth. What is more likely is that Musk actually thinks Hyperloop is great (its his idea after all) and would have wanted investment in it.
> shit on Elon
I prefer not to shit on people most of the time.
Musk is the outcome of a South Africa/American way of thought that is more in line with the US avg then most people who advocate for public transport. That is the sad reality.
And the problem in the US road system or the US bad public transport can 100% not be blamed on him. There are many people with far more responsibility that deserve to be shit on far more.
accidents per mile driven isn't a good metric. If you half the number of miles driven (through public transit/walking/biking) and have 50% more accidents per mile driven then you've still saved a bunch of people's lives.
Not GP, but the underlying problem to me is repeat zero-sum interactions in a crony capitalism based system. All the shit stems from the winners that emerge from these repeated games making short term decisions to either allow them to exit or play again.
I spent 6 months driving a 2019 Hyundai with lane assist and radar cruise control. Personally it's almost perfect. If you added some smart road features - to improve lane assist and better sign readability (to drop the speed as I enter town or curve and increase when I leave one). I don't need full self driving, just an improvement on speed control and some lane assist.
Would smart roads be expensive? RFID responders seem super cheap compared to how much actual asphalt costs. Authorities are currently unable to remotely control flows, speeds and safety which is completely bonkers.
> Would smart roads be expensive? RFID responders seem super cheap compared to how much actual asphalt costs. Authorities are currently unable to remotely control flows, speeds and safety which is completely bonkers.
Yeah its not actually that easy. Go and look into train signaling. And cars are not even able to do coupling.
Making a train system operate like a super-railway with cars is crazy difficult and has never been done before.
> That said, I think even just basic machine-readable road metadata like signs, speed limits, lanes etc would improve matters.
That assumes that sign reading is the primary issue with AI cars and it isn't.
If you are willing to go to that expense and infrastructure investment, why not just build thing like trains (trams, subways, S-Bahn) and things like Trolley buses and so on.
They are far more space efficient, you get much higher threwput then your untested fancy road infrastructure.
Another issue with your solution is that there are 1 billion cars out there that do not have any of these things so after decades of working on it, you will still have decades where most cars don't use any of this stuff.
I'm not saying that it is a primary issue, but all of these issues come up with Teslas occasionally - just read some of the comments here.
More importantly, even if this kind of stuff doesn't enable full self-driving, it can still be used to make basic driving assistance much better - think of even the most basic cars having lane assist on most road, for example.
I don't think it's a significant expense, either. I'm not talking about ripping up roads and replacing all signs in one fell swoop. But e.g. when they redid the highway near me recently, it got little reflectors to mark the lanes. What if, say, every Nth of those had some kind of RFID thing in it? Or when they replace signs, why not put up a new one that broadcasts what it is? None of this is particularly expensive.
As to why not trains and buses - because they do not actually offer the same features as cars, and many people want those features. If you want to argue against cars on principle, it's a different conversation entirely. I do think that more quality public transport is desperately needed, but, speaking as someone who used it nearly exclusively for the first 25 years of my life (and in places that are designed around it, unlike US), it does not replace a car.
We’re not talking about road safety. Road safety relies on a baseline amount of driving safety (car hardware + car software). If a car doesn’t stop at a stop sign, no amount of “safe roads” prevent it from mowing down a cyclist.
Saying this is like saying we should ignore sex offenders in favor of reworking our society. Instead, we should solve both problems rather than bicker over priorities.
Ok we should be solving both problems, I agree, but the reality is agencies and political processes have limited capacity. And how much society and influential people in that society is consumed with one thing, the less they do about the other.
And I would say that I'm not proposing to rework society in sociological sense, but rather to throw out the standard engineering standards and replace them with better standards.
And actually it does matter even in the case you suggest. If a car rolls into an intersection, what speed that car is matters. It matters from what points it is clear that the car is out of control. It matters if there are speed bumps or something along those lines that can send strong signals to a driver.
If you have all raised intersection then the top speed of cars will simply be lower and if somebody human or AI makes a mistake that lead to a crash, that crash will be at far lower speed.
And proper road design also leads to less intelligence and fancy car design being required. I rather get hit by a shitty designed unsafe old car at 20mph then a fancy new car at 30mph.
Lets be real here, automated driving or not, having actually save roads helps prevent death and harm in many cases.
The hyper focus on high tech software by all the agencies engaged in 'automotive security' is totally wrongly focused. What they should actually do is point out how insanely unsafe and broken the infrastructure is, specially for people outside of cars.
See: "Confessions of a Recovering Engineer: Transportation for a Strong Town"