Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Slate Op-Ed

This site may earn commission on affiliate links.
Slate has a new article up demanding that the Biden Administration impose more regulatory discipline on FSD and AP.

The Biden administration should crack down on Tesla. (slate.com)

I don't think regulators should single out Tesla. But I do think that consistent federal standards on FSD would be a good idea. And these standards should apply to all companies. Consistent standards would only help the industry develop better FSD because companies would know what to expect and what to strive for. For example, it would be good to have a coherent standard of what level of safety FSD needs to achieve. And the consumer would benefit as well since the consumer would get better info on what a particular FSD system is capable of. Consistent federal standards would avoid FSD being a "wild wild west" where your L2 is different than my L2 or my L4 is not as safe as your L4 or your L5 is not really L5.

I do wish Tesla implemented a camera based driver monitoring system. I think it would help make driver attention more reliable and thus help prevent accidents from drivers not paying attention.

Lastly, if Tesla released FSD to every car in the US and removed all the nags, letting people stop paying attention, and we got reports of many accidents every day because FSD was not ready, then yeah, I think regulators should shut down Tesla's FSD. But I don't think Tesla would do that. The truth is that Tesla does seem to be taking a methodical and safe approach to releasing FSD by deploying FSD Beta to small number of cars first and expanding slowly to more cars as the software improves. I suspect Tesla will continue to roll out FSD Beta very slowly, making sure it is safe enough first. Tesla does not want to risk getting shut down if there were a ton of accidents.

We should wait and see what FSD does when it is released wide to a large number of cars. If there are accidents on FSD, they should be investigated of course to see if FSD was responsible or not.
 
Last edited:
I think Tesla should be able to pursue their strategy (even though I don't think it will work). The onus should be on them to prove that FSD, beta or not, actually improves safety. They should not release FSD to everyone until they can do that. I would hope that they already have a plan on how to evaluate that internally. They should make it public and have a 3rd party evaluate the data as well.
 
  • Like
Reactions: diplomat33
What a sloppy article. First, they quote a 'concerned' group:

An association representing truckers warned that “while [FSD] may be a fun experiment for Tesla’s customers, public roads are our members’ workplace.”

Hmmmmm .... what would a truckers' association have against autonomous vehicles? It must be because they're dangerous right? Not because they're going to take jobs away from its members.

Also, can they stop using photos of pre-facelift (AP1) cars for these articles?
 
I don't think regulators should single out Tesla. But I do think that consistent federal standards on FSD would be a good idea. And these standards should apply to all companies. Consistent standards would only help the industry develop better FSD because companies would know what to expect and what to strive for. For example, it would be good to have a coherent standard of what level of safety FSD needs to achieve. And the consumer would benefit as well since the consumer would get better info on what a particular FSD system is capable of. Consistent federal standards would avoid FSD being a "wild wild west" where your L2 is different than my L2 or my L4 is not as safe as your L4 or your L5 is not really L5.

I do wish Tesla implemented a camera based driver monitoring system. I think it would help make driver attention more reliable and thus help prevent accidents from drivers not paying attention.

Lastly, if Tesla released FSD to every car in the US and removed all the nags, letting people stop paying attention, and we got reports of many accidents every day because FSD was not ready, then yeah, I think regulators should shut down Tesla's FSD. But I don't think Tesla would do that. The truth is that Tesla does seem to be taking a methodical and safe approach to releasing FSD by deploying FSD Beta to small number of cars first and expanding slowly to more cars as the software improves. I suspect Tesla will continue to roll out FSD Beta very slowly, making sure it is safe enough first. Tesla does not want to risk getting shut down if there were a ton of accidents.

We should wait and see what FSD does when it is released wide to a large number of cars. If there are accidents on FSD, they should be investigated of course to see if FSD was responsible or not.
How is driver attention monitored in any other car sold? How is it monitored in a Chevy Malibu or a BMW 3 series or ( you take your pick)? We've all seen the woman next to us putting on her make up going to work or the guy on his cell phone. How is their attention monitored? Thousands of accidents occur every day due to distracted driving, acting as if it's a Tesla only problem does not solve the problem.
 
  • Like
Reactions: 365gtb4 and DanCar
How is driver attention monitored in any other car sold? How is it monitored in a Chevy Malibu or a BMW 3 series or ( you take your pick)? We've all seen the woman next to us putting on her make up going to work or the guy on his cell phone. How is their attention monitored? Thousands of accidents occur every day due to distracted driving, acting as if it's a Tesla only problem does not solve the problem.

Those cars don't monitor the driver because the driver is driving manually.

It is not just a Tesla problem. All advanced L2 systems should have driver attention. That's because advanced L2 only work if the driver is paying attention.
 
Last edited:
How is driver attention monitored in any other car sold? How is it monitored in a Chevy Malibu or a BMW 3 series or ( you take your pick)? We've all seen the woman next to us putting on her make up going to work or the guy on his cell phone. How is their attention monitored? Thousands of accidents occur every day due to distracted driving, acting as if it's a Tesla only problem does not solve the problem.

GREAT point. I would rather have an L2 self driving vehicle with an inattentive driver than a 'dumb' vehicle with an equally inattentive driver. The *vast* majority of vehicle accidents today are the result of distracted drivers. Driver attention monitoring should be mandatory in all cars!
 
I would rather have an L2 self driving vehicle with an inattentive driver than a 'dumb' vehicle with an equally inattentive driver.

L2 is not self driving.

Second, the big difference is that L2 gives you the illusion that you don't need to pay attention. With L2, it is not always apparent that you need to pay attention until it is too late. But if you don't pay attention when driving manually, it is definitely more obvious that it is unsafe.
 
L2 is not self driving.

Second, the big difference is that L2 gives you the illusion that you don't need to pay attention. With L2, it is not always apparent that you need to pay attention until it is too late. But if you don't pay attention when driving manually, it is definitely more obvious that it is unsafe.

I definitely take your point on this. I guess my question is - what would the argument be *against* making driver monitoring systems mandatory in all vehicles, not just those with driver assist/FSD?

On the one hand, you can argue that people are more likely to get distracted in a vehicle with driver assist/FSD. On the other hand, you can argue that the risk of an inattentive driver is MUCH higher with a regular car, which will not even attempt to prevent an accident.
 
  • Like
Reactions: 82bert and drift
I definitely take your point on this. I guess my question is - what would the argument be *against* making driver monitoring systems mandatory in all vehicles, not just those with driver assist/FSD?
Driver monitoring will be required on all vehicles in the EU by 2022. Council of the European Union Adopts New Regulations to Significantly Reduce Road Casualties
Plenty of cars in the US have driver monitoring system now but you can turn them off. In the US our politics are different so I'd expect it will only be required when using advanced level 2 systems.
On the one hand, you can argue that people are more likely to get distracted in a vehicle with driver assist/FSD. On the other hand, you can argue that the risk of an inattentive driver is MUCH higher with a regular car, which will not even attempt to prevent an accident.
Yep, one of those answers is the truth. It would be nice to know!
 
  • Informative
Reactions: AlanSubie4Life
The bigger point is that inattentive drivers without any assist features are certainly more dangerous than those with driver assist on, so why not monitor all drivers (assist features or not)? The reality is that overwhelmingly most accidents are due to human error and frequently the result of inattentive driving. Every day I see someone cruising down the interstate texting a friggin novella.
 
  • Like
Reactions: DanCar
certainly
On the one hand, you can argue that people are more likely to get distracted in a vehicle with driver assist/FSD. On the other hand, you can argue that the risk of an inattentive driver is MUCH higher with a regular car, which will not even attempt to prevent an accident.
I would also point out that a Tesla and almost all other cars today will take measures to prevent an accident even if you're not paying attention without any automation system enabled. I have no doubt that ADAS reduce accidents. I have my doubts about Autosteer+TACC... though I still like using it.
 
One problem I see is that a lot of academics want to design the perfect system that everyone needs to follow rather than just look at the data and see if a system is safer than humans. I think Tesla are doing this really well, they focus on proving their safety statistically rather than fulfilling some ASIL D standard. Rapidly improve their system and gather tons of data and at one point just release the statistics that their current version is 10x safer than average humans. That is what actually matters.
 
I think Tesla should be able to pursue their strategy (even though I don't think it will work). The onus should be on them to prove that FSD, beta or not, actually improves safety. They should not release FSD to everyone until they can do that. I would hope that they already have a plan on how to evaluate that internally. They should make it public and have a 3rd party evaluate the data as well.

The onus should be on them to prove that high acceleration and high maximum speed actually improves safety.

The onus should be on them to prove that cruise control actually improves safety.

The onus should be on them to prove that adaptive cruise control actually improves safety.

The onus should be on them to prove that automatic steering actually improves safety.

The only thing I can see being mandated is driver monitoring, and Tesla's working on it.

I'm sure that Tesla understands that the NHTSA understands that there is enough of a problem with driver distraction and fatigue, with most vehicles _not_ having driver assistance, that the biggest thing needed is better driver monitoring in _every_ vehicle.

Because they _want_ companies to add more driver assistance to reduce crashes, and don't have the balls to mandate driver monitoring in all vehicles, I expect stronger driver monitoring eventually to be mandated in any vehicle with advanced driver assistance, with the hope that all vehicles end up with advanced driver assistance anyway,
 
What a sloppy article. First, they quote a 'concerned' group:

An association representing truckers warned that “while [FSD] may be a fun experiment for Tesla’s customers, public roads are our members’ workplace.”

Hmmmmm .... what would a truckers' association have against autonomous vehicles? It must be because they're dangerous right? Not because they're going to take jobs away from its members.

Also, can they stop using photos of pre-facelift (AP1) cars for these articles?

What they might also have against it, is that two careless Tesla drivers have been killed in incidents where truckers were making illegal left turns across traffic.
 
I definitely take your point on this. I guess my question is - what would the argument be *against* making driver monitoring systems mandatory in all vehicles, not just those with driver assist/FSD?

It's more about what driving mode the car is in. Cars that are in fully autonomous driving mode don't need any driver attention. But I agree that there should be a driver attention system when driving manually or when driving with a ADAS.

So all cars that are not fully autonomous or that have the option to drive manually should have a driver attention system. Basically, all cars that have a steering wheel and pedals should be required to have a driver attention system.

If you have a robotaxi with no steering wheel or pedals, the very act of removing the controls, is implying that you don't need driver attention since.

On the one hand, you can argue that people are more likely to get distracted in a vehicle with driver assist/FSD. On the other hand, you can argue that the risk of an inattentive driver is MUCH higher with a regular car, which will not even attempt to prevent an accident.

Different driving modes need driver attention but maybe for different reasons.

Manual driving should require a driver attention system to make sure the driver is engaged since the driver is doing 100% of the driving. The risk of accident is highest if the driver is not paying attention.

With basic L2, it is only doing some basic tasks like lane keeping and cruise control. It is only automating some tasks. It is not actually driving or responding to the road. It can prevent some accidents in some cases but the system cannot respond to many situations. You still need an alert driver who can respond to all those situations it cannot handle. So you need a driver monitoring system to make sure the driver is alert.

With advanced L2, it can do more steering and braking tasks but the system is not designed to actually drive on its own. Yes, it may prevent some accidents but it cannot respond to many situations. You need an alert driver to respond to situations it cannot handle. And this system starts to get close to almost autonomous in some instances that the driver may get overconfident and get lulled into a false sense of security. So monitoring the driver attention is very important to prevent the driver from not paying attention when the system actually needs them the most.

With FSD Beta, the system is designed to do all the driving but it is not reliable yet. You need a "safety driver" to intervene when the system might make a mistake. Yes, it may prevent some accidents but it might cause accidents too. You don't know when those failures might happen so you definitely want to pay attention at all times. So a driver attention system is important, especially since the driver won't think they need to pay attention.

Lastly, with a fully validated FSD system, the car is able to do all the driving on its own in a given ODD to an acceptable level of safety. So no driver attention system is required.