> Reading a stop sign and deciding whether to stop is a City-Streets function, and that’s FSD. Basic Autopilot is built as a highway lane-keeping system, and it does not respond to stop signs. So a driver-assist car verified engaged as it ran a stop sign was, most likely, running FSD.
This just isn't true, "Traffic Light and Stop Sign Control" is a feature of basic autopilot on Model 3s.
It differs based on when you purchased the car and package. I have “Advanced Autopilot” (which is no longer for sale in my market) that does not stop for stop signs or traffic lights.
Earlier this year they stopped selling Autopilot altogether in the US and Canada.
Too much variation to make a call on whether their version sips have been stopping under these conditions.
My 2024 Model Y has "Autopilot" mode that does not stop for stop signs. It just keeps the car in the lane and maintains distance from the car in front.
The article is pretty clickbait-y, acknowledging it is unknown whether or not the car was using FSD or Autopilot, and then going on to speculate it probably was FSD, and then again saying "That's not a certainty".
>Because Tesla redacted the software version, we can’t say for certain whether this was basic Autopilot or “Full Self-Driving.”
I would like to know why Tesla (or any automaker) is allowed to redact this information.
Most US regulatory agencies give businesses broad latitude to self-redact their trade secrets in public filings, because otherwise they would have to take time away from their regulatory mission to engage in a bunch of debates about what is or is not a trade secret. The versions would be unredacted in the private reports given to the NHTSA.
People who think the software version isn't a valid trade secret can still ask for public disclosure, and the Washington Post has done so (https://electrek.co/2025/06/05/tesla-admits-suffer-financial...). If Tesla weren't allowed to self-redact, then the NHTSA wouldn't be able to see these reports at all until the courts settle this issue.
I don't know anything about this specific case but this exact thing will be a major barrier for autonomous vehicles to overcome. Instead of benchmarking AVs against the current, human-caused number of vehicular deaths - people will naturally benchmark AVs to 0 and treat every instance of death as a failure of AVs, even if the reality is that AVs end up 99% safer than human controlled vehicles.
As it should be. “Safer on average” is not a defense for failing something as basic as stopping at a stop sign. The bar for not running one and killing someone is not unreasonably high.
Safer than average is totally defensible, because every year the margin of what is "average" changes.
I'm just pulling numbers out of my ass here, but if say, one person is killed by an automobile every 100m miles, if the next year the average is 0.999 per 100m miles, that's a good thing provided the trend continues. We should not step over a dime to pick up a penny.
We do tend to impose them on individual drivers who make egregious and consequential failures, that’s the point. If every car running the same software is functionally the same ‘driver’ this class of failure absolutely should trigger a moratorium.
>We do tend to impose them on individual drivers who make egregious and consequential failures, that’s the point.
Do we? It's a common saying that if you want to kill someone with no penalties, have them run over with a car.
Then there are examples like these all over the place. Not to mention distracted driving and DUIs don't even get your license suspended in many or most jurisdictions.
>The Audi, which Jones had bought one month before the crash, was the third car he had totaled in a crash within 11 months. Speed was a factor in all three collisions, but police did not cite Jones in the first two crashes, prosecutors said.
Also see the 80-year-old woman in SF who killed an entire family at a bus stop, while speeding, while going the wrong way on a one way street, and was sentenced to probation and community service and banned from driving for three years.
No, because you can punish bad drivers, remove them from the driving pool, put them in jail, or in some way achieve justice (this may not always happen, but at least it's posible). There is no way for FSD to have any accountability, and so no justice can be served. "Sorry, your son was just unlucky that he was killed by the AI, but we can't do anything about it" is an intolerable breakdown of justice. Same problem we have with qualified immunity.
"Sorry, the driver could have turned on FSD, but decided to drive [drunk/tired/distracted] and killed your son. Thankfully we can ban them from driving for a while."
This could play out a million times at the population level. If you can decrease the chances of needing justice in the first place, shouldn't that be considered?
But you can take their license, fine them, jail them.
Corporations just pay a fine and move on. I don’t want that to be the standard for road accidents and frankly outside of this techie niche no one does.
It should be a critical part of the evaluation. If you set the bar higher than “safer on average” then you’re going to delay the switch, and every month of delay will cost more lives. “Safer on average” is exactly when you should start using the new thing, even if it’s uncomfortable and even if it’s a low bar.
No it shouldn’t. Low risk behaviors should equal low risk outcomes. If it’s safer on average yet causes low risk behaviors to suddenly become unpredictably dangerous, folks will very reasonably push back
Often we do not. The Freakonomics podcast had an episode titled The Perfect Crime where they basically say if you want to kill somebody, hit them with your car because that’s often not punished beyond a ticket.
>According to evidence presented from law enforcement at trial, on September 30, 2024, Burke was operating his patrol vehicle at a high rate of speed without emergency lights or siren activated when it crashed with another vehicle. Burke's partner was injured in the crash, and 34-year-old Cedric Hayden Jr. and 33-year-old Dejuan Pettis were killed.
Not sure why you are getting downvoted here though yes, the system is not perfect, in some states more so than in others. it is a fine line between giving someone who deserves it a 2nd chance vs. someone with 17 priors. but guessing you would also agree that allowing "self-driving" tech to kill people on the road without any repercussions isn't all that ideal either?
You can't compare autonomous systems that drive hundreds of thousands of vehicles to a single person. You have to compare it to hundreds of thousands of people.
American drivers kill ~100 people every day behind the wheel of a car. If autonomy were 50% safer and was deployed at such scale that it replaced 50% of all hours driven it would save 25 lives every day. It would also kill 25 people every day.
Sadly, many people will clutch their pearls at inevitable headlines like Bloodthirsty Elon Sacrifices 25 Innocents on the Altar of Profit Every Day! and, their emotions sufficiently primed, they will happily pull the plug on autonomy, sending 25 people per day to their deaths and feeling great about themselves as they do it.
Ok, in my country, you actually get your license removed and jail time if you kill someone (usually 2 years, up to 10 if you're drinking).
Let's say Tesla kill 50% less than a human driver, I'd agree to give Tesla's board 50% of the sentence for each person autopilot kills. As an incentive to make the cars kill less.
Imho part of isuse here is that you cannot "partial AV" it.
You cannot have a car that will EVER need you to intervene or else it does not work because of awful human nature. The amount of tesla uber drivers ive had literally browsing tiktok, watching youtube UNTIL the car beeps that it is a green light detected, then they immediately hit the accelerator before looking up is ridiculous. You cannot have a half ass "autopilot" driving experience where you are expected to zone out and then be expected to take control within a millisecond during uncertainty. It does not help that FSD, AP, aAP, bAP, the model, whether the car is visual only all apply differently
Another huge problem is that error rate it not error shape. We have an enormous body of knowledge and process for guessing what might cause problems for human drivers, how to deter the problems, and how to mitigate the damage.
For example, you'll have a pretty good idea how our fellow humans will react to rumble-strips if I describe them or play a recording. I don't need to consult a datasheet or design an AI test suite with a zillion simulations and a bunch of closed-course tests.
What percentage of road accidents involve broken down old cars, drugs and alcohol, irresponsible and risky behavior?
Because most of us don’t have that problem, and insurance stats suggest we’re a hell of a lot safer drivers than a hypothetical aggregate including stolen car joyrides by teens and 50 year old teetotalers in a well maintained, modern car.
Or, and hear me out, a crash has to occur before an event is logged as a crash. To be clear, I haven’t even read the article yet so I don’t know any of the details, but that’s not a conclusion we can draw without understanding how the data is logged.
I've had a pretty wild ride from being hit at probably less than that speed.
I'm sure you've seen video of the police performing a PIT maneuver. That's basically what happened to me--tapped by a car starting from a stop sign. They got a scrape, I got totaled. And, yes, it works as advertised--totally confusing, my foot somehow found the brake but I have no recall of doing so.
Tesla now has ~40 level 4 automated taxis in operation with no pedals or steering wheel. When these have an accident they can't blame a non-existent driver for mashing a non-existent accelerator pedal. And they are their own insurance company. A few fatalities in the Cybercab could squelch the whole rollout, even if the stats show it's safer than human drivers.
As a human driver with declining capabilities who will need level 4 autopilot before long to stay independent, I'm rooting for FSD.
If they hadn't scammed millions into paying for the FSD scam starting Oct 2016 (see their "Driver is just there for legal reasons" video) - then there would be no headlines for a driver assist system occasionally failing like this.
> Reading a stop sign and deciding whether to stop is a City-Streets function, and that’s FSD. Basic Autopilot is built as a highway lane-keeping system, and it does not respond to stop signs. So a driver-assist car verified engaged as it ran a stop sign was, most likely, running FSD.
I don't understand this part. My understanding is the basic Autopilot is very similar to the what on other cars would be adaptive cruise control plus lane finding/lane keeping.
If so, isn't the expected behavior that if you do not manually intervene at stop lights and stop signs it will remain engaged and drive right on through?
Yes, that's what autopilot will do. Even though I knew that with my tesla, I thought it could be confusing because the visualization does show traffic lights and stop signs and other cars/pedestrians. So a driver might think that if the computer obviously knows about the stop sign, it would stop there.
I can't help but think we need a more extensive "Flight Data Recorder" sort of object for, at least, self-driving cars if not all of them. Event Data Recorder's are coming and mandated for 2021+ but I think they need to go further, IMO. Police in my area already do not bother responding, and insurance just does 50/50 unless theirs dash cams. Why not just make them integrated and mandatory along with other telemetry?
Humans can go to prison [1], and depending on the state there are specific vehicular homicide laws or they are could be charged under normal homicide laws, specific statutes exist because vehicular laws can have higher punishment than typical negligence/reckless standards.
In California the charge can be anything from vehicular manslaughter to second degree murder, that means 15 years-to-life without parole, depending on other factors such prior(murder) convictions or if the victim is a peace officer and so forth.
In comparison self-driving systems do not have any punishment mechanism in the penal code. There is not even a well established liability standard, or ban/suspension threshold or any regulatory body equipped to audit a self driving system specifically [2]
[1] if found guilty - we may disagree on the mechanisms, effectiveness ,conviction rate and other details, nonetheless there is enforcement and there are going people in prison for this type of crime everyday.
[2] yes NTSB can do investigations but only for accidents (and give recommendations) and also has just 400 employees and $150M budget to cover everything from aviation to marine or pipeline and rail accidents.
But they don't. Thousands of people are being killed at intersections, yet unless there is some proven intent, it's extremely rare to go to jail.
Maybe this might be valid if this was a realistic thing and every bad driver who pulls out of a side street and kills someone suffers serious consequences, but they don't and it's not.
You are talking about enforcement. I am saying there are no AI laws on books, there is no established liability for the self driving software developer, the liability is still to the driver solely and completely.
First lets have those laws and regulations for AI and we can THEN talk about conviction rates for humans(or agents) and effectiveness of enforcement.
Right now human and agents are in two different leagues of enforcement civil and criminal, so they cannot be compared and agents should be held to much higher standard than humans until that changes.
self driving is safer than human driving, why should we apply urgent attention to AI driver safety. Human drivers are killing hundreds of people daily.
Oh good. I may be put at unnecessary risk because people have higher standards for automation, but at least the guy who killed me will go to jail, maybe.
If safety is the objective we should be prioritizing public transit(automated even better) and not discussing agent/human driving a personal vehicle, here we value freedom[1] over safety, my comment is only a reflection of that ethos.
[1]Whether we like it or not individual freedom reflects in everything from gun safety(unique in the developed world) to the state of public transport in the richest country in the world.
It’s not whataboutism when you’re comparing the exact same actions and the two things can be substituted for each other.
Whataboutism is “how dare you complain about our gulags when you treat your black people like slaves.” It is not “our system kills people but it kills fewer people than what we currently have.” Especially when killing people is a result of being on the bleeding edge of what’s possible rather than being a deliberate choice to hurt people.
I think, pragmatically, we have to accept that death is the outcome for us all, and trending towards fewer avoidable deaths and longer fuller lives is approximately the aim. Nonetheless deaths from a technology is inevitable.
I believe what you're trying to talk about are excess deaths when a technology is less capable than the incumbent? I'm not sure the stats, but I'd suspect FSD is already superior to the often distracted, sometimes drunk, phone in hand driver that is the norm today.
If chatgpt is to be trusted, Waymo and Tesla FSD are already massively safer than human drivers.
“[REDACTED, MAY CONTAIN CONFIDENTIAL BUSINESS INFORMATION]”
So any company has the option to redact information if they feel like it. Tesla has chosen to provide as little information as possible, claiming business secrecy, and no one can compel them otherwise.
It's because Musk fired his PR department. Tesla have no ability to damage control stories like this because it's no one's job at the company.
A normal firm would have a heads up warning when requests for comment hit their media inboxes and internally could start collecting data and developing a response: thanks to Elon they don't. The story hits first and goes completely unchallenged until investigators get involved.
Regardless of what you think of Tesla or Musk, they repeatedly kick own goals on basic product brand management.
How the hell is Tesla (or any other car co) allowed to redact fields in a safety report as “confidential business information"? and how the hell are they still allowed to call this "FULL SELF DRIVING"
I teach Tesla FSD as a cautionary tale of malicious business practices, how to deploy robots in the least safe way possible, and an anti-example of how to be a good engineer. It should be a crime to have as little a grasp on engineering and to have as much input as Musk does.
As an avid fsd user, your list makes me more comfortable using fsd than not. If there were more you’d parrot it around, but in the end, fsd is safer than humans at the wheel. As much as you dislike it because of Musk, fsd will continue to march forward and improve and millions will use it. You are beta testing fsd even if you don’t want to be.
> You are beta testing fsd even if you don’t want to be.
Exactly, that's precisely the wrong way to do ethical, responsible engineering. It's one thing to put your own life in the hands of a beta test, but it's another to put all of our lives at risk without our consent.
Distinguish a beta test from a learner driver or first year licensed driver. Why should I share the road with anyone who isn’t an experienced and competent driver?
Consent. We've consented to people learning to drive. Everyone is aware of those risks and they act accordingly. Often, to make other drivers aware of the driver's learning status, it's customary to affix a sticker to the car.
On the other hand, the public generally is not aware that these FSD machines are operating without critical safety equipment, and if they knew and were properly informed I'm not sure they would agree to be part of the test. I think Elon Musk knows that as well which is why he hasn't bothered to ask.
Moreover, I'm shocked and horrified by the number of people going around "beta testing" new versions of FSD. You'll see someone go to an area where a previous FSD version is known to fail just to see if the new one works. How is that a good idea at all? The last time you were here the car lurched into oncoming traffic unexpectedly and went down a one way, so let's see if it does it again, fingers crossed?
If we're using that definition of consent, it is an absolute unequivocal red herring. Roads are full of vehicles and full of drivers which, if you understood exactly what and who they are, you would absolutely refuse to consent sharing the road with. I didn’t consent to share the road with someone who failed to tighten their wheel nuts. I didn’t consent to share the road with that tourist in a rental car. I didn’t consent to sharing the road with a deer. I didn’t consent to sharing the road with Mercedes Drive Pilot or Ford BlueCruise or Mercedes SuperCruise or Nissan ProPILOT or any vehicles operated by Zoox or Waymo.
The dangerous you're talking about have been around for a long time, have been depicted in many forms over decades in all kinds of popular media media, have actually impacted people which has birthed reforms and new safety measures, and that is all part of the constant education and training the public receives.
None of that is done for FSD or tesla cars. People don't know how they work, what the dangers are, what their failure modes are, how to approach them, and no one is telling them!
If you want to lament a general lack of awareness and progressive education, then I'm fully on board. Perhaps we should require operators to pass an additional driver test to be legally permitted to supervise any L2+ system from any manufacturer. No BlueCruise or SuperCruise unless you've got the extra marking on your license.
But as far as "consent" is concerned, if you wish to use the public roads, consent necessarily means accepting the roads as a commons. This means, at minimum, accepting the presence of anything the relevant government authority deems legal.
I'm not really talking about formal education here.
> This means, at minimum, accepting the presence of anything the relevant government authority deems legal.
All that means if you have enough money you can put whatever death machines on the road you want. I don't think that's a good way to manage a road network, and will not ultimately lead to greater safety for all.
We do not generally hold drivers responsible for slips in judgement when driving. Only in cases of extreme recklessness is it normal for a driver to be 'held responsible'.
It is silly that we hold self driving cars to an astronomically higher standard. We all know they are safer than human drivers. And that is true even if they kill 50 people a day.
You can disagree, and I do too, but it is the basis of an interesting discussion.
Personally I think that fully autonomous vehicles need to be at minimum 100x safer than humans in order to maintain a social license to operate. I’m not saying the law needs to codify that. I’m saying that society could never accept a system which was merely 10x safer than humans. Imagine every vehicle in North America was replaced with a robotaxi and road deaths absolutely plummeted from 40,000 to just 1,000 per year. That’s still an article every day about “Elon Musk's river of blood”.
Supervised systems do not need such extreme safety, because the human driver remains responsible. But here’s an idea: if you want to use supervised FSD, that’s fine, but every collision gets automatically uploaded to a repository that is searchable by law enforcement and insurance companies.
> Reading a stop sign and deciding whether to stop is a City-Streets function, and that’s FSD. Basic Autopilot is built as a highway lane-keeping system, and it does not respond to stop signs. So a driver-assist car verified engaged as it ran a stop sign was, most likely, running FSD.
This just isn't true, "Traffic Light and Stop Sign Control" is a feature of basic autopilot on Model 3s.
https://www.tesla.com/ownersmanual/2017_2023_model3/en_eu/GU...
It differs based on when you purchased the car and package. I have “Advanced Autopilot” (which is no longer for sale in my market) that does not stop for stop signs or traffic lights.
Earlier this year they stopped selling Autopilot altogether in the US and Canada.
Too much variation to make a call on whether their version sips have been stopping under these conditions.
Well that's news to me because my model 3 in autopilot definitely does not stop for traffic lights or stop signs.
My 2024 Model Y has "Autopilot" mode that does not stop for stop signs. It just keeps the car in the lane and maintains distance from the car in front.
The article is pretty clickbait-y, acknowledging it is unknown whether or not the car was using FSD or Autopilot, and then going on to speculate it probably was FSD, and then again saying "That's not a certainty".
>Because Tesla redacted the software version, we can’t say for certain whether this was basic Autopilot or “Full Self-Driving.”
I would like to know why Tesla (or any automaker) is allowed to redact this information.
Most US regulatory agencies give businesses broad latitude to self-redact their trade secrets in public filings, because otherwise they would have to take time away from their regulatory mission to engage in a bunch of debates about what is or is not a trade secret. The versions would be unredacted in the private reports given to the NHTSA.
People who think the software version isn't a valid trade secret can still ask for public disclosure, and the Washington Post has done so (https://electrek.co/2025/06/05/tesla-admits-suffer-financial...). If Tesla weren't allowed to self-redact, then the NHTSA wouldn't be able to see these reports at all until the courts settle this issue.
> I would like to know why Tesla (or any automaker) is allowed to redact this information.
That's an easy one, donate enough money to elected officials who would be the only ones that could prevent this from happening :)
I don't know anything about this specific case but this exact thing will be a major barrier for autonomous vehicles to overcome. Instead of benchmarking AVs against the current, human-caused number of vehicular deaths - people will naturally benchmark AVs to 0 and treat every instance of death as a failure of AVs, even if the reality is that AVs end up 99% safer than human controlled vehicles.
As it should be. “Safer on average” is not a defense for failing something as basic as stopping at a stop sign. The bar for not running one and killing someone is not unreasonably high.
Safer than average is totally defensible, because every year the margin of what is "average" changes.
I'm just pulling numbers out of my ass here, but if say, one person is killed by an automobile every 100m miles, if the next year the average is 0.999 per 100m miles, that's a good thing provided the trend continues. We should not step over a dime to pick up a penny.
I agree; that’s why I think we need a moratorium on human drivers while we figure this all out.
We do tend to impose them on individual drivers who make egregious and consequential failures, that’s the point. If every car running the same software is functionally the same ‘driver’ this class of failure absolutely should trigger a moratorium.
I think this is a fantastic distinction to make, and one I haven’t seen before.
>We do tend to impose them on individual drivers who make egregious and consequential failures, that’s the point.
Do we? It's a common saying that if you want to kill someone with no penalties, have them run over with a car.
Then there are examples like these all over the place. Not to mention distracted driving and DUIs don't even get your license suspended in many or most jurisdictions.
https://www.seattletimes.com/seattle-news/law-justice/teen-w...
>The Audi, which Jones had bought one month before the crash, was the third car he had totaled in a crash within 11 months. Speed was a factor in all three collisions, but police did not cite Jones in the first two crashes, prosecutors said.
Also see the 80-year-old woman in SF who killed an entire family at a bus stop, while speeding, while going the wrong way on a one way street, and was sentenced to probation and community service and banned from driving for three years.
What is this Tesla (or its non-driver) going to get?
more retail investors buying the stock
This also solves the rising oil and gas prices issue as well!
It should be a defense for FSD to exist though, right? Net more alive people.
No, because you can punish bad drivers, remove them from the driving pool, put them in jail, or in some way achieve justice (this may not always happen, but at least it's posible). There is no way for FSD to have any accountability, and so no justice can be served. "Sorry, your son was just unlucky that he was killed by the AI, but we can't do anything about it" is an intolerable breakdown of justice. Same problem we have with qualified immunity.
"Sorry, the driver could have turned on FSD, but decided to drive [drunk/tired/distracted] and killed your son. Thankfully we can ban them from driving for a while."
This could play out a million times at the population level. If you can decrease the chances of needing justice in the first place, shouldn't that be considered?
What you're talking about is who to blame, not how to keep more people alive.
You can update software though, you can’t update a negligent driver.
But you can take their license, fine them, jail them.
Corporations just pay a fine and move on. I don’t want that to be the standard for road accidents and frankly outside of this techie niche no one does.
It should be a critical part of the evaluation. If you set the bar higher than “safer on average” then you’re going to delay the switch, and every month of delay will cost more lives. “Safer on average” is exactly when you should start using the new thing, even if it’s uncomfortable and even if it’s a low bar.
No it shouldn’t. Low risk behaviors should equal low risk outcomes. If it’s safer on average yet causes low risk behaviors to suddenly become unpredictably dangerous, folks will very reasonably push back
Sorry, I’m not following. What are the low risk behaviors you’re referring to and what makes them suddenly become unpredictably dangerous?
Walking / driving across an intersection where the cross traffic has a stop sign?
That's unpredictably dangerous today, with human drivers.
Humans fail to stop at stop signs all the time. Their failure to stop kill people as well.
And we take away people's ability to drive when they kill people.
Often we do not. The Freakonomics podcast had an episode titled The Perfect Crime where they basically say if you want to kill somebody, hit them with your car because that’s often not punished beyond a ticket.
This trial just happened where there was no punishment:
https://www.macombgov.org/node/55/news/former-warren-police-...
>According to evidence presented from law enforcement at trial, on September 30, 2024, Burke was operating his patrol vehicle at a high rate of speed without emergency lights or siren activated when it crashed with another vehicle. Burke's partner was injured in the crash, and 34-year-old Cedric Hayden Jr. and 33-year-old Dejuan Pettis were killed.
Then there was this one a couple years ago:
https://en.wikipedia.org/wiki/Killing_of_Jaahnavi_Kandula
https://www.theriolindanews.com/2025/04/22/529861/license-to...
https://calmatters.org/investigation/2025/04/license-to-kill...
It’s in the US. The “no punishment” might be related to the offender being a policeman. Amazing it went to trial, to be honest.
Not sure why you are getting downvoted here though yes, the system is not perfect, in some states more so than in others. it is a fine line between giving someone who deserves it a 2nd chance vs. someone with 17 priors. but guessing you would also agree that allowing "self-driving" tech to kill people on the road without any repercussions isn't all that ideal either?
You can't compare autonomous systems that drive hundreds of thousands of vehicles to a single person. You have to compare it to hundreds of thousands of people.
American drivers kill ~100 people every day behind the wheel of a car. If autonomy were 50% safer and was deployed at such scale that it replaced 50% of all hours driven it would save 25 lives every day. It would also kill 25 people every day.
Sadly, many people will clutch their pearls at inevitable headlines like Bloodthirsty Elon Sacrifices 25 Innocents on the Altar of Profit Every Day! and, their emotions sufficiently primed, they will happily pull the plug on autonomy, sending 25 people per day to their deaths and feeling great about themselves as they do it.
Ok, in my country, you actually get your license removed and jail time if you kill someone (usually 2 years, up to 10 if you're drinking).
Let's say Tesla kill 50% less than a human driver, I'd agree to give Tesla's board 50% of the sentence for each person autopilot kills. As an incentive to make the cars kill less.
Imho part of isuse here is that you cannot "partial AV" it.
You cannot have a car that will EVER need you to intervene or else it does not work because of awful human nature. The amount of tesla uber drivers ive had literally browsing tiktok, watching youtube UNTIL the car beeps that it is a green light detected, then they immediately hit the accelerator before looking up is ridiculous. You cannot have a half ass "autopilot" driving experience where you are expected to zone out and then be expected to take control within a millisecond during uncertainty. It does not help that FSD, AP, aAP, bAP, the model, whether the car is visual only all apply differently
Teslas do very badly when benchmarked against human caused deaths in similar (luxury) vehicles.
Another huge problem is that error rate it not error shape. We have an enormous body of knowledge and process for guessing what might cause problems for human drivers, how to deter the problems, and how to mitigate the damage.
For example, you'll have a pretty good idea how our fellow humans will react to rumble-strips if I describe them or play a recording. I don't need to consult a datasheet or design an AI test suite with a zillion simulations and a bunch of closed-course tests.
What percentage of road accidents involve broken down old cars, drugs and alcohol, irresponsible and risky behavior?
Because most of us don’t have that problem, and insurance stats suggest we’re a hell of a lot safer drivers than a hypothetical aggregate including stolen car joyrides by teens and 50 year old teetotalers in a well maintained, modern car.
This part is wild to me:
If true, how could the driver of the other car have been killed?And all 4 Tesla occupants were injured. Something doesn't add up.
It's almost like, and hear me out, Tesla's internal diagnostics might have been manipulated.
Or, and hear me out, a crash has to occur before an event is logged as a crash. To be clear, I haven’t even read the article yet so I don’t know any of the details, but that’s not a conclusion we can draw without understanding how the data is logged.
No one outside Germany would ever do that ! /s
2-way stop and the other guy hit the Tesla at 40-60 no seatbelt?
That would track with the injuries in the Tesla as well.
Gotta come up with the energy somehow. If the Tesla doesn't have it the other car needs to.
You raise a very good point. I didn't consider the other car.
I've had a pretty wild ride from being hit at probably less than that speed.
I'm sure you've seen video of the police performing a PIT maneuver. That's basically what happened to me--tapped by a car starting from a stop sign. They got a scrape, I got totaled. And, yes, it works as advertised--totally confusing, my foot somehow found the brake but I have no recall of doing so.
possible if it accelerated after crashing
Tesla now has ~40 level 4 automated taxis in operation with no pedals or steering wheel. When these have an accident they can't blame a non-existent driver for mashing a non-existent accelerator pedal. And they are their own insurance company. A few fatalities in the Cybercab could squelch the whole rollout, even if the stats show it's safer than human drivers.
As a human driver with declining capabilities who will need level 4 autopilot before long to stay independent, I'm rooting for FSD.
If they hadn't scammed millions into paying for the FSD scam starting Oct 2016 (see their "Driver is just there for legal reasons" video) - then there would be no headlines for a driver assist system occasionally failing like this.
Scam or not, "Robot kills human" was always going to be news.
> Reading a stop sign and deciding whether to stop is a City-Streets function, and that’s FSD. Basic Autopilot is built as a highway lane-keeping system, and it does not respond to stop signs. So a driver-assist car verified engaged as it ran a stop sign was, most likely, running FSD.
I don't understand this part. My understanding is the basic Autopilot is very similar to the what on other cars would be adaptive cruise control plus lane finding/lane keeping.
If so, isn't the expected behavior that if you do not manually intervene at stop lights and stop signs it will remain engaged and drive right on through?
> most likely, running FSD.
"most likely running FSD" so was it running FSD or are we speculating?
Yes, that's what autopilot will do. Even though I knew that with my tesla, I thought it could be confusing because the visualization does show traffic lights and stop signs and other cars/pedestrians. So a driver might think that if the computer obviously knows about the stop sign, it would stop there.
I can't help but think we need a more extensive "Flight Data Recorder" sort of object for, at least, self-driving cars if not all of them. Event Data Recorder's are coming and mandated for 2021+ but I think they need to go further, IMO. Police in my area already do not bother responding, and insurance just does 50/50 unless theirs dash cams. Why not just make them integrated and mandatory along with other telemetry?
Why didn't the driver stop and why does it say "a Tesla ran a stop sign" and not "a driver in a Tesla ran a stop sign" ?
Because media isn’t a charity, it’s a profit-seeking business model. Outrage is profitable.
Do you know how humans run stop lights and kill people, how does it compare?
Humans can go to prison [1], and depending on the state there are specific vehicular homicide laws or they are could be charged under normal homicide laws, specific statutes exist because vehicular laws can have higher punishment than typical negligence/reckless standards.
In California the charge can be anything from vehicular manslaughter to second degree murder, that means 15 years-to-life without parole, depending on other factors such prior(murder) convictions or if the victim is a peace officer and so forth.
In comparison self-driving systems do not have any punishment mechanism in the penal code. There is not even a well established liability standard, or ban/suspension threshold or any regulatory body equipped to audit a self driving system specifically [2]
[1] if found guilty - we may disagree on the mechanisms, effectiveness ,conviction rate and other details, nonetheless there is enforcement and there are going people in prison for this type of crime everyday.
[2] yes NTSB can do investigations but only for accidents (and give recommendations) and also has just 400 employees and $150M budget to cover everything from aviation to marine or pipeline and rail accidents.
But they don't. Thousands of people are being killed at intersections, yet unless there is some proven intent, it's extremely rare to go to jail.
Maybe this might be valid if this was a realistic thing and every bad driver who pulls out of a side street and kills someone suffers serious consequences, but they don't and it's not.
You are talking about enforcement. I am saying there are no AI laws on books, there is no established liability for the self driving software developer, the liability is still to the driver solely and completely.
First lets have those laws and regulations for AI and we can THEN talk about conviction rates for humans(or agents) and effectiveness of enforcement.
Right now human and agents are in two different leagues of enforcement civil and criminal, so they cannot be compared and agents should be held to much higher standard than humans until that changes.
self driving is safer than human driving, why should we apply urgent attention to AI driver safety. Human drivers are killing hundreds of people daily.
Oh good. I may be put at unnecessary risk because people have higher standards for automation, but at least the guy who killed me will go to jail, maybe.
If safety is the objective we should be prioritizing public transit(automated even better) and not discussing agent/human driving a personal vehicle, here we value freedom[1] over safety, my comment is only a reflection of that ethos.
[1]Whether we like it or not individual freedom reflects in everything from gun safety(unique in the developed world) to the state of public transport in the richest country in the world.
We could do both. We can make cars safer and push on public transit.
You're right, we should ban cars.
Their point, presumably, was that we shouldnt be anti-tech but we have to compare the statistics.
Their point is whataboutism to somehow defend autonomous deaths as acceptable.
It's not 'whataboutism'. It's an identical situation, and the statistics relate to the same events.
It’s not whataboutism when you’re comparing the exact same actions and the two things can be substituted for each other.
Whataboutism is “how dare you complain about our gulags when you treat your black people like slaves.” It is not “our system kills people but it kills fewer people than what we currently have.” Especially when killing people is a result of being on the bleeding edge of what’s possible rather than being a deliberate choice to hurt people.
I think, pragmatically, we have to accept that death is the outcome for us all, and trending towards fewer avoidable deaths and longer fuller lives is approximately the aim. Nonetheless deaths from a technology is inevitable.
I believe what you're trying to talk about are excess deaths when a technology is less capable than the incumbent? I'm not sure the stats, but I'd suspect FSD is already superior to the often distracted, sometimes drunk, phone in hand driver that is the norm today.
If chatgpt is to be trusted, Waymo and Tesla FSD are already massively safer than human drivers.
In many jurisdictions you can run people over and kill them and as long it was a normal lapse in judgement or mistake they cannot charge you.
Only if you intentionally drive recklessly, for example speeding and racing, is it easy to charge a driver for manslaughter.
At least we can hold them directly accountable by the law.
We hold the driver accountable.
AI slop to the max, what a bad article.
Now on the issue itself: Why doesn't Tesla report the version and whether it was FSD or Autopilot?
Because for some reason Tesla is exempt from reporting specifics about their safety reporting. Can’t possibly imagine why…
According to https://fuelarc.com/news-and-features/redacted-tesla-asked-f...
The redactions have the note:
“[REDACTED, MAY CONTAIN CONFIDENTIAL BUSINESS INFORMATION]”
So any company has the option to redact information if they feel like it. Tesla has chosen to provide as little information as possible, claiming business secrecy, and no one can compel them otherwise.
It's because Musk fired his PR department. Tesla have no ability to damage control stories like this because it's no one's job at the company.
A normal firm would have a heads up warning when requests for comment hit their media inboxes and internally could start collecting data and developing a response: thanks to Elon they don't. The story hits first and goes completely unchallenged until investigators get involved.
Regardless of what you think of Tesla or Musk, they repeatedly kick own goals on basic product brand management.
It's electrek, 95% chance any front page Tesla negative story is from them.
so they allow auto drive in non-approved areas?
How the hell is Tesla (or any other car co) allowed to redact fields in a safety report as “confidential business information"? and how the hell are they still allowed to call this "FULL SELF DRIVING"
Will be talking about this in my robotics class. Here's my current indictment of FSD, hasn't been updated in a year but I'll add this one to the list:
2016 - Self-Driving Tesla Was Involved in Fatal Crash, U.S. Says - https://www.nytimes.com/2016/07/01/business/self-driving-tes...
2016 - Tesla working on Autopilot radar changes after crash - https://www.freep.com/story/money/cars/2016/07/16/tesla-auto...
2017 - NTSB Issues Final Report and Comments on Fatal Tesla Autopilot Crash - https://www.caranddriver.com/news/a15339472/a-tesla-crash-bu...
2019 - Tesla didn’t fix an Autopilot problem for three years, and now another person is dead - https://www.theverge.com/2019/5/17/18629214/tesla-autopilot-...
2021 - Inside Tesla as Elon Musk Pushed an Unflinching Vision for Self-Driving Cars - https://www.nytimes.com/2021/12/06/technology/tesla-autopilo...
2021 - Tesla announces transition to ‘Tesla Vision’ without radar, warns of limitations at first - https://electrek.co/2021/05/25/tesla-vision-without-radar-wa...
2022 - Former Head Of Tesla AI Explains Why They’ve Removed Sensors; Others Differ - https://www.forbes.com/sites/bradtempleton/2022/10/31/former...
2022 - Tesla Dropping Radar Was a Mistake, Here is Why - https://www.idtechex.com/en/research-article/tesla-dropping-...
2023 - Tesla reportedly saw an uptick in crashes and mistakes after Elon Musk removed radar from its cars - https://archive.is/mL7zy
2023 - Elon Musk Overruled Tesla Engineers Who Said Removing Radar Would Be Problematic: Report - https://insideevs.com/news/658439/elon-musk-overruled-tesla-...
2023 - How Elon Musk knocked Tesla’s ‘Full Self-Driving’ off course - https://www.washingtonpost.com/technology/2023/03/19/elon-mu...
2023 - The final 11 seconds of a fatal Tesla Autopilot crash - https://www.washingtonpost.com/technology/interactive/2023/t...
I teach Tesla FSD as a cautionary tale of malicious business practices, how to deploy robots in the least safe way possible, and an anti-example of how to be a good engineer. It should be a crime to have as little a grasp on engineering and to have as much input as Musk does.
As an avid fsd user, your list makes me more comfortable using fsd than not. If there were more you’d parrot it around, but in the end, fsd is safer than humans at the wheel. As much as you dislike it because of Musk, fsd will continue to march forward and improve and millions will use it. You are beta testing fsd even if you don’t want to be.
> You are beta testing fsd even if you don’t want to be.
Exactly, that's precisely the wrong way to do ethical, responsible engineering. It's one thing to put your own life in the hands of a beta test, but it's another to put all of our lives at risk without our consent.
Distinguish a beta test from a learner driver or first year licensed driver. Why should I share the road with anyone who isn’t an experienced and competent driver?
Consent. We've consented to people learning to drive. Everyone is aware of those risks and they act accordingly. Often, to make other drivers aware of the driver's learning status, it's customary to affix a sticker to the car.
On the other hand, the public generally is not aware that these FSD machines are operating without critical safety equipment, and if they knew and were properly informed I'm not sure they would agree to be part of the test. I think Elon Musk knows that as well which is why he hasn't bothered to ask.
Moreover, I'm shocked and horrified by the number of people going around "beta testing" new versions of FSD. You'll see someone go to an area where a previous FSD version is known to fail just to see if the new one works. How is that a good idea at all? The last time you were here the car lurched into oncoming traffic unexpectedly and went down a one way, so let's see if it does it again, fingers crossed?
Utter recklessness, far worse than a 16 year old.
If we're using that definition of consent, it is an absolute unequivocal red herring. Roads are full of vehicles and full of drivers which, if you understood exactly what and who they are, you would absolutely refuse to consent sharing the road with. I didn’t consent to share the road with someone who failed to tighten their wheel nuts. I didn’t consent to share the road with that tourist in a rental car. I didn’t consent to sharing the road with a deer. I didn’t consent to sharing the road with Mercedes Drive Pilot or Ford BlueCruise or Mercedes SuperCruise or Nissan ProPILOT or any vehicles operated by Zoox or Waymo.
The dangerous you're talking about have been around for a long time, have been depicted in many forms over decades in all kinds of popular media media, have actually impacted people which has birthed reforms and new safety measures, and that is all part of the constant education and training the public receives.
None of that is done for FSD or tesla cars. People don't know how they work, what the dangers are, what their failure modes are, how to approach them, and no one is telling them!
If you want to lament a general lack of awareness and progressive education, then I'm fully on board. Perhaps we should require operators to pass an additional driver test to be legally permitted to supervise any L2+ system from any manufacturer. No BlueCruise or SuperCruise unless you've got the extra marking on your license.
But as far as "consent" is concerned, if you wish to use the public roads, consent necessarily means accepting the roads as a commons. This means, at minimum, accepting the presence of anything the relevant government authority deems legal.
I'm not really talking about formal education here.
> This means, at minimum, accepting the presence of anything the relevant government authority deems legal.
All that means if you have enough money you can put whatever death machines on the road you want. I don't think that's a good way to manage a road network, and will not ultimately lead to greater safety for all.
If you disagree with your government, that is something you need to take up with them.
We do not generally hold drivers responsible for slips in judgement when driving. Only in cases of extreme recklessness is it normal for a driver to be 'held responsible'.
It is silly that we hold self driving cars to an astronomically higher standard. We all know they are safer than human drivers. And that is true even if they kill 50 people a day.
If you blow through a stop sign and cause an accident you will be held responsible, no '' needed.
what an assertion
You can disagree, and I do too, but it is the basis of an interesting discussion.
Personally I think that fully autonomous vehicles need to be at minimum 100x safer than humans in order to maintain a social license to operate. I’m not saying the law needs to codify that. I’m saying that society could never accept a system which was merely 10x safer than humans. Imagine every vehicle in North America was replaced with a robotaxi and road deaths absolutely plummeted from 40,000 to just 1,000 per year. That’s still an article every day about “Elon Musk's river of blood”.
Supervised systems do not need such extreme safety, because the human driver remains responsible. But here’s an idea: if you want to use supervised FSD, that’s fine, but every collision gets automatically uploaded to a repository that is searchable by law enforcement and insurance companies.
I think you're right about having to be 100x safer, but it's definitely irrational and a bad thing.
Sadly correct.
Ah yes, we would never hold a driver accountable for running a stop sign and killing someone. It doesn't meet the bar for seriously reckless
Plenty of examples of this. In many jurisdictions even accidentally rolling through a stop sign does not even come close to 'serious recklessness'.
https://www.gulfcoastnewsnow.com/article/12-year-old-dead-af...
Yeah, but you asserted that we do not hold drivers responsible for slips in judgement.
We regularly hold drivers responsible for minor infractions that are otherwise consequence free.
The thing is it's a hypothetical 50 humans a day killed by the same driver.