Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Blame the user for not making sure the flawed autonomous system doesn’t crash. I wonder how many more crashes would happen if people didn't intervene to prevent FSD from crashing. Sounds like the number of crashes would be even higher

this is exactly why unattended Tesla FSD is not yet permitted in most jurisdictions

I'll point to this again for anyone looking for real numbers and analysis that aren't just a Tesla press release:

 
  • Like
Reactions: Thehangmn
scalability is irrelevant to the question of which system is actually safer
Scalability is important as things get underway. Tesla is poised for scalability. And for safety it depends if immobilizing your fleet or hitting a telephone pole is “less worse”.
 
Scalability is important as things get underway. Tesla is poised for scalability. And for safety it depends if immobilizing your fleet or hitting a telephone pole is “less worse”.

level of scalability does not effect how safe the system is one way or another

and again, it doesn't matter how scalable it is if it's not allowed on the roads because it is not yet safe enough
 
😂. Again no goal posts being moved,

You saying it constantly doesn't make it true.

you cannot see the future, who knows what the bottom is. Indicators are mixed.

You seem to not understand what bottom means. Bottom is the lowest price reached before the stock begins a meaningful upward move. That's where you went wrong.

You’re the one expressing confidence, I’m the one saying that time will tell and it’s too early to say.
No. I expressed literal fact of what happened. Two consecutive quarters of improved numbers defined the bottom that already happened.

Cyber truckout be sold out because they aren’t making any.

They aren't making any yet they literally started delivering a new trim a few weeks ago. That new trim went from June 2026, to April 2027, to just 2027 in a matter of a few days of launching.

Literally you're wrong.

If the cybercab is in production where is that amazing scaling? None currently on the road?
Cybercab just started a couple of months ago. Production starts slow then gets faster and faster. It's literally scaling 😂
Plenty of videos of it on the road and hundreds shown in parking lots.

What happened to “the numbers don’t lie?” Cybertruck is a dud and you know it.
Model y is selling well for now.

You're literally wrong. See above for the explanation.

I can help someone that is incapable or unwilling to understand.

Says a lot when you refuse to answer the questions.

😂. I am changing my wording

It's so obvious that you are.

So the camera inference can’t be trusted
So the cameras shouldn’t be trusted. Lidar make sure the pole was seen at all.
And why did Waymo put more weight into the camera system over its lidar system when calculating the damage score? See where this is heading? You're almost there.

Conjecture. Maybe camera inference is just not enough
If you looked at a frame coming straight from the camera and can make out the telephone pole, camera inference is more than enough to tell you that the car is about to collide. Not conjecture, but a simple fact.

Any Waymo crash must be sensor contention because Elon says right?
Nope. Never said any. You're trying to strawman. It's not working.

And every Tesla crash is because…
I'm not making any generalizations of *every* Tesla crash and neither am I making any generalizations of *any* waymo crash.


You can pretend that is the issue of the cybertruck hitting the pole, but that’s a different crash. Cybertruck running FSD crashed into a pole at high speed. Review the thread, I know you were gone for a bit there so maybe you forgot.

I don't see anywhere showing hard data that it was not the driver overriding FSD. Sounds to me it's conjecture.

If all Waymo crashes
never said all waymo crashes. more strawmanning.
 
level of scalability does not effect how safe the system is one way or another

and again, it doesn't matter how scalable it is if it's not allowed on the roads because it is not yet safe enough
We are going under the assumption they will all get there and Tesla would be in a better position.
 
There's plenty of data publicly available

How many crashes did FSD avoid that would have happened without it. The website you cited is not the epitome of "plenty of public data available". You're only presenting a one sided argument. 🙄
 
The website you cited is not the epitome of "plenty of public data available". You're only presenting a one sided argument. 🙄

I’m guessing you neither read the full article nor investigated its sources. It’s based on and references plenty of publicly available data, including but not limited to data from National Highway Traffic Safety Administration.

There’s plenty more:

 
Last edited:
I’m guessing you neither read the full article nor investigated its sources. It’s based on and references plenty of publicly available data, including but not limited to data from National Highway Traffic Safety Administration.

There’s plenty more:


So based on your stated goal of perfection, airplanes should’t fall out of the sky? Or fleets shouldn’t become immobilized. I’m guessing you didn’t read the article I posted or else FSDs “crashes” pale in comparison to the US average of 5 million. I’ll take my chances with FSD. And at least the fleet won’t become immobile.

And the one sided picture doesn’t detail how many accidents were avoided. Because bad news rules?
 
So based on your stated goal of perfection, airplanes should’t fall out of the sky?

And the one sided picture doesn’t detail how many accidents were avoided. Because bad news rules?

I have no stated goal of perfection. You must be mixing me up with someone else

I’ve pointed you toward some good actual journalism on the subject since you seem interested in it. I’ll leave it to you whether you want to enlighten yourself or stick to the talking points from your elon fanboy site fed from elon's press releases and social media posts

Hope you bought some $700 floor mats to keep the site running! 😂
 
Last edited:
You saying it constantly doesn't make it true.
Let me see how this works…
You saying that constantly doesn’t make it true.
You seem to not understand what bottom means. Bottom is the lowest price reached before the stock begins a meaningful upward move. That's where you went wrong.
you’re talking about the stock now? Who is moving goal posts? We are talking earnings, deliveries. Tesla stock price is already way overvalued.

No. I expressed literal fact of what happened. Two consecutive quarters of improved numbers defined the bottom that already happened.
So Tesla is up and to the right forever? Interesting. Again your confidence is interesting. I suspect the true bottom has yet to come.
They aren't making any yet they literally started delivering a new trim a few weeks ago. That new trim went from June 2026, to April 2027, to just 2027 in a matter of a few days of launching.
Yeah? What are the volumes? Last I heard 10% of total demand was coming from SpaceX.
Literally you're wrong.
Define success for the Cybertruck. It sells nowhere near the volumes of the Y or 3. It’s an order of magnitude difference. They are losing money on it
Cybercab just started a couple of months ago. Production starts slow then gets faster and faster. It's literally scaling 😂
Plenty of videos of it on the road and hundreds shown in parking lots.
So how many of the 50 tesla cabs are cyber cabs?
You're literally wrong. See above for the explanation.
Explain where I’m wrong. Tesla has 50 cabs on the road… none of which are cyber cabs
Says a lot when you refuse to answer the questions.
I keep answering, just can’t get through to you. It’s like you are plugging your ears
It's so obvious that you are.
So do you get it yet? Or do I have to try again? Im trying my best to help you here. I’ve maintained the same meaning, I think you are purposely trying to twist the words because you’ve got nothing better to work with.

And why did Waymo put more weight into the camera system over its lidar system when calculating the damage score? See where this is heading? You're almost there.
You don’t know that’s the case. frankly your premise that (if they agreed and saw the pole, the score would be higher)
assumes the score is read-out of sensor agreement, but it isn't. its a readout of estimated harm.
Imagine a plastic bag floating in the lane. Both the camera and Lidar could detect it clearly and agree completely that it is there, but the damage score would still be Zero, because it's a plastic bag. So even if we accept what you say, the error isn't in sensor contention, it's in damage weighting. Waymo changed their software to assign all poles a higher damage score, and add the missing road edge to the map (the bigger underlying issue). to suggest that this is all because the lidar saw it and cameras didn't is conjecture on your part. They both could have seen it and the damage score could still be incorrectly assigned. Your assumptions are problematic here, you don't know how Waymo uses the Camera and lidar data to assign the damage score, you are assuming that the camera inference was poor, and this led to a low damage score. There is no evidence to suggest that is the case, and even if it is... it's not a sensor contention issue anymore.
If you looked at a frame coming straight from the camera and can make out the telephone pole, camera inference is more than enough to tell you that the car is about to collide. Not conjecture, but a simple fact.
The cybertruck didn't seem to get that when it crashed into a telephone pole, perhaps your assumptions are not accurate?

Nope. Never said any. You're trying to strawman. It's not working.
you've blamed a lot on "sensor contention" without actually backing it up. just assumptions and conjecture. find me a source that confirms it, and you might just have something. otherwise, you've got nothing. Just Elons words, and your assumptions.
I'm not making any generalizations of *every* Tesla crash and neither am I making any generalizations of *any* waymo crash.
You saying that doesn’t make it true.
I don't see anywhere showing hard data that it was not the driver overriding FSD. Sounds to me it's conjecture.
FSD is funny that way, always easy to blame the user. harder to run fully autonomous taxis. (as Elon is finding out)
never said all waymo crashes. more strawmanning.
oh good, I'm glad you're coming to see the light.
 
I have no stated goal of perfection. You must be mixing me up with someone else

I’ve pointed you toward some good actual journalism on the subject since you seem interested in it. I’ll leave it to you whether you want to enlighten yourself or stick to the talking points from your elon fanboy site fed from elon's press releases and social media posts

Hope you bought some $700 floor mats to keep the site running! 😂
🤣 man, not you too!

Look, we are talking about continuous improvement obligation. I was specifically talking about tesla removing redundant sensors and going vision only. The burden is on tesla to demonstrate that the residual risk of removing these sensors is acceptable and that rejected redundant systems were rejected for defensible reasons, not to assert it and not to let the public found out via an NHTSA probe.
Tesla has made its architecture choice unilaterally, declined to publish the analysis that would justify it, historically redacted the crash data that would let outsiders test it, and framed the quote question in Elon's rhetoric "lidar is a crutch" "freaking stupid" which is the language of ideology, not a ranked risk queue.
So if Tesla bets on camera, they should prove that it is at least as safe as the alternative they refused, and back it up with evidence that others can check. But Tesla has done the betting, and skipped the proving. Elon insisting that Waymos recent voluntary recall is because of sensor contention is again, ideology and not data based science.
He is welcome at any time to prove his position with data. Will he ever do it?

I get that perfection is not attainable, and allowing imperfect systems is the reality of human existence. The point is that we should never stop improving the systems, and disregarding sensors because they are "freaking stupid" without providing any proof that rejecting them is just as safe, is what we refer to as "assertion based safety".
 
🤣 man, not you too!

Look, we are talking about continuous improvement obligation. I was specifically talking about tesla removing redundant sensors and going vision only. The burden is on tesla to demonstrate that the residual risk of removing these sensors is acceptable and that rejected redundant systems were rejected for defensible reasons, not to assert it and not to let the public found out via an NHTSA probe.
Tesla has made its architecture choice unilaterally, declined to publish the analysis that would justify it, historically redacted the crash data that would let outsiders test it, and framed the quote question in Elon's rhetoric "lidar is a crutch" "freaking stupid" which is the language of ideology, not a ranked risk queue.
So if Tesla bets on camera, they should prove that it is at least as safe as the alternative they refused, and back it up with evidence that others can check. But Tesla has done the betting, and skipped the proving. Elon insisting that Waymos recent voluntary recall is because of sensor contention is again, ideology and not data based science.
He is welcome at any time to prove his position with data. Will he ever do it?

I get that perfection is not attainable, and allowing imperfect systems is the reality of human existence. The point is that we should never stop improving the systems, and disregarding sensors because they are "freaking stupid" without providing any proof that rejecting them is just as safe, is what we refer to as "assertion based safety".

I’m going to chalk this up to you not actually reading my posts or the journalism I’m pointing to

I more or less agree with everything you are saying. I think you may have gotten something mixed up here

All that I’m doing here is pointing out the ridiculousness of the Tesla fanboys arguments and claims, while pointing them to actual information that they ignore
 
Last edited:
Let me see how this works…
You saying that constantly doesn’t make it true.
😂 Me saying facts that are backed up by official sources makes it true.

you’re talking about the stock now? Who is moving goal posts? We are talking earnings, deliveries. Tesla stock price is already way overvalued.

😂 Calling "bottom" applies to practically any stats, which is typically used in stock price. Delivery numbers aren't treated differently.

So Tesla is up and to the right forever?

😂 No, I just explained bottom doesn't mean forever. LITERALLY just explained it to you and you still can't understand it. This is incredible.

Yeah? What are the volumes? Last I heard 10% of total demand was coming from SpaceX.

😂 Volume that maxed out their production where you will get one delivered in 2027 if you ordered today

Define success for the Cybertruck.

😂 Moving goal posts once again, you claimed they aren’t making any. Yet they're making as much as they can from their line and delivering every single one. Even the used CTs were sold out at one point. Also saw a post of one of the original founders edition selling more than they bought it for.

The statement "Cyber truck be sold out because they aren’t making any." is objectively false.

So how many of the 50 tesla cabs are cyber cabs?

😂 That's a different question than the original claim. Moving goal posts once again.

Explain where I’m wrong.

I did. Literally said "see above for the explanation

I keep answering,
.

Answer the questions.
1. Is Waymo less than perfect?
2. Do you accept Waymos being on the road?

Basic yes or no questions. If you do not answer these questions with either yes or no, then it's clear you haven't answered at all.

So do you get it yet?

I get you are changing what you're saying after the fact.

I’ve maintained the same meaning

😂 It's so false.

You don’t know that’s the case.

Sure I do.

frankly your premise that (if they agreed and saw the pole, the score would be higher)
assumes the score is read-out of sensor agreement, but it isn't. its a readout of estimated harm.

If they trust more on the camera system and the camera system's inference returned nothing or something that's able to be driven through, it's clear they put less weight on the telephone pole coming from the lidar data.

Imagine a plastic bag floating in the lane. Both the camera and Lidar could detect it clearly and agree completely that it is there, but the damage score would still be Zero, because it's a plastic bag. So even if we accept what you say, the error isn't in sensor contention, it's in damage weighting.

Actually across time, even without cameras, lidar can pick up an object to be classified as a plastic bag. Because an object that flies/floats like a plastic bag can easily be classified as something that won't cause damage at all. A camera would confirm this but even if the camera system failed to infer it completely, lidar's inference would suggest it's a low damage scoring object and drive straight through.

Meanwhile, a tall pole, fixed in position, there's no question lidar can pick up as a telephone pole. But because it was still classified as a low damage score, it's clear that their system put more weight into the camera system that failed to pick up on the pole, despite the camera sensors can clearly see it is a telephone pole.

You gave a terrible example.

The cybertruck didn't seem to get that

Based on the erroneous assumption that driver didn't press on the accelerator, which in many cases it has happened.

you've blamed a lot on "sensor contention" without actually backing it up.

I've backed it up with how it works.

Meanwhile you seem to just assume it was FSD that caused a cybertruck crash with zero evidence to back that up.

just assumptions and conjecture.

I've showed you many times where a driver lied about FSD/autopilot causing crashes when in reality, the data showed the driver pressed the accelerator.

find me a source that confirms it,
It's axiomatic, but happy to hear other explanations. Your plastic bag example didn't disprove what I said.

Just Elons words, and your assumptions.

And waymo's explanation. You know. The part about damage scoring and how they fused the data and run inference

You saying that doesn’t make it true.
Point to where I made a generalization of *every* tesla crash and generalizations of *any* waymo crash. Direct quote please. If you cannot provide this, your argument failed.

FSD is funny that way, always easy to blame the user.
1. That's your argument? "FSD is funny"? 😂 conjecture.
2. many times the data has showed the user to be lying about fsd/autopilot, even one went to jail for it.
oh good, I'm glad you're coming to see the light.

Still waiting for where I said anything about "all waymo crashes". Direct quote please. Show me.
 
I have no stated goal of perfection. You must be mixing me up with someone else

I’ve pointed you toward some good actual journalism on the subject since you seem interested in it. I’ll leave it to you whether you want to enlighten yourself or stick to the talking points from your elon fanboy site fed from elon's press releases and social media posts

Hope you bought some $700 floor mats to keep the site running! 😂
Did you not read the posted link? I posted some actual journalism that’s a respected publisher. I’ll leave it to you to decide the relative merits and if you find yourself needing an autonomous ride spend more money than necessary. 😀
 
Did you not read the posted link? I posted some actual journalism that’s a respected publisher. I’ll leave it to you to decide the relative merits and if you find yourself needing an autonomous ride spend more money than necessary. 😀

😂

Teslarati is not a “respected publisher” it’s an Elon fan site. Parroting elons social media posts, press releases and talking points is not journalism
 
  • Haha
Reactions: I7guy
Answer the questions.
1. Is Waymo less than perfect?
2. Do you accept Waymos being on the road?

Basic yes or no questions. If you do not answer these questions with either yes or no, then it's clear you haven't answered at all.

Answering yes to those questions doesn’t mean that Tesla FSD should be allowed on the road, which I assume is the argument you trying to make

That conclusion in no way follows from the premise

It’s like arguing:

P1 drinking alcohol isn’t perfectly healthy

P2 drinking bleach isn’t perfectly healthy

C: if we allow the sale of alcohol based drinks we should also allow the sale of bleach based drinks
 
  • Haha
Reactions: I7guy
Calling "bottom" applies to practically any stats, which is typically used in stock price.

“Tesla shares have underperformed the Nasdaq by 30.2% year-to-date (Tesla is -18% vs the Nasdaq's +12%) due to Musk having hyped up the rollout of driverless robotaxis more than usual last year, only to have a few unsupervised cars driving around a small geofence for several hours a day. If he fails to come up with a good excuse or a new, believable timeline, the stock should sell off.”
 
Answering yes to those questions doesn’t mean that Tesla FSD should be allowed on the road, which I assume is the argument you trying to make

That conclusion in no way follows from the premise

It’s like arguing:

P1 drinking alcohol isn’t perfectly healthy

P2 drinking bleach isn’t perfectly healthy

C: if we allow the sale of alcohol based drinks we should also allow the sale of bleach based drinks
How many accidents did FSD avoid? It’s just as important to know how FSD saved the day as much as caused a crash. Right?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.