FSD 14 now rolling out to employees

[Jhodgesatmb]

Oh it does, and often. Doesn't learn for the next time either.

Mine would make the same right turn into a left turn lane every single day if I let it. It would get over every single day into oncoming construction we've had up for 6 months that says right lane ending every time, unlike a person that has encountered it many times. It's just dumb code quite often still. It doesn't even read road speed limit signs like my 2022 Lexus does reliably.

That is what I hope to see improve someday soon maybe with 14.

It’s not code friend; it’s reinforcement learning. And no, it doesn’t learn autonomously but is fed scenarios and told when how it responds is correct. FSD updates include new scenarios that have been trained, but it is also possible for there to be ‘two steps forward and one step back’ updates.

Sponsored

 

[dalton108]

but it is also possible for there to be ‘two steps forward and one step back’ updates.

Common, even.

 

[gtgarner]

Oh it does, and often. Doesn't learn for the next time either.

Mine would make the same right turn into a left turn lane every single day if I let it. It would get over every single day into oncoming construction we've had up for 6 months that says right lane ending every time, unlike a person that has encountered it many times. It's just dumb code quite often still. It doesn't even read road speed limit signs like my 2022 Lexus does reliably.

That is what I hope to see improve someday soon maybe with 14.

What? Make a right turn into a left turn lane? What? Thats not even what I said. What does that even mean?

i know you aren't bringing up teammate driving. Are you kidding? It only drives on selected/pre-defined expressways.

Its a shame we have to weed through these kind of things.

 

[gtgarner]

I don’t think I’m understanding your argument, your arguing that humans suck compared to FSD, I agree with that, but you also said that FSD doesn’t need get better(or that it’s perfect for you, something along those lines) and I was saying there is a lot of improvements that can be made to FSD. Like Predictive Travel.

Wait...quote me saying that FSD should not get better.

Which post did i say that?

 

[Gaximus]

Wait...quote me saying that FSD should not get better.

Which post did i say that?

I was saying that you mentioned that you didn't know how it would get better, not that it should not get better. And I was giving examples on how it could be better.

I don't know how 14 could be significantly better than anything prior. FSD is great.

 

[gtgarner]

I was saying that you mentioned that you didn't know how it would get better, not that it should not get better. And I was giving examples on how it could be better.

Read. Read. Read ALL of my words.

Read what I was responding to. I was responding to "significantly".

My point was the word significantly. Significantly? How can it get significantly better? Better yes. Significantly better - no.

I even gave an example of the best basketball player ever. Who can get significantly better than Michael Jordan? better - maybe, but significantly? not ever.

 

[Gaximus]

Read. Read. Read ALL of my words. Read what I was responding to. I was responding to "significantly".

My point was the word significantly. Significantly? How can it get significantly better? Better yes. Significantly better - no.

That's a matter of opinion, I think that predictive travel is significant. I also think that going unsupervised will be a significant improvement, and that is where FSD is going.

 

[gtgarner]

That's a matter of opinion, I think that predictive travel is significant. I also think that going unsupervised will be a significant improvement, and that is where FSD is going.

I agree its a matter of opinion.

I don't think either will be significant. It will be better, but significant? no.

In comparing it to humans - humans hardly do predicative anything especially when texting and doing makeup and watching videos.

Do you realize who is driving cars. Its more than people like you and I. .

 

[HaulingAss]

I have been one of Tesla's greatest supporters, an original public shareholder, have never been short, have been along for the entire ride. I have personally instroduced and escorted Tesla execs to two statehouses and successfully assisted them in overcoming auto dealership resistance to Tesla showrooms. You have no idea who I am. I have a public persona, would be at personal risk for spreading FUD, in several ways. Let's keep it to substance please. Content, not personality.

Dave, I always stick to substance over personality, it is you who are holding up your Tesla "public persona" as some kind of proof that you couldn't possibly spread anti-Tesla FUD. However, people who derive income from their association with a Tesla persona are, in my observation, the ones most likely to sell false narratives. Putting substance above personality is exactly what I'm doing.

Let's look at your most recent post (to this one):

You actually accept this "safest vehicles" BS claim? It has been thoroughly debunked. Tesla has the highest fatality rate of any auto manufacturer based on miles driven. Tesla has the highest fatal accident rate of all car brands, followed by Kia, Buick, Dodge and Hyundai. Tesla's claim uses vastly distorted statistics that compare highway driving, where fatalities are the lowest.

You are actively pushing the disproven FUD claim that Tesla's have the highest fatality rate of any manufacturer based on miles driven. This is textbook anti-Tesla FUD, and you are propagating it in online forums. This contradicts your claim that you are a big Tesla supporter. Either you are one of the most gullible "Tesla fans" I know of, or you are simply using your Tesla fan persona to be more effective at spreading anti-Tesla narratives. Let's deal in facts, OK?


A 2024 study by iSeeCars, using NHTSA data from 2017-2022, claimed Tesla had the highest brand-level rate at 5.6 fatal accidents (with at least one occupant death) per billion vehicle miles traveled, ahead of Kia (5.5), Buick (4.8), Dodge (4.4), and Hyundai (3.9).

However, this study has been widely criticized and debunked for methodological flaws, including significantly underestimating vehicle miles traveled (VMT) for Tesla models—leading to inflated rates. For example, Tesla's VP of Vehicle Engineering, Lars Moravy, called the study "clickbait" and noted that actual Tesla data shows over 7 billion miles for the Model Y and 19 billion for the Model 3 by the end of 2022, far higher than iSeeCars' estimates.

Independent analyses of raw NHTSA FARS data confirm the study's Model Y rate is overstated by nearly 4x, with a more accurate figure around 2.5-2.86 per billion miles—close to the overall U.S. average of 2.8.

Tesla vehicles consistently earn top crash test ratings from NHTSA (5 stars) and IIHS (Top Safety Pick+), and IIHS driver death rates (a different metric) often show Tesla models with very low or zero deaths in available data, contrasting with higher-risk small cars and muscle cars.

Using a debunked study that grossly under-estimated the actual number of miles travelled to grossly over-state the death rate does not reflect well on your public persona. And, yes, I know exactly who you and your son are, even though you think you are anonymous. You two are not the only father/son public personas who consistently throw shade on Tesla while pretending to be fans.

Tesla knows exactly how many miles their fleet has travelled, the iSee study did not use the actual data of the miles travelled to calculate the Tesla death rate, they used their own inaccurate estimates (likely purposefully underestimating the miles travelled to wildly inflate the death rate relative to other manufacturers). You don't have to be particularly intelligent to understand that this happens all the time.

Furthermore, your comment, "You actually accept this "safest vehicles" BS claim?" focuses on the safety of the cars themselves, not the drivers. It's widely known that muscle cars have a high fatality rate because they are more likely to drive at higher speeds, with higher rates of acceleration and this kind of behavior increases the fatality rate. Serious studies on the safety of different brands/models of cars use statistical techniques to normalize for different demographics and driver behaviors. That is the only way to accurately assess the safety of a car model or brand without the problem of different driver behaviors skewing the results. By amplifying a widely debunked study that was created for the purpose of throwing shade on Tesla, you become part of the problem, not the solution.

Why do you twist the facts and amplify clearly bogus studies that throw shade on Tesla while simultaneously claiming to be one of Tesla's biggest fanboys? It's not a good look, although I can see you clearly make money doing it. If you could explain why you are pushing this bogus study that purports Tesla's have the highest fatality rate of any manufacturer, then I'm all ears.

The ironic thing is only 8 minutes elapsed between your "You actually accept this "safest vehicles" BS claim? Tesla has the highest fatal accident rate of all car brands " and the claim that you are a huge Tesla supporter. Do you not believe in measuring the death rate by the actual miles travelled?

 

[jimbobb23]

I don't know how 14 could be significantly better than anything prior. FSD is great.

Thats like saying that someone could be significantly better than Michael Jordan. How could that be possible?

It's pretty great and the last 10% might take another couple years before it can drive on its own perfectly. But right now I think they could have it do highways at level 3 no problem. If it seems construction back to level 2. But it's basically perfect on highways.

 

[GnarlyDudeLive]

I would just love to see how a DMV would evaluate FSD. There is absolutely no way they should fail FSD looking at how they pass some humans.

I can add some context to your thoughts. My Father-in-law is newly retired: He taught sports and Driver's Education. Not only did he teach Drivers Ed at the local high school but he also wrote the entire college course book and taught Drivers Ed at the local university for other teachers wanting to teach it as well. So basically, he taught the teachers. He has had multiple multi hour rides with me and riding shotgun and was very impressed at just about every aspect of my CT self driving. I eventually asked him how he would compare FSD to his students that PASSED his classes. He laughed pretty hard and said "Way better". Anytime we go places as a family group, he always wants me to take the CT just so he can watch it drive. He does all the counting and stuff that the drivers ed instructors do. Its pretty amusing.

 

[gtgarner]

I can add some context to your thoughts. My Father-in-law is newly retired: He taught sports and Driver's Education. Not only did he teach Drivers Ed at the local high school but he also wrote the entire college course book and taught Drivers Ed at the local university for other teachers wanting to teach it as well. So basically, he taught the teachers. He has had multiple multi hour rides with me and riding shotgun and was very impressed at just about every aspect of my CT self driving. I eventually asked him how he would compare FSD to his students that PASSED his classes. He laughed pretty hard and said "Way better". Anytime we go places as a family group, he always wants me to take the CT just so he can watch it drive. He does all the counting and stuff that the drivers ed instructors do. Its pretty amusing.

Excellent news.

Thanks for taking the time to inform/educate us based on a professional.

 

[REM]

Dave, I always stick to substance over personality, it is you who are holding up your Tesla "public persona" as some kind of proof that you couldn't possibly spread anti-Tesla FUD. However, people who derive income from their association with a Tesla persona are, in my observation, the ones most likely to sell false narratives. Putting substance above personality is exactly what I'm doing.

Let's look at your most recent post (to this one):



You are actively pushing the disproven FUD claim that Tesla's have the highest fatality rate of any manufacturer based on miles driven. This is textbook anti-Tesla FUD, and you are propagating it in online forums. This contradicts your claim that you are a big Tesla supporter. Either you are one of the most gullible "Tesla fans" I know of, or you are simply using your Tesla fan persona to be more effective at spreading anti-Tesla narratives. Let's deal in facts, OK?


A 2024 study by iSeeCars, using NHTSA data from 2017-2022, claimed Tesla had the highest brand-level rate at 5.6 fatal accidents (with at least one occupant death) per billion vehicle miles traveled, ahead of Kia (5.5), Buick (4.8), Dodge (4.4), and Hyundai (3.9).

However, this study has been widely criticized and debunked for methodological flaws, including significantly underestimating vehicle miles traveled (VMT) for Tesla models—leading to inflated rates. For example, Tesla's VP of Vehicle Engineering, Lars Moravy, called the study "clickbait" and noted that actual Tesla data shows over 7 billion miles for the Model Y and 19 billion for the Model 3 by the end of 2022, far higher than iSeeCars' estimates.

Independent analyses of raw NHTSA FARS data confirm the study's Model Y rate is overstated by nearly 4x, with a more accurate figure around 2.5-2.86 per billion miles—close to the overall U.S. average of 2.8.

Tesla vehicles consistently earn top crash test ratings from NHTSA (5 stars) and IIHS (Top Safety Pick+), and IIHS driver death rates (a different metric) often show Tesla models with very low or zero deaths in available data, contrasting with higher-risk small cars and muscle cars.

Using a debunked study that grossly under-estimated the actual number of miles travelled to grossly over-state the death rate does not reflect well on your public persona. And, yes, I know exactly who you and your son are, even though you think you are anonymous. You two are not the only father/son public personas who consistently throw shade on Tesla while pretending to be fans.

Tesla knows exactly how many miles their fleet has travelled, the iSee study did not use the actual data of the miles travelled to calculate the Tesla death rate, they used their own inaccurate estimates (likely purposefully underestimating the miles travelled to wildly inflate the death rate relative to other manufacturers). You don't have to be particularly intelligent to understand that this happens all the time.

Furthermore, your comment, "You actually accept this "safest vehicles" BS claim?" focuses on the safety of the cars themselves, not the drivers. It's widely known that muscle cars have a high fatality rate because they are more likely to drive at higher speeds, with higher rates of acceleration and this kind of behavior increases the fatality rate. Serious studies on the safety of different brands/models of cars use statistical techniques to normalize for different demographics and driver behaviors. That is the only way to accurately assess the safety of a car model or brand without the problem of different driver behaviors skewing the results. By amplifying a widely debunked study that was created for the purpose of throwing shade on Tesla, you become part of the problem, not the solution.

Why do you twist the facts and amplify clearly bogus studies that throw shade on Tesla while simultaneously claiming to be one of Tesla's biggest fanboys? It's not a good look, although I can see you clearly make money doing it. If you could explain why you are pushing this bogus study that purports Tesla's have the highest fatality rate of any manufacturer, then I'm all ears.

The ironic thing is only 8 minutes elapsed between your "You actually accept this "safest vehicles" BS claim? Tesla has the highest fatal accident rate of all car brands " and the claim that you are a huge Tesla supporter. Do you not believe in measuring the death rate by the actual miles travelled?

Again, glad that I blocked that FUDster months ago. He lives in a completely different plane of reality than the rest of us.

 

[noobla3]

I can add some context to your thoughts. My Father-in-law is newly retired: He taught sports and Driver's Education. Not only did he teach Drivers Ed at the local high school but he also wrote the entire college course book and taught Drivers Ed at the local university for other teachers wanting to teach it as well. So basically, he taught the teachers. He has had multiple multi hour rides with me and riding shotgun and was very impressed at just about every aspect of my CT self driving. I eventually asked him how he would compare FSD to his students that PASSED his classes. He laughed pretty hard and said "Way better". Anytime we go places as a family group, he always wants me to take the CT just so he can watch it drive. He does all the counting and stuff that the drivers ed instructors do. Its pretty amusing.

Did he count the follow distance of the Cybertruck and the car in front of it?

 

[GnarlyDudeLive]

Did he count the follow distance of the Cybertruck and the car in front of it?

So many times. Lol

Sponsored