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

Tesla.com - "Transitioning to Tesla Vision"

This site may earn commission on affiliate links.
The media is reporting on the fact that NHTSA called Tesla, consulted with them and then based on that, removed AEB from NHTSA's own NCAP list, and reduced the safety rating of the cars based on this.
No media is responding to any forum post.
Tesla was unable to convince NHTSA that the AEB on the current cars met NHTSA's definition for AEB. Yes, maybe that just means it's untested, but that is still a news-worthy event that Tesla made a large enough change to the cars that a regulator had to remove a feature from the list until it's tested. Who knows if it will past the test? If you care about NHTSA safety ratings, you can't buy a new car that just came out until NHTSA rates it. In this case, if you care about AEB, you also can't buy a 3/X until NHTSA tests it. It's functionally a new car, and it's up to the consumer to understand if they will trust Tesla that it is there before a regulator verifies.

All sounds very, very newsworthy for a company that is so focused on safety yet is removing safety features (even if supposedly temporary).

To me this all really comes down to lack of communication from Tesla.

We don't know why they pulled the radar BEFORE features were at parity. If it was a support chain issue they could just acknowledge it, and I think most people would be understanding.
We don't know from the Tesla Vision post on Tesla's website what the current state is of FCW, AEB, etc. Sure I have no problems believing they are in there, but they're waiting on official testing. But, that's because I'm pretty comfortable separating bs Musk from factual Musk.
We don't know how they're going to use the Radar on cars once they release Tesla Vision on HW3 cars that have Radar.
The validity of any news I read about Tesla is questionable because Tesla has no PR department for journalist to send in inquiry to before posting anything.
Tesla doesn't include any useful information on their release notes. Usually its the same release notes for 4+ updates where you're just blindly installing something having no idea of what bugs were fixed.
You can't just message or email them. I don't see any mechanism using the app or the website to do so. The only thing I see is the ability to "chat with us" between 8am, and 3pm PST. Sure I could call, but why would I want to waste my time calling on a non-critical inquiry to answer some question or concern?

To me the lack of communication is a strong indication that you are in a relationship with someone lying to you. Because the hardest thing about a lie is maintaining the lie so its best to not say anything.
 
We don't know why they pulled the radar BEFORE features were at parity. If it was a support chain issue they could just acknowledge it, and I think most people would be understanding.

Exactly. To top it off they pulled radar from 3/Y but not the new S/X, which doesn't actually make a ton of sense if they are 100% sure of actually going vision-only. What is the purpose of providing radar in the S/X if it won't be used?

The official word is:

Why aren’t Model S and Model X transitioning now?
Model 3 and Model Y are our higher volume vehicles. Transitioning them to Tesla Vision first allows us to analyze a large volume of real-world data in a short amount of time, which ultimately speeds up the roll-out of features based on Tesla Vision.

This makes it sound like the new cars are beta test vehicles and they aren't sure of the approach yet.
 
If it was a support chain issue they could just acknowledge it, and I think most people would be understanding.
For any normal company, this would be an understandable challenge they are facing like any other automotive company is right now.

However Elon had to make a big point in their earnings reports about now nimble and pivot-y they are, and how that helped them avoid shutdowns, and how they did this without impacting customers. That was only a few weeks ago.

Acknowledging they are removing radar because of a supply chain issue really hurts that narrative. Much better to frame it as a long term plan, with "temporary" impacts to customers.

We know Tesla is big on narrative- this is the same company that went back and purged previous articles explaining how useful radar is from their blog because it didn't fit the current story. Totally normal. Totally cool.
 
To me this all really comes down to lack of communication from Tesla.

That's true of FUD in general though. Blame Tesla for:

1) not immediately commenting on all "suspected" autopilot deaths
2) not explaining all parked cars on their lots
3) not explaining every decision they make immediately, in detail
4) the impact of any fires on any of their properties
5) not giving us a daily car production #s
6) providing exact # of chips they're getting per day
7) providing a website about how many car ships are going to Europe per month
8) list goes on

Fact is, the blame should be on people jumping to conclusions too quickly. This whole vision-only thing was hours old before everyone was saying it's horrible, safety is compromised, etc.

Why create FUD before more details even come out? Much of this has been based on pure speculation with negative bias.
 
Last edited:
We don't know why they pulled the radar BEFORE features were at parity.
Because "shipping" is a feature.

EVERY company in the world works to get out things as quickly as possible and decide on a MVP. Well, atleast tech companies do.

In every project I've done - if we waited for feature parity (esp. with competitors) we'd never ship anything.
 
In every project I've done - if we waited for feature parity (esp. with competitors) we'd never ship anything.
Can you describe a product you worked on where you had a shipping and functioning product, advertised that to customers, and then removed internal hardware to save cost, while saying the software to fix the deficiency is "coming soon"?

There is nothing normal about this "shipping"- what Tesla did in this case was remove one <$100 module that was fully functional and integrated from a $50K car, while asking their customers to be patient as they work it out. It literally takes work for them to remove this from the car. It took work to remove parity, not achieve it. That's why it's clear what happened here is driven by more than normal development practices.
 
Fact is, the blame should be on people jumping to conclusions too quickly. This whole vision-only thing was hours old before everyone was saying it's horrible, safety is compromised, etc.
Tesla's own announcement of Vision only says safety is compromised. If Tesla didn't want that to be the message, they had full control of it. I applaud their transparency, but you can't call it FUD or jumping to conclusions when it comes right from Tesla.

I also imagine you are fully OK with discussing possible positive news around Tesla right when it comes out, even if it's not 100% proven yet. Like Elon tweets about how feature X is right around the corner.

not immediately commenting on all "suspected" autopilot deaths
It took Elon about 24 hours to comment on the recent Florida incident. They are very quick to respond when they can distance themselves from it being autopilot. It's still interesting that he said they have data that the car was not on AP, but the NHTSA preliminary investigation still makes no note of any kind of OTA information, and discusses damages to the onboard loggers.
 
Much of this FUD'y whining would be appropriate 2 months after the first vision-only cars are delivered. Not now...

If the features reach parity after 1-2 months, big deal. We don't know yet.

That's why it's FUD'y right now. Too premature.
 
  • Like
Reactions: Doge-1
If the features reach parity after 1-2 months, big deal. We don't know yet.
So basically, when it's negative news, you want the world to give Tesla a pass for a few months, and see if they fix it, and then if they don't let's discuss it then. After thousands of people already paid for their non-refundable cars.

When Elon tweets "button goes out to everyone in 10 days" or "vision only solves phantom braking" - it's worth discussing right away though. Unless of course you have the opinion that those statements are likely untrue based on a large amount of history. Then that would be FUD. Please hold your opinions until 60 days after a 10 day tweet, a 2 month delay is no big deal.

Elon shows off a new Roadster or Cybertruck and says it's coming in 2 years? Endless news cycles, yes please, the media is awesome. Elon says it can do 0-60 in 1.9 seconds? Publish that ALL DAY LONG. Someone runs some math and says that's really unlikely and even Tesla's own website has an asterisk? FUD FUD FUD, wait until the cars have been out for a few months before you dare push on the narrative.
 
Last edited:

Direct quote from Elon. They are currently in a 'not approved' state until tested but they are there and active. This is a regulatory thing that will likely be old news next week.
It's a bit more than just a "regulatory thing" dude. The NHTSHA has *no* confidence that the safety features that relied on radar will work with an all-camera system, and there's TONS of industry reporting that those features, will, in fact, suffer greatly without radar (not to mention LIDAR)...particularly in low-light situations and particularly with Tesla's shitty 2015 cameras. This is a stunningly poor move by Tesla, and one that I suspect is due to something they're not telling us.

I'll say this...if Tesla's software stops using the radar that is already in my own Model Y, I'm selling it as a safety issue. There's already enough "camera blocked" or "degraded" messages now. Forget about not having radar in bad weather Holy sh*t, this is a bad move.

My wife's Volvo XC40 all-electric Recharge has Model-X like features and quality, with radar, for a Model Y price. Sure, the real-world range is only 208 miles, but it's got fast charging options.
 
This is a stunningly poor move by Tesla, and one that I suspect is due to something they're not telling us.

Typical FUD'y logic here though. It's possible Tesla did this because of some radar shortage or some other negative reason, but we have no real assessment of the vision-only cars yet. Perhaps it will be like Elon says, and they're actually safer than vision+radar, especially within the next month or two.
 
Perhaps it will be like Elon says, and they're actually safer than vision+radar, especially within the next month or two.
Source Elon says it will be safer?

Perhaps, like Elon says, we'll have 1M robotaxis by the end of this year and will be able to sleep in our cars, or we'll have a FSD button soon, or we'll see that despite the lack of visible progress they have actually been making enormous strides internally. Perhaps. Please just wait 2 months from whatever day it is today before you decide. Free beer tomorrow too. But your money for the car NOW please. No refunds. DOGE to the moon. Diamond hands.
 
Typical FUD'y logic here though. It's possible Tesla did this because of some radar shortage or some other negative reason, but we have no real assessment of the vision-only cars yet. Perhaps it will be like Elon says, and they're actually safer than vision+radar, especially within the next month or two.
I don't know -- or much care -- what "FUD'y" logic means.

What I do care about is that legions of actual automotive experts with years of experience are commenting on this in the media as an almost inexplicable move, including some of the very posts with links to articles in this (and other) threads about this topic.

I have 25+ years of experience working on projects that directly involve computer vision (CV) techniques at a major, well-known aerospace industry company. I agree with what experts are saying about this, rather than fanbois on a Tesla forum. And what they're saying, almost universally, is that removing additional sources of range, speed, and angle data...particularly when the costs of automotive-grade radar and LIDAR are plummeting...is idiotic. The very fact that *every* other automaker is moving in the exact opposite direction based on their own testing and analysis...and have higher safety test ratings than Tesla...should tell you something.

So, how about refuting *that* logic? We'll take our answer here....now. Go ahead.
 
saftey board removing aeb from model 3 and Y is an assessment...
No, no, no, you don't get it. They just haven't had time to test them yet. Regulators are slow, which is why Tesla didn't tell them about this and they had to find out about the change from a post on Tesla's website and call them the next day. Now they are just being FUDy and over-reacting. They should let Tesla sell cars to the public for a few months before they evaluate it to see how it is then. Perhaps it works fine. This is just the way Tesla works, what are they worried about?
 
Even if vision-only is not as good as vision+radar for a few months, it'll probably be better than other OEMs' implementations:


Again, I don't know. We don't know. Wait a few months to get a better picture of the new system's weaknesses or issues.
 
Even if vision-only is not as good as vision+radar
1622226062879.gif
 
Are you aware that Tesla themselves removed AEB from the standard functions list on the autopilot page when they announced Tesla Vision?
It only came back when someone pointed it out.
Then, when the NHTSA called them, they were unable to convince them that the cars still had AEB.

How is this FUD and anything but Tesla's own fault?

Also, the only evidence we have that AEB still exists is a screen shot, where the option is still on the screen. As we keep getting told, this doesn't mean it's the same AEB as before, maybe it doesn't pass the tests needed to legally call it AEB.
That is not what NHTSA's statement said. NHTSA never actively called Tesla. It was Tesla that contacted NHTSA about a production change with the removal of radar and as a matter of policy, NHTSA does not give check marks for features they have not tested yet on cars with significant hardware changes (using the NHTSA NCAP tests I pointed out, that you previously continually suggested did not exist). Elon just recently said next week the testing will be done, so we'll see very soon.

Goes to show you can spin facts to fit whatever argument you want.
 
Last edited: