Elon: FSD Beta tweets

[EVNow]

I like how you left it ambiguous whether you were talking about your own statement or the one you quoted. Clever; it's like recursion.

Esteemed members of this forum no doubt can figure out that it is Elon's statement that has been misreported and abused by haters / TSLAQ like OP for years

 

[Terminator857]

I wonder if this is due to actual improvement in Tesla Vision, or they just had too many complaints from Elon? ...

Somewhere I read that the minimum distance will increase at higher speeds. Seems like a smart move.

 

[diplomat33]

https://twitter.com/x/status/1526685434910212102

Looking forward to hearing about how they broke through the local maxima. Maybe robotaxi prototype will be shown?

I think AI Day #2 will focus on NN they are working on, like vector space or behavior prediction on vision-only. Elon and Karpathy will talk about how they are solving difficult AI problems and encourage talent to apply. Elon might tease some updates on Tesla Bot. I think they will also show or tease HW4.

Personally I find these events a bit frustrating. On one hand, I really like the technical stuff and learning about the machine learning that Tesla is doing. But these events also create a lot of hype that gets me more excited and then I want to know when will we get FSD that does not require driver supervision. Then reality settles in when I take my Model 3 for a spin and FSD Beta does something dumb. Plus, I can't even try FSD Beta 10.12 yet because my Model 3 is still in the body shop. Honestly, I would almost prefer Tesla not do these events and just tell me when FSD Beta does not require driver supervision anymore. Stop teasing me if it is still going to be awhile before FSD is actually L4.

 

[DukeofURL]

Personally I find these events a bit frustrating. On one hand, I really like the technical stuff and learning about the machine learning that Tesla is doing. But these events also create a lot of hype that gets me more excited and then I want to know when will we get FSD that does not require driver supervision.

It's very exciting to get a peek at the cutting edge work they're doing, but I agree it's harsh to step back into the reality that we won't see it for 2-3 years. All in all, I like to know what I can expect (or at least hope for), and set a general level of expectation better than "3 months maybe, 6 months definitely".

 

[EVNow]

Can you clarify what you mean by misinformation? Seems like he literally did say that

https://twitter.com/x/status/1148070210412265473

See the first tweet. "Capability" - explained in the annual general meeting that year as well.

ps : And has been debated to death here.

 

[JupiterMan]

But these events also create a lot of hype that gets me more excited and then I want to know when will we get FSD that does not require driver supervision. Then reality settles in when I take my Model 3 for a spin and FSD Beta does something dumb. Plus, I can't even try FSD Beta 10.12 yet because my Model 3 is still in the body shop. Honestly, I would almost prefer Tesla not do these events and just tell me when FSD Beta does not require driver supervision anymore. Stop teasing me if it is still going to be awhile before FSD is actually L4.

Now imagine having purchased FSD and living in the UK watching all this ...!

not complaining though.. love watching the progress from afar!

 

[DrChaos]

Stop teasing me if it is still going to be awhile before FSD is actually L4.

I think L4 is very doubtful ever on current sensor set and lack of semantic mapping.

Vision-only is a mistake. The problem with the old radar was that it was too low resolution, not that it was radar. There are much better radar chipsets now with AESA like scanning & imaging capability (and will be less expensive and more reliable than lidar). For optical data, they should use higher resolution cameras (current ones are worse than eyes by quite a bit, they're like 10 year old cell phone cameras, about iphone 4 level) and in stereo.

What Is 4D Imaging Radar?

4D imaging radar is high-resolution, long-range sensor technology that offers significant advantages over 3D radar, particularly when it comes to identifying the height of an object. This technology is important in the development of advanced driver-assistance systems (ADAS) for some Level 2 and...

www.aptiv.com

Vayyar Imaging - Home - Vayyar

One sensor.Infinite opportunities One sensor.Infinite opportunities One sensor.Infinite Opportunities. We started with a vision of detecting cancer within breast tissue Vayyar is developing a safe, affordable and portable way of screening for early-stage breast cancer using radio frequency...

vayyar.com

RFISee – Developing the world’s first 3D imaging-radar-on-1-chip, providing the next generation ears and eyes for autonomous vehicles, drones, robots etc.

rfisee.com

4d Imaging Radar - Arbe

Arbe is the world’s 1st company to demonstrate ultra-high-resolution 4D imaging radar with post-processing and SLAM. 100 times more detailed than any other system in the market, Arbe has repositioned radar from supporting camera and Lidar to the backbone technology of the automotive sensor suite.

arberobotics.com

There seem to be a bunch of Israeli startups---there is strong expertise in military radar and signal processing and it seems to be filtering to commercial applications.

 

[D Good]

So whatever happened to 10.12?

 

[sleepydoc]

I think L4 is very doubtful ever on current sensor set and lack of semantic mapping.

Vision-only is a mistake. The problem with the old radar was that it was too low resolution, not that it was radar. There are much better radar chipsets now with AESA like scanning & imaging capability (and will be less expensive and more reliable than lidar). For optical data, they should use higher resolution cameras (current ones are worse than eyes by quite a bit, they're like 10 year old cell phone cameras, about iphone 4 level) and in stereo.

What Is 4D Imaging Radar?

4D imaging radar is high-resolution, long-range sensor technology that offers significant advantages over 3D radar, particularly when it comes to identifying the height of an object. This technology is important in the development of advanced driver-assistance systems (ADAS) for some Level 2 and...

www.aptiv.com

Vayyar Imaging - Home - Vayyar

One sensor.Infinite opportunities One sensor.Infinite opportunities One sensor.Infinite Opportunities. We started with a vision of detecting cancer within breast tissue Vayyar is developing a safe, affordable and portable way of screening for early-stage breast cancer using radio frequency...

vayyar.com

RFISee – Developing the world’s first 3D imaging-radar-on-1-chip, providing the next generation ears and eyes for autonomous vehicles, drones, robots etc.

rfisee.com

4d Imaging Radar - Arbe

Arbe is the world’s 1st company to demonstrate ultra-high-resolution 4D imaging radar with post-processing and SLAM. 100 times more detailed than any other system in the market, Arbe has repositioned radar from supporting camera and Lidar to the backbone technology of the automotive sensor suite.

arberobotics.com

There seem to be a bunch of Israeli startups---there is strong expertise in military radar and signal processing and it seems to be filtering to commercial applications.

From what I understand part of the problem with radar is not just resolution but also calibration with the other sensors. If you can’t reliably calibrate two sets of input it’s hard to use them together.

 

[diplomat33]

From what I understand part of the problem with radar is not just resolution but also calibration with the other sensors. If you can’t reliably calibrate two sets of input it’s hard to use them together.

Radar vision calibration is not a problem. All the major AV companies do it.

 

[Terminator857]

... Stop teasing me if it is still going to be awhile before FSD is actually L4.

Is he still saying L5 by next year? I think hoping for L3 is better regardless of what Elon says. L3 in stop and go traffic on limited access freeway. Do we need pothole detection to work first? I'm optimistic mobileye is going to have mass L3 by 2025, so Elon has 3 years, otherwise Tesla will be behind. @Bladerskb If Merc actually does L3 like they say they are going to do (this year in the U.S.?) then I'll be tempted to buy a merc.

 

[momo3605]

Is he still saying L5 by next year? I think hoping for L3 is better regardless of what Elon says. L3 in stop and go traffic on limited access freeway. Do we need pothole detection to work first? I'm optimistic mobileye is going to have mass L3 by 2025, so Elon has 3 years, otherwise Tesla will be behind. @Bladerskb If Merc actually does L3 like they say they are going to do (this year in the U.S.?) then I'll be very tempted to buy a merc.

I don't recall when was the last time he said L5 by next year. Recently the main thing i've heard Elon say is FSD Beta by end of year. We all know that means at best it'll be a summon like release. Shaky L2 driver assistance that most people won't use when they realize it's not good. (At least that's a realistic goal lol)

 

[diplomat33]

I'm optimistic mobileye is going to have mass L3 by 2025

Mobileye is saying mass L4 by 2025 on the Zeekr 001.

 

[EVNow]

Mobileye is saying mass L4 by 2025 on the Zeekr 001.

Really ? Earlier they were saying L4 by 2024.

CES: Mobileye, Zeekr Collaborate on Consumer AV

Together, Mobileye and Zeekr aim to deliver the world’s first consumer autonomous vehicle with L4 capabilities by 2024.

www.intel.com

With an anticipated debut in China in 2024, the new all-electric Zeekr vehicle is expected to be the world’s first consumer AV with L4 autonomous capability.​

 

[diplomat33]

Really ? Earlier they were saying L4 by 2024.

CES: Mobileye, Zeekr Collaborate on Consumer AV

Together, Mobileye and Zeekr aim to deliver the world’s first consumer autonomous vehicle with L4 capabilities by 2024.

www.intel.com

With an anticipated debut in China in 2024, the new all-electric Zeekr vehicle is expected to be the world’s first consumer AV with L4 autonomous capability.​

I stand corrected. Thanks. They plan to debut consumer L4 in China in 2024. If they are successful, we could see mass L4 by 2025.

 

[CaptUAL]

So whatever happened to 10.12?

They are trying to get the cameras to work....any month now..

 

[2101Guy]

speaking of Elon tweets. Horn via airbag on refresh models. "Just a firmware update". Have we heard any more on this promise as well?

 

[EVNow]

If they are successful, we could see mass L4 by 2025.

Not sure what you mean by "mass". Are you talking 10k / 100k / million cars in 2025 ?

 

[diplomat33]

Not sure what you mean by "mass". Are you talking 10k / 100k / million cars in 2025 ?

"mass" means wide spread, large numbers. I don't know how many cars Zeekr is planning to produce in 2025. So it is hard to put a number to it. But certainly, when Zeekr starts mass producing consumer L4 where they are making like thousands per day, I would probably consider that "mass".

 

[EVNow]

"mass" means wide spread, large numbers. I don't know how many cars Zeekr is planning to produce in 2025. So it is hard to put a number to it. But certainly, when Zeekr starts mass producing consumer L4 where they are making like thousands per day, I would probably consider that "mass".

2021 EV deliveries in China by the numbers: How some of the market's major players grew

A breakdown of 2021 EV deliveries from some of the growing and best known automakers currently developing vehicles in China.

electrek.co

With about 6,000 deliveries of its 001 sedan in 2021 (deliveries began in late October), Zeekr is aiming for 70,000 in 2022 and 650,000 by 2025.

Yes, by any measure that would be mass production. If Tesla can't get FSD to work by then, Zeekr can pull it off, MobilEye would have beaten Tesla to mass produced self-driving consumer cars with wide ODD.