I'm wondering how good or bad does v9 need to be for people to be pleased or disappointed. If it is no where near robotaxi good, will it still blow your mind?
Last edited:
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Almost ready with FSD Beta V9
- Thread starter DanCar
- Start date
I’m going to use this and see how it goes.
maybe with my wife?
The force is strong among many on this forum.
Is he though? I've been hearing "Elon has gotten better" for 6 years now from apologists. It's the generic argument made for why it will be better this time. But what's your evidence? Recently Elon said:
We'd have a FSD button
We'd have "vector birds eye view"
The FSD beta would be in wide release by the end of 2020
All superchargers will be powered by carbon free power by end 2021 (5th time Tesla has said this)
So why is he not "not lying, just too enthusiastic" this time too?
Cheburashka
Active Member
Remember Elon said that Smart Summon will blow our minds and the holiday update was fire?
Elon has a long history of lies and mistruths when it comes to this stuff. For all we know, v9 could actually be worse.
Or he holds himself to one thing he said- which is that they won't release v9's pure vision approach until it's safer than vision + radar. And thus, they never release v9.
Personally, software that is not on my car never blows my mind. People argue that Waymo is not ahead because you can't "buy a waymo" and then point to Tesla's super narrow beta as if Tesla is selling any kind of FSD to anyone that wants it. The irony is that if they do release v9 and it's "safer than radar" but they still aren't willing to give it to everyone, they are telling you just how crappy the current stuff is.
Cheburashka
Active Member
Well, if they use v8.2 as the bar for vision + radar, it doesn't seem like it would be too hard to get something that is better. I guess it all boils down to what their definition of "safety" is. If it's anything like Waymo level, then we'll never see it. If it's more in line with "you still have to hold the wheel at all times and drive when the car can't" then I don't see why they won't release it.
mtndrew1
Active Member
I believe Elon is more of a “believes his own BS” guy than a liar.
Either way it’s exasperating.
powertoold
Active Member
It's actually a very intriguing aspect of Elon to me. I run a software business, so I kind of understand the challenges, but yes, I also don't understand why Elon is so loose with his enthusiasm. What is true is that essentially no one in the world can deal with the level of responsibility and detail that Elon does on a daily basis. I can barely handle the responsibility of a small software business, and despite all my strengths and effort, I can't imagine myself handling (mentally / emotionally) 20% of what Elon does.
Elon deals with life and death responsibilities on a daily basis. He also deals with large, powerful, dangerous machines. Imagine looking astronauts in the eye and telling them you're doing everything you can to keep them safe (knowing the immense complexity of rockets and software). It's the same with AP and the safety of Tesla's cars. Ultimately Elon will feel that he's responsible and that he wants the best for his customers. That sort of responsibility will break any normal human down.
To compensate, perhaps Elon likes to imagine a better and "sooner" future.
Last edited:
I may have naive on this, but I bought my Model 3 thinking in the future it can take over my daily commute to work and allow me to rest or do work in the car. So I will feel good about V9 if it shows progress toward that, even if the actual goal won't be achieved for a couple more years.
BrianBigNFun
Giver of snark
I was promised flying cars around this time in the 80's, still trying to process that pain before I move on to FSD
Soda Popinski
Member
I predict FSD wide release will be relatively safe, but overly cautious as a driver. It'll probably miss a lot of turns because it can't get into the proper lane in time, requiring re-routing. It'll probably drive slower than advisable, and be fairly annoying.
There's not though.
Rather than re-write content, read this:
Why LiDAR is Doomed
LiDAR has often been heralded as an essential part of safe autonomous driving, and it is generally an accurate and reliable distance sensor. However, Tesla has proven that cameras can be trained to "see" just as well as LiDAR. What is ultimately required for autonomous driving is not merely...www.voltequity.com
You have to be kidding me voltequity? I have read and debunked that person on twitter several times.
Literally every sentence in that article is either wrong or a balant lie.
They write about Waymo "This means that the trips these cars provide are limited to 30 minutes"... "having to charge the autonomous car for 5 hours after every 1 hour trip."
This is who you are getting your information from? This is absolute none-sense.
Something easily disproved by the fact that the same Waymo can been used for hours.
That's without even talking about the fact that Waymo's compute is made up of Edge TPU that is years ahead of Tesla's FSD.
TPU is vastly more powerful and more efficient than the FSD Computer
100% complete lies. This isn't even opinion, these are straight lies that can be easily proven by downloading any of the pleural of lidar datasets available.
Lidar is used to detect and classify ALL road users and dynamic and static objects. Virtually everything but the color of traffic lights.
They removed the below from their site.
Basic code? Tesla/Karpathy uses these same codes. They use off the self NN architectures. For several years, Tesla was using GoogleNet (inception) from 2014 now they use a slightly modified MS's ResNet 50. I could go on and on.
Waymo/Deepmind/Google Brain on the other hand, creates their own NN architecture. Almost all NN innovation and breakthroughs the last decade came from Google.
"years-old processor." Are you aware how peer review research in ML is done?
I posted the results from the latest per reviewed paper from 2020 comparing lidar to pseudo-lidar.
This is the same paper posted on VE website. The problem is they didn't even read the paper and don't have a clue what its talking about.
Every single SDC is using pseudo-lidar. Its still not close to lidar accuracy especially long distance.
Are you saying that Tesla is training with lidar point clouds? If Tesla is using lidar for anything other than validation, then their entire so-called data "moat" is worthless. They can't use the data from the fleet to train with lidar to give accurate depth.
From @VoltEquity
Basic Arguments Against LiDAR:
Expensive, Useless in Bad Weather, Power Hungry, and Ugly
1) @voltequity says "LiDAR is expensive -- both initial costs and maintenance costs."
This is a flat out lie. Over 10 cars will have one or multiple lidars by end of 2022. Over 5 cars will have lidar(s) by the end of 2021. The lidars going into mass production in 2021 and 2022 cost well under $500. These are the highest resolution lidars.
2) @voltequity says "Furthermore, we must also account for the cost of maintenance. Because LiDAR uses moving parts, it is easier for it to break or malfunction and, thus, it is more expensive to maintain."
This is also false. This is again is using the SOTA of lidar in 2010 to define the SOTA of lidar in 2021. There are 6+ lidars going into production 2021/2022 are not prototype lidars. These lidars are automotive grade. ISO 26262 functional safety compliant, operating and storage conditions of ~-40°-~85°C (~-40°-~176°F) with over 50,000 hours performance reliability (that's over 170 years of human driving), Ingress protection, shocks, etc. There is no maintenance cost. This is pure FUD.
3) @voltequity says "LiDAR is power hungry..decreasing the car’s driving range and thus needing to recharge the car more often."
Lets fact-check this. Automotive grade production lidars use on average around 14 watts. Mobileye's SDC with 3x luminar is 45 Watts or 0.045 kWh. 3) An average EV car uses 3 KWH per 10 miles. Mobileye's Lidars will eat up less than 0.15 miles of range every hour.
This is the classic definition of FUD.
Livox 12 Watts
https://livoxtech.com/horizon/specs
Luminar 15 Watts
https://wired.com/story/lidar-cheap-make-self-driving-reality/…
Innovision under 40 Watts
https://oemoffhighway.com/electronics/sensors/product/21073236/innovusion-inc-innovusion-cheetah-lidar-system…
Innoviz One 15-23 Watts
https://just-auto.com/interview/ces-israeli-start-up-bags-capital-accelerates-lidar-innovation_id192792.aspx…
VoltEquity, is a typical example of using 10 years old outdated tech/info in your argument. Not only has lidar gotten 100x cheaper (compared to 75k in 2010), it has also gotten 100x better. aka FUD.
4) @voltequity says "LiDAR is unable to work well in bad weather conditions. Since LiDAR uses visible lasers in order to measure distance, LiDAR is unable to work well in bad weather conditions, such as heavy rain, snow, and fog -- whereas radar still works in such conditions. LiDAR is essentially blind in bad weather. For this reason, autonomous cars using LiDAR still have to use radar in order to drive in such bad weather conditions, further adding to their costs."
The argument debates devoid into radar is awesome. Funny now that Tesla is not using radar, the new updated page will now say "radar is a stupid crutch". Proving there's no intelligent thought, its just people regurgitating whatever Elon tells them.
However, the argument that "Lidar is unable to work well in bad weather conditions" is also completely false.
5) @voltequity says "4. LiDAR is ugly. While this does not matter from an engineering or safety perspective, it does matter from a consumer’s perspective when choosing a car. There is no way around how ugly LiDAR sensors look when they are placed atop a car."
Two lidars, where are they?
One Lidar?
One Lidar?
Where's the lidar?
Placements depends on the OEM. You don't need a dome like you see on various robot-taxi prototypes.
6) @VoltEquity says "The comparison should be about the full package of sensors each company uses: In other words, how does Tesla’s full package of sensors (cameras, radar, ultrasonics) compare to other companies and their full package of sensors (LiDAR, cameras, radar, maybe ultrasonics)? That is the debate."
When talking about coverage alot of Tesla evangelist lists the 8 camera sensors and mention its 360 degree in comparison to other competitors. But when it comes to Radar they always put something similar to "YES" for both. That's far from true. Its actually.
Company | Radars | Field of View | Resolution | Radar Type | Object Classification | Range | Year
Tesla | 1 | Narrow Forward only | very very low | 2D | No | 150m | 2010
Waymo | 5 | 360 surround | ultra high imaging | 4D imaging | Yes | 300m+ | 2020
7) @VoltEquity incorrectly uses a image of a high resolution imaging radar, which is misinformation because that's not the radar Tesla uses
As you can see what Tesla uses is practically useless.
8) @VoltEquity says "LiDAR, cameras, and radar are poor at detecting objects that are very close, so Tesla uses ultrasonic sensors to help with objection detection at less than 8 meters (= 26 feet)."
This is also wrong. Not only that but ultrasonic are not even a self driving sensors, they are parking sensors for ADAS that have a very high fail rates that's why almost no SDC company uses them. They use surround 4 parking cameras and/or lidars.
I could keep going. But its pointless.
Last edited:
Which year? I'm thinking 2025 for wide release in the U.S. Still ahead of the competition in the U.S. Suspect China will be first due to lawsuit issues. Relatively safe? Safer than an average human? That means a lot of lawsuits and a lot of headlines: "Tesla killed ..." "Tesla controlled robot car maimed ...." "Robot attacked ..." "Tesla robot murdered family of ..."
Last edited:
I was promised world destruction from climate change around this time 15 years ago. Still trying to process that joy before moving on to FSD.
So does the CEO of Airbus, Ford, BMW, Audi, GE, Seimens, Pfizer, the EPA, the NRC.......
You know how they handle that? By being conservative and listening to the data and their HUNDREDS OF THOUSANDS OF EMPLOYEES. Not by tweeting hopes and dreams and telling people they will have cancer cured by the end of this year because he was "too enthusiastic in the moment."
I'm sick of the story that Elon is the Technoking of all of this crap and shoulders it all himself. You think he's ever done a FMECA? You know what I want out of my safety critical CEO's? Having enough self control to critically look at the situation and work carefully, not taking an Ambien and some whiskey and going on a tweet bender. And to actually believe in science, not deny the existence of COVID. You actually think he deeply cares about human safety? You think he worries about it more than the actual woman who designs airbag algorithms or the dude that designs thermal control systems for batteries and knows that statistically, some day, some where, a decision they made will likely injure or kill someone?
And then let's just forget about medical doctors that literally hold people's lives in their hands.
Last edited:
These people are far gone. To them Elon is their God that they pray to, infallible, omnipotent and omniscience. Facts doesn't matter to them. All they see is elon.
powertoold
Active Member
These other CEOs don't "own" their products (they didn't found the company; they're mostly there to push the bottom line). I know how this feels because I own my software product (b2b, workflow critical software), and the sense of responsibility is different. It's like being a nanny vs your own child.
You've clearly never worked with safety critical systems. The idea that people working on fly by wire systems or radiation dosing machines don't feel "ownership" of their products and don't deeply worry about the harm they can cause if they fail is insanely disrespectful to the hundreds of thousands of engineers and others around the world that work hard every day to make the world safer and have deep ethical and emotional considerations to deal with every line of code they write or ciruit they design.
I've worked in aerospace with CEO's that do build safety critical products, and they do very much "own" their products as well. You tend to do that when you, your family, and your friends will be the ones flying on those airplanes.
Tell me, how is Elon's job more responsible or more "life and death" than a brain surgeon?
You know what Elon said when asked about the risk of COVID? "Everybody dies." Yep, the mark of a true humanitarian that cares about human safety and bears the weight of it more than any other human.
Last edited:
Similar threads
- Article
