Fully agree. I just wish Musk could do the same thing with the same nuance. This all started with my comment that it's really not the sign of someone who has deeply thought about "safety" to say both "be paranoid" about this beta L2 system we are giving you and "safety is our top priority" in the same tweet.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
OFFICIAL BUTTON WATCH
- Thread starter Drrobmiller
- Start date
Saying I am familiar with ISO 26262 and how you analyze safety means that I "clearly" stated my option matters more than others?
And people say I over-read into Elon's communications!
Anyway, I will agree I have become over invested in this as well. I've stated my options, probably poorly at times, and probably too much. Off we go to waiting for tonight, and hopefully V9 is an amazing, clear step on the path to L3.
Last edited:
powertoold
Active Member
Only 11 more hours folks, hold your horses
Back to the topic described at the title of this post: Tonight, who's supposed to received FSD Beta V9? Only the EAP crowd, right? Not general public, us mere mortals? Just so I align my expectations. Not sure if I'll be happy with a new UI in my Car screen on in my Computer screen. haha
FloridaJohn
Member
Not even the EAP crowd. Only the internal Tesla employees and approximately 70 actual customers. Same group that was previously testing FSD.
So, I've seen that 70 number thrown around a lot. What was the source on that? Not that I don't believe y'all I'm just curious of the origin.
FloridaJohn
Member
From the memo of the meeting with the California DMV and the Tesla AP team.
Tesla is more cautious about Full Self-Driving timeline with regulators than the public, based on new memo
YouTube stars must be drooling. Their revenue these past few months has surely suffered.
Knightshade
Well-Known Member
Also you-
So which is it?
Do they need to test in a lot of situations and then make adjustments and updates to fix issues in those situations-- or is bringing up different situations an "excuse"?
100% nonsense.
For one, adjusted for sales, Tesla is involved in the fewest fatalities of major car brands.
The only ones that beat it are brands rarely driven (and basically never in hard situations- stuff like Ferrari, Bugatti, etc)
For another if making more cars was their primary goal they'd use smaller batteries and not spend so much effort designing them to have the lowest chance of occupant injury of any car on the market
Sometimes you make excellent points man,
Other times I wanna ask you to show us Where On the Doll Elon Touched You.
Cheburashka
Active Member
Elon could put half these arguments to rest by giving us the button which he promised. Instead he took our money and all we got for it is some YouTube videos and a green light chime.
Daniel in SD
(supervised)
They must think that beta FSD is unsafe when used by the broad public (or will be perceived by the public and regulators as unsafe). What I don't understand is what is going to change. What is the criteria for release? I don't see how they're going to make many city maneuvers easy enough to monitor.
rxlawdude
Active Member
No, they must want to ensure it is as safe as possible; thus limited numbers of BETA testers initially.
Daniel in SD
(supervised)
They say they believe it will be possible to achieve unsupervised safety greater than the average human. Are they waiting for that?
Many people around here think they will do a broad release before that. How do you measure safety when supervised by untrained customers? If you do believe it's safe, how will you convince the public?
I'd like to see where this data comes from. I searched myself, and the latest IIHS data is from 2017, and Tesla is not listed at all. Other searches all included lists without Tesla at all. Most of these are by car model, not by brand though, and brand is kind of a cheat when Tesla has so few vehicle designs and other manufacturers have 20+.
Tesla has had a fatality in every single model they sell. 88 in total in fact. However, there are 7 vehicles from 2017 that have had *zero* fatalities in the USA. One of those is the VW Golf, a small, popular car. The Nissan leaf has had only 5 fatalities per million registered vehicle years. Tesla, by my quick math, has had about 15 vehicle deaths per million registered years. That's not even in the top 20. But that's a guess from the numbers I know, so I await the better analysis.
As always, never saying a Tesla is not a very safe car. Saying that safety is not always their top priority as a company, and now questioning a bit that they are the unquestionable safest cars sold since you brought it up.
Safety is measured in rate. Incidents per mile, hour, exposure, use, etc.
You do not achieve safety by only giving it to a limited number of people. Systems do not change their level of safety because of how many people are using them. The system is the system, and each copy of it is independent from the others. You would just have fewer instances per calendar time, but also less users and benefit, so there is no net change.
You do lower overall human risk by exposing less people to an unknown system.
However, if you think you can estimate your level of exposure from 71 drivers over 1 month, you are either bad at statistics, or your target level of safety is remarkably low. All you are doing with 71 people is making sure it's not so atrocious that you'll immediately have bad PR if you were to release it more widely. Hence the NDA.
As a simple rule of thumb FYI, you use the rule of three. Example:
If we have 71 drivers, and they each drive 2000 miles on City Streets Autosteer. None of them have accidents. That's 142,000 miles with no accidents.
You divide the 142,000 by three- 47,000. You now have a 95% confidence that your accident rate is 47,000 miles or more. But you have no idea what the actual rate is. It could be 47,001. It could be 17M. In fact, you can never really know what your accident rate is until you have some accidents. But you also don't need an accident if you can get enough data to go beyond your safety target.
In order for Tesla to have a good sense their rate is 1:2.05M or more (better than a human in a Tesla with no AP by Tesla's own numbers), Tesla needs at least 6M miles with no accidents. That's only 85,000 miles per beta tester, all on city streets autosteer (no highways).
They better get crackin', it's going to be a while.
Then, once you're done with that, you now have to deal with the fact that your test was synthetic, because by your own test design, you picked expert and safe drivers. Your normal user population is not this, so your data does not apply. You have no idea what the rate will be when given to your broad user base. This is like testing a new drug on 18 year old men exclusively, and then once they have no adverse reactions claiming that it's safe for pregnant 40 year old women.
Now, you might say that Tesla has 2,071 testers, not just 71. We know those 2,000 testers are Tesla employees. By definition, they are concentrated near Tesla sites. This means their data is less useful than truly broad, random data. You cannot have 2,000 people test autonomy in/near Fremont, and claim that you have any sense of the incident rate to use in Denver.
So basically, these 71 or 2,071 drivers tell you very little about your system if your goal is 1:2M miles. They tell you a LOT if your goal is 1:1000 miles. There's a reason aerospace uses simulation and design assurance, not real world testing to demonstrate safety. Doing it in the real world is unaffordable when your goal is 1:1B hours.
Last edited:
rxlawdude
Active Member
You can't fix stupid. People use beta versions without heeding all the cautionary statements.
Let's see what beta 9 looks like.
