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

FSD Beta Videos (and questions for FSD Beta drivers)

This site may earn commission on affiliate links.
#FSDBeta 6 version 2020.44.15.4
This is a demonstration of some very difficult unprotected left turns. The results were not very impressive, but realize I am putting the software though very challenging situations.


Basically failed every scenario, which is understandable given the difficulty. The problem is how it’s failing: it shouldn’t be stopping in the highway. It’s failure mode should be to just sit at the stop sign and wait for a big enough gap both ways or ask for user intervention instead of creeping forward into the highway slowly.

Tesla might be better off just trying to avoid unprotected left turns like this for now. Adjust their FSD nav to just turn right and make a U, when they enable U’s, or just reroute around it. Or, if that’s too difficult, throw up warnings and ask for user intervention.

edit - 4:36 - nice improvement at the end where it went into the correct lane where in a previous vid I saw it start to turn into the opposing lane instead before Mr Cook took over.

edit 2 - would be curious to know what mspisars disagreed with.
 
Last edited:
  • Disagree
Reactions: mikes_fsd
Any input on the format for the videos? Any recommended scenarios you would rather see? Some people are doing whole drives .. some people are doing voice overs, some are doing a lot of tagging and post production .. I feel like more shorter and less produced content is valuable.. but I am open to input.

Edit: I will do what I can with the time I have. This isn't my day job, but its still fun for me.
 
I think you’re being extremely optimistic. From what I’ve seen, the release is nowhere close to being ready for a public release. It is EXTREMELY IMPRESSIVE, but I liken it to my home full of Amazon echos, being extremely impressive 9/10 times makes that one time it ****s up even more annoying because you come to expect it to be impressive and then one time instead of the fifteen minute timer you thought you set, instead it set a fifty minute timer and your chicken gets over cooked and dry to the point of inedibility. The same holds true here, IMO, where ironically being extremely impressive 99/100 times makes the one time the system ****s up more annoying, but instead of setting an inaccurate timer or playing the wrong song, it’s the driver not paying attention while their five thousand pound vehicle does something wrong while going at speed. The consequences here are a lot more serious. 99/100 means at least a small percentage of people will start to take the system for granted and we have a FSD BETA accident, or god forbid a fatality and a deluge of public relations nightmares comes down on Tesla.

A limited release let’s Tesla personally vet people and personally emphasize that they’re all testers and to be aware of the dangers and limitations of the system over the phone and then have those same people sign a bunch of documents also emphasizing the risks of using the system like Mr Cook said they did with him. By having everyone do that, it’s less likely people will take the system for granted. It’s easier for them to control a small population to try to minimize the chances of an individual user misusing the system. If they did a wide general release, then it would be very difficult to impossible for Tesla to do the same for everyone who would receive the update, and having a pop up or even a number of pop ups come up on the screen before enabling the system where the user has to confirm doesn’t have the same effect, as anyone who has ever skipped past the dozens of EULAs everyone who installed software/apps on their computer/smartphone can attest. I dare say the vast majority of people don’t read the text of the EULAs and just hit confirm.

I hope you’re right, because I do want the system on my car. I’ve delighted in the progression of visualizations on my car and I use AP at nearly every opportunity, but my biggest fear of Tesla doing a quick general wide release would be a quick, enforced clawback a few months down the road and huge scrutiny permanently gimping the capability of my car because of some tragedy. I just don’t see it happening because I think Tesla well understands the risks of public over-reaction and the potential for government crackdown, and will be very sure to do everything they can to perfect the system before a wide release, which likely means multiple waves of slow increases of users as they stop getting enough relevant data from current beta testers to occupy their team before any kind of wide release, which is likely to take a great deal more time than by the end of the year.
I don't disagree with you regarding your analogy at all. I think you're spot-on.

However, the current AP implementation is EXACTLY like this. Maybe even worse. You mention that you use AP every change you get. I do too. And Tesla allows us to do this but includes disclaimers in the user manual explaining the risks that we are taking.

I see FSD as an improved extension of this. At the end of the day, the driver is responsible for ensuring that the car is operated safely. And it will continue to be this way until they remove pedals and steering wheels from the cars.

I understand that FSD is more capable than the current AP. And I can understand how someone could make the argument that increased capability results in increased risks. But at the end of the day, whether the car is simply trying to keep it between the lanes or whether it's trying to make turns for you, the drive always has to be in control. That's no different in AP than it is in FSD.

If they plan on waiting to roll this out until they've got all the kinks worked out then I agree with you - we won't see this for another year or more. I just don't see them doing that based on their history. Look at Advanced Summon in its current state. That thing is no where NEAR perfect yet they rolled it out. And for that, you aren't even IN the car!
 
Any input on the format for the videos? Any recommended scenarios you would rather see? Some people are doing whole drives .. some people are doing voice overs, some are doing a lot of tagging and post production .. I feel like more shorter and less produced content is valuable.. but I am open to input.

Edit: I will do what I can with the time I have. This isn't my day job, but its still fun for me.
I think you're doing a great job. If you do a whole drive then definitely speed up the boring parts.

I think that Teslatino and AIDRIVR are doing great stuff with their recent content. Multiple styles can work.

But I DO agree with you - short, action-packed is way better than long videos with dead spots.
 
I see FSD as an improved extension of this. At the end of the day, the driver is responsible for ensuring that the car is operated safely. And it will continue to be this way until they remove pedals and steering wheels from the cars.

Today's FSD beta testers have to sign waivers and essentially agree to the greater risk. While that is not practical with a wide release, I began to think about what Tesla could do, such as having you "sign" your name on the screen each time you start your car with AP enabled. Having the camera for driver monitoring could also be used, especially initially, to put any inattentive drivers in "jail", maybe no autopilot for the rest of the day, or something severe enough that it gets your attention.

I think there could be some ways to better enforce attentiveness. I admit I am more inattentive on the freeway, but rarely so when on city streets. I know that is not good enough, especially when you can be inattentive and still keep driving.

I do hope though that this comes sooner than later for those drivers willing to be additional beta testers.
 
  • Like
Reactions: Chazman92
Today's FSD beta testers have to sign waivers and essentially agree to the greater risk. While that is not practical with a wide release, I began to think about what Tesla could do, such as having you "sign" your name on the screen each time you start your car with AP enabled. Having the camera for driver monitoring could also be used, especially initially, to put any inattentive drivers in "jail", maybe no autopilot for the rest of the day, or something severe enough that it gets your attention.

I think there could be some ways to better enforce attentiveness. I admit I am more inattentive on the freeway, but rarely so when on city streets. I know that is not good enough, especially when you can be inattentive and still keep driving.

I do hope though that this comes sooner than later for those drivers willing to be additional beta testers.
But they don't have any of this today and we can use AP just about anywhere. So I don't really see the difference with FSD except that it appears to actually perform BETTER than AP today when used on city streets.
 
But they don't have any of this today and we can use AP just about anywhere. So I don't really see the difference with FSD except that it appears to actually perform BETTER than AP today when used on city streets.
Better != safer.
It obviously does unprotected left turns better than the released software. It sure doesn't look safer yet than just doing it yourself. The only way it might be safer now is that it puts you in a hypervigilant state because you fear for your life.
 
  • Like
Reactions: AlanSubie4Life
Any input on the format for the videos? Any recommended scenarios you would rather see? Some people are doing whole drives .. some people are doing voice overs, some are doing a lot of tagging and post production .. I feel like more shorter and less produced content is valuable.. but I am open to input.

Edit: I will do what I can with the time I have. This isn't my day job, but its still fun for me.

I enjoy your videos a lot and appreciate that they aren’t just cruising down streets that I can do with my current autopilot.

I would find it handy if you could make a number of “named” intersections that fail now so we can keep track of over time as the car becomes more capable.

Steven Peeters does this in his autopilot progress videos and it’s been fun watching his specific locations go from failing completely to working perfectly over the years.

So maybe identify the really tricky stuff and then make note of the progress over time on a few of them so we can compare releases.
 
Any input on the format for the videos? Any recommended scenarios you would rather see? Some people are doing whole drives .. some people are doing voice overs, some are doing a lot of tagging and post production .. I feel like more shorter and less produced content is valuable.. but I am open to input.

Edit: I will do what I can with the time I have. This isn't my day job, but its still fun for me.
I definitely enjoy your videos. Well done with editing. Could you possibly ask the system to navigate you to Gap and maybe grab a pair of denims? I realize it’s Florida and all. But...
 
  • Like
Reactions: 82bert
I could be completely wrong. But I suspect that a small release population has more to do with allowing them to do some hand-to-hand combat before rolling this out more broadly to let the system to do the work.

I still expect a release within 3-8 weeks.
Not how it works.

# of outside testers is directly proportional to the maturity of the sw / confidence of the engr team.

When Musk recently talked about releasing FSD before end of 2021, I think he was talking about FSD Beta going GA - not robotaxi (that he always says is 1 year away).
 
Better != safer.
It obviously does unprotected left turns better than the released software. It sure doesn't look safer yet than just doing it yourself. The only way it might be safer now is that it puts you in a hypervigilant state because you fear for your life.
I hear ya. But my comment wasn't that it's better/safer than doing it on your own. It was that it's better/safer than using the current AP software. If they'll release that thing to the wild then I don't see the downside of releasing the current FSD to the wild either.
 
Not how it works.

# of outside testers is directly proportional to the maturity of the sw / confidence of the engr team.

When Musk recently talked about releasing FSD before end of 2021, I think he was talking about FSD Beta going GA - not robotaxi (that he always says is 1 year away).
You may be right. I read his comment to mean actual L5.

My desire to have FSD in my vehicle is likely clouding my thoughts though.

For me personally.... the current iteration of FSD appears to run circles around AP for city driving so there's only upside to having it in my vehicle.

I'm admittedly an edge case though. I'm perpetually an early adopter and lover of all things tech-related. I understand and appreciate the baggage that comes with that. Therefore I expect FSD to make mistakes and I'm ready for it. If the general public expects to read magazines while FSD dries them to work then we will have major issues.
 
  • Like
Reactions: jayman
Haha.. ok put long pants on.. point taken.
....and I don't like Khaki's so could wear something more stylish and befitting of a FSD Beta tester? Example of the proper Youtuber attire below.:eek:o_O:D:D

Screen Shot 2020-12-03 at 6.58.48 PM.png

Screen Shot 2020-12-03 at 6.57.35 PM.png
 
It was that it's better/safer than using the current AP software.

But you'd never use the current AP software to make unprotected left turns, so that means the current AP software is likely safer when talking about that particular use scenario.

They're not going to release this to the wild any time soon. 3 months definitely (not), 6 months maybe (probably not).
 
  • Like
Reactions: HighZ
Ok .. I would wear that if you sent it too me.. but other than that.. dunno how to make that happen.. So I got a no on Khaki's .. get some Gap denim .. but definitely no shorts..

Geez .. back to the video editor.

EDIT: By the way those are my beekeeping trousers to keep the honeybees off me. Im a beekeeper.
 
  • Like
Reactions: byeLT4