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

FSD V9 First Impressions - General Public Access Seems Way Off

This site may earn commission on affiliate links.
Hey all -

Curious to get everyone's first impressions of V9 Beta and timeline release for general public based on what they have seen on youtube? Me personally - After what I have seen, "A month or so" as Elon has described looks to me like it could be a solid 6 months out before this is released to the public. There just seems to be so many edge cases and craziness with city streets that if I was Tesla legal team in charge of managing risk, I would be shaking in my boots knowing this is going out to the public in a month. The number of interventions people have had to take in these videos looks to be a major liability once this goes out broadly. Thoughts?
 
Hey all -

Curious to get everyone's first impressions of V9 Beta and timeline release for general public based on what they have seen on youtube? Me personally - After what I have seen, "A month or so" as Elon has described looks to me like it could be a solid 6 months out before this is released to the public. There just seems to be so many edge cases and craziness with city streets that if I was Tesla legal team in charge of managing risk, I would be shaking in my boots knowing this is going out to the public in a month. The number of interventions people have had to take in these videos looks to be a major liability once this goes out broadly. Thoughts?

For me, FSD Beta V9 is an ok L2 "door to door" that is intended for suburban areas, low traffic, well marked intersections with traffic lights. It is similar to NOA but for non-highways. Just as NOA is intended to automate steering to navigate you on the highway from on ramp to off ramp, FSD Beta is intended to automate steering on non-highways to get you from A to B. In ideal conditions, it will do all the steering and braking and you just supervise and confirm certain decisions with the stalk. But for more complicated situations like narrow city streets and downtown city driving, FSD Beta will do poorly. FSD Beta is definitely not an autonomous driving system. You must pay attention. And there are situations it cannot handle at all.

In terms of a general release, I think we could see a general release in 6 months to a year. When they do release it to the general public, I think they will nerf it. I think they will keep the stalk confirmations and may also geofence it to only certain types of roads or reroute to avoid certain cases like a difficult unprotected left (like we saw in one video). I expect FSD Beta will be a L2 system for a long time.
 
I don't think reducing the number of interventions will make the system safer. For example a system that drove perfectly except that every 10k miles it hit a pedestrian would be incredibly unsafe because people would get complacent and not pay attention while using it.
Obviously if they actually achieve superhuman capability that's completely different but that doesn't look likely for the foreseeable future.
I honestly have no idea how they're going to release it. I would love to be a fly on the wall and find out what their criteria is.
 
  • Like
Reactions: Matias and thedm96
Hey all -

..., "A month or so" as Elon has described looks to me like it could be a solid 6 months out before this is released to the public. There just seems to be so many edge cases and craziness with city streets that if I was Tesla legal team in charge of managing risk, I would be shaking in my boots knowing this is going out to the public in a month.
Elon doesn't listen to lawyers. He does what he wants. It is going to be optional and perhaps not recommended with plenty of disclaimers. Tesla is not going to recommend people use this.
The number of interventions people have had to take in these videos looks to be a major liability once this goes out broadly. Thoughts?
Keeps people attentive as is needed. This is not a system for dummies. This is a system for people who will place constant and close attention. Driving is that way, but some people don't know this.
 
  • Disagree
Reactions: gearchruncher
Yes, but how would you keep this out those people's hands? How do you ensure that only people who will place constant and close attention will use the system?
By making them acknowledge the requirement for them to be vigilant and ready to take over at any moment.

You can't fix stupid people who, in spite of warnings, do dumb things.
 
...do dumb things and then sue for not stopping them. Though, I think today's influencers would gladly trade the views and likes for crashing and getting hurt. Maybe even dying.

By making them acknowledge the requirement for them to be vigilant and ready to take over at any moment.

You can't fix stupid people who, in spite of warnings, do dumb things.
 
By making them acknowledge the requirement for them to be vigilant and ready to take over at any moment.

You can't fix stupid people who, in spite of warnings, do dumb things.
But you can regulate vehicles that are used on public roads.
I think what Tesla is doing with FSD beta is violating the regulations for autonomous vehicle testing in California. I suspect the DMV is giving them a very long leash because of their clout in the state and because there haven't been any high profile accidents. I would like Tesla to make a convincing case that FSD beta doesn't reduce road safety (when used by real people, not hypothetical people) before releasing it to the general public.
It will be interesting to see what happens, the suspense is killing me. haha.
 
By making them acknowledge the requirement for them to be vigilant and ready to take over at any moment.

You can't fix stupid people who, in spite of warnings, do dumb things.
If that is the solution, then prepare for a lot of bad publicity for Tesla. I would also assume some unnecessary fatalities, and maybe some restrictions placed on Tesla by governing bodies. We certainly aren't going to get to "safer than most humans" with that plan.
 
Yes, but how would you keep this out those people's hands? How do you ensure that only people who will place constant and close attention will use the system?
There will eventually be some darwin award winners. Perhaps someone will get roasted alive trying to show off the system to youtube. Maybe the headlines and videos will keep people scared enough. On the other side of the argument: Tesla will publish stats showing how many lives the system has saved and youtubers will publish videos of how the system miraculously saved their lives.
 
I suspect the DMV is giving them a very long leash because of their clout in the state and because there haven't been any high profile accidents.

I would like Tesla to make a convincing case that FSD beta doesn't reduce road safety (when used by real people, not hypothetical people) before releasing it to the general public.
It will be interesting to see what happens, the suspense is killing me. haha.
How many real people use current FSD functions on highways (you know, where that feature is supposed to be used), and how many fatalities in the past 5 years directly attributed to FSD/AP/EAP?

I'll take those odds anytime. But then again, I read and follow instructions and heed warnings.
 
There will eventually be some darwin award winners. Perhaps someone will get roasted alive trying to show off the system to youtube. Maybe the headlines and videos will keep people scared enough. On the other side of the argument: Tesla will publish stats showing how many lives the system has saved and youtubers will publish videos of how the system miraculously saved their lives.
Maybe. As a counter example we can look at Uber's self driving development program that involved a trained safety driver. That did not help further Uber's goals towards self driving robotaxis. When cars start causing preventable deaths, eventually the government steps in.
 
How many real people use current FSD functions on highways (you know, where that feature is supposed to be used), and how many fatalities in the past 5 years directly attributed to FSD/AP/EAP?

I'll take those odds anytime. But then again, I read and follow instructions and heed warnings.
There haven't been nearly enough miles driven with beta FSD to generate conclusive safety data. I look forward to Tesla providing it when there is.
Their safety data for regular Autopilot isn't very convincing to me for reasons I'm sure you've seen discussed a million times on this forum.
 
  • Like
Reactions: gearchruncher
Now we're using Uber as a proxy for Tesla? SMH.
They're both operated by humans. Why couldn't the Uber AV testing fatality occur in a Tesla?
That single fatality pushed Uber's safety record to far worse than the human average of 1 per 100 million miles. Obviously it could be a fluke or it could be evidence that testing autonomous vehicles on public roads is relatively dangerous and needs strict oversight by manufacturers.
 
Im even just talking about the short term (say, within a year). FSD level 5, robo taxi's, even level 4 is still a ways out. FSD features on the highway work great and there havent been mainly high profile accidents. However, city and streets introduce

A. Pedestrians
B. Unprotected left turns
C. Right on red - with pedestrians crossing
D. All other edge cases

I just cant see Tesla releasing this beta to the general public anytime in 2021. Happy to hear arguments differently. If they do, they open the flood gates to high profile accidents that could set them back years. In some ways, at some point, someone needs to lead the way in this. But I cant think of a scenario they give general public access to this beta in 2021.
 
  • Like
Reactions: Matias and daktari
Im even just talking about the short term (say, within a year). FSD level 5, robo taxi's, even level 4 is still a ways out. FSD features on the highway work great and there havent been mainly high profile accidents. However, city and streets introduce

A. Pedestrians
B. Unprotected left turns
C. Right on red - with pedestrians crossing
D. All other edge cases

I just cant see Tesla releasing this beta to the general public anytime in 2021. Happy to hear arguments differently. If they do, they open the flood gates to high profile accidents that could set them back years. In some ways, at some point, someone needs to lead the way in this. But I cant think of a scenario they give general public access to this beta in 2021.
For anyone, myself included, that has FSD (not apart of EAP), I wouldn't get your hopes up anytime soon. "A month or so" seems like a pipe dream
 
Keeps people attentive as is needed. This is not a system for dummies. This is a system for people who will place constant and close attention. Driving is that way, but some people don't know this.
To be a useful driver assist feature, you have to be able to give it some trust. Autopilot on freeways is fairly unstressfull, same with autopilot in heavy stop and go traffic. In these situations, the feature is reliable enough to take stress off of driving while remaining safe; making the feature useful.

But FSD looks incredibly stressful... far, far, more stressful than just driving. Yeah you can "use it responsibly", but that requires too much vigilance right now. There is no good reason why Tesla would release this as a feature to anyone outside of a very tightly controlled group. Even that is pretty risky, and Tesla is obviously only going that far as a desperate measure to retain what little trust people have left that FSD even exists in any capacity.

Until you can trust the system to not run into a concrete pillar or oncoming traffic if you take the eyes off the road for one second, it should not be released in any capacity to the public. Anything less than that is insanity.

I still have hope we'll have something useful to use within the next 6 to 12 months; but FSD beta 9 isn't it and should never see public release.