Did you not see the autonomous day driving video? They are testing now in company cars. It is running in the background mode in customer cars. Some of it is undoubtedly running on background in HW 2.0-2.5 cars.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Is it worth $3k for FSD for EAP owners
- Thread starter Dariushd
- Start date
they've had a similar video with the model X for a couple years now... The promise of FSD is still some ways away and if you factor in regulatory approval in each country, I don't think we will see it soon. Also with regulatory intervention/standardization, I don't think cameras and sensors will be enough
diplomat33
Average guy who loves autonomous vehicles
Autopilot is only a driver assist system now because the software is still incomplete. Since AP cannot do all the driving, the driver must compensate, hence why it is a driver assist system. Once the AP software is complete where it can handle all the driving reliably, then by definition, it won't be a driver assist anymore, it will be FSD.
Once the AP software is complete where it can handle all the driving reliably-------4 months, 8 months, 16 months, 36 months, 48 months from now?, Once doesn't mean it will happen anytime soon. Why pay for something that does not exist now. Some of us feel this is just an illegal capital raise.
FSD functions will be available soon. It will still require an engaged driver until data is gathered. The data will be gathered fast. Actual driverless operation will depend on regulatory approval. Driving without "nags" will probably not need regulatory approval so long as there is a qualified driver in the seat. What you see now in customer cars is not the level that Tesla is at. They are much farther along.
diplomat33
Average guy who loves autonomous vehicles
That's your prerogative of course. But when I get the traffic light feature and automatic city driving, I will get to use them while you will have to watch them on youtube.
That's your prerogative of course. But when I get the traffic light feature and automatic city driving, I will get to use them while you will have to watch them on youtube.
I don’t get it. What practical benefit do you get from these features that could justify the expense if you cannot truly rely on traffic light detection? You still have to be fully alert - at least as much as if you are driving without the aid.
I would stress out that the car suddenly thinks the light is green (or the light teleported away, like the cars in the display do now) and happily starts driving into cross-traffic.
Just yesterday, after succumbing to the AP praise on this forum did I try it again and while it worked with perfect lane markings it didn’t take long until it made a dangerous decision requiring me to disengage it.
It’s a super stressful “aid”. And if others feel differently then I’m scared that they don’t truly understand how poorly it still performs.
Don't forget that those 'months' are on Elon Standard Time and the new Elon [Gregorian] calendar!
diplomat33
Average guy who loves autonomous vehicles
I always pay attention but AP has been wonderful for me lately. It is not stressful for me at all. When the car can handle traffic lights and intersections, it will make my AP experience even more automated. Not FSD yet of course, but definitely less stressful and more convenient.
diplomat33
Average guy who loves autonomous vehicles
I paid $2000 for FSD. I am happy to buy it in advance and get features via OTA update as they come out. But I don't think I will have to wait 48 months to get the features.
The traffic light feature is already in shadow mode now which means it is just a few months away from possible release. Even if Tesla misses their "coming later this year" deadline (possible), I doubt that it would be 48 months away from general release.
Automatic city driving is a different question of course. We have less info on when it will be released. But I also highly doubt that we will have to wait 48 months for its general release. If we assume that Tesla misses their "coming later this year" deadline, we should at least get a piece of "automatic city driving" next year.
I agree 100% with the above. I suspect that you expect the "when" to be "real soon now" and I've seen that with Tesla, "real soon now" can be in two years, or five years.
I find that the current level of AP/EAP is enormously useful and to me it's worth what I paid. I think that was $5K. The next step up that I would pay for is Level 3, which is still not what Elon promised my car would be capable of with no more than an OTA firmware update!
This whole thread really comes down to: Do you expect truly useful features, not already in EAP, to come "soon" or do you think they will take several years? I actually think that regulatory approval will come rapidly after the system has demonstrated that it's safer than a human driver. I think that true FSD is a decade away.
Here's what it comes down to for me: Right now my car can drive itself (with me being fully alert) 90% of the time. NoA/city with traffic light and stop sign recognition would increase that to 93% of the time, with the last 7% having to wait for true Level 5 FSD. To go from doing all the driving, to the car driving 90% of the time (under my oversight) was worth $5K. Increasing the time the car drives itself (still with my oversight) from 90% to 93% is not worth another $3K. To get more money from me, that 90% has to go from level 2 to Level 3. That is, not require my constant oversight. That's worth another $5K or even $10K.
diplomat33
Average guy who loves autonomous vehicles
I do not expect true FSD to be "real soon now". But I do expect some useful AP features "soon". For example, the traffic light feature is already in shadow mode. So we should at least get that feature in a few months. I consider that to be "soon".
I find the current EAP/NOA very useful right now as well.
I think true FSD is less than a decade away but not "soon". However, I do consider useful AP features, not already in EAP, to come "soon". As I mentioned above. the traffic light feature is in shadow mode now so it should only be a few months away. I base that on the fact that NOA went from shadow mode to NOA without confirmation in roughly 6 months according to the Autonomy Investor Day presentation. So that gives us an approximate idea of how long a feature may take to go from shadow mode to full roll out.
Last edited:
There is a misconception about terms. FSD means full self driving. That does not mean autonomous. It's a significant distinction. FSD will be here soon. Autonomous is unknown as much data has to be obtained in monitored FSD mode and regulatory approval(s) are needed. Why is this so difficult to understand? The process has been explained a lot. To become autonomous, FSD has to accumulate billions of miles under supervision. Tesla is well positioned to do this because of the hundreds of thousands of cars that will be operating FSD under supervision. Musk has said FSD will be "feature complete" this year. That does not say it will be released to the general owners this year but it probably will go out to early access and HW 3 cars. All of the features are probably being worked on by groups simultaneously; sign recognition, stop lights, parking lots, 4 way stops, 90 º turns, etc. Some features of FSD may be released to HW 2-2.5 cars if they have the power to do it. We don't know all the details but what we do know is most all of what Musk says will happen does happen. It is NOT vaporware.
boiler81
Member
Existing NOA fails in a hard rain, and TACC doesn't work when it's snowing, so here in Washington State we can't count on the existing sensor suite/Autopilot features. So...why buy FSD when it can't the existing cameras and radar don't function in adverse weather?
"Useful" is a subjective determination. You might find stop light recognition at Level 2 to be "useful." I don't consider it "useful" until it is Level 3.
Differences of preference and viewpoint.
I really hope you are right on the time line. I think it will take much longer.
"Autonomy" as understood in the context of self-driving cars comes in five recognized levels. Definitions differ slightly, but basically, Level 1 Autonomy is cruise control; level 2 autonomy is adaptive cruise control plus lane-keeping, driver is required to be alert at all times; level 3 is similar but the driver need not pay attention, the car will alert the driver with enough advance warning to take over; level 4 autonomy the driver can sleep in the back seat, if the car cannot handle the situation it will find a safe place to stop and wake the driver; and level 5 autonomy is full self-driving, where the car does not need a human to be in the car and the presence of driving controls is optional. All these are levels of autonomy. Any of them can be geofenced at the manufacturer's discretion, limiting their usefulness
As for Musk's promises: He keeps his promises, but never when he says. He promised that the 2018 Model 3 had all the hardware needed for the car to function as a robotaxi, needing only an OTA firmware upgrade when the software became available and received regulatory approval. He had to admit he was wrong and to his credit has promised a hardware upgrade to those who have paid for the feature. But then he seems to have backed off on his promise and is now promising instead a set of driver-assist features that will still require a driver who must remain alert. IOW, Level 2 autonomy. Some people feel that stoplight and stop sign recognition that works much of the time and only requires their intervention occasionally is worth $3,000. Others don't feel it is worth that.
The car is already capable of Level 2 autonomy 80% or 90% of the time. Increasing that by 3 or 4% with the addition of the new promised features alters my driving by an insignificant margin. Raising the existing AP features to Level 3 would be an enormous boon. We're talking about features that sound really slick, but cover a tiny part of a typical drive. As long as they're Level 2 they're worth very little.
The problem that Tesla is going to have is that with AP on the highway an alert driver has a lot of time to react to situations the car cannot handle. Stoplights, stop signs, and left turns across traffic give the driver far less time to react to situations the car cannot handle. I really hope I'm wrong but I don't think those features are going to progress to level 3 for five years or more. And at Level 2 they're not worth it to me. Others may feel differently.
AlanSubie4Life
Efficiency Obsessed Member
It is true that this is Tesla’s intent. But to the topic of this thread, remember that Tesla is not promising L3+ autonomy to people who purchase FSD today. It’s really important when talking to people making this decision that they understand that they are purchasing only a “guarantee” of an L2 system (city NoA, sign/light recognition/response). There is no obligation on Tesla’s part to deliver more, if it does not work out as planned. These people buying today may well be required to hold the wheel and stay engaged in the driving task for the life of their vehicle, without any failure to deliver on Tesla’s part - Tesla would be 100% delivering with a fully complete L2 system.
Again, more specifically for owners considering purchasing: it’s entirely possible that you will get L2 city NoA, while future buyers of new Tesla vehicles get something better: hardware AND sensing that can do L3+. And you are not guaranteed to get what future buyers get. Hypothetically, you could be watching YouTube videos of people in their true self driving L4/L5 Teslas in a few years, while today’s buyers of FSD are stuck at L2! This would be a success for Tesla, and everybody will have what they bought, all obligations satisfied.
I agree with most of what you say, and the above may well apply to some older FSD buyers (because they bought a different promise from what I understand). But for people considering the $3k upgrade available today, there is no reason for Tesla to offer any refund as long as they deliver city NoA (L2, driver attention required) and sign/light recognition/response in a “reasonable” timeframe.
You’re missing a couple key points here: the billions of miles of data must unquestionably demonstrate a safety level far better than human drivers. It’s not enough to just gather the data.
Second point you are missing: Tesla clearly states in the purchase page for the $3k option that they will not allow these features to be used without “human supervision” (meaning L3+) if the data does not show it is safer.
If it turns out to not be safer with current hardware, people who buy this will end up with (hopefully) L2 driver-monitored (engaged driver) city NoA. Assuming that does not get prohibited by regulators of course (separate discussion). If Tesla delivers that city NoA & traffic signal/sign recognition/response, they are done, promise is kept, even if the system cannot be demonstrated to be safer than a human driver and thus cannot proceed to a higher automation level.
Anyway, that’s what they are selling today - it’s off topic to discuss prior promises in this thread so I won’t speculate what will happen to earlier FSD buyers.
Last edited:
there is no reason for Tesla to offer any refund as long as they deliver city NoA (L2, driver attention required) and sign/light recognition/response in a “reasonable” timeframe--------this can be viewed as a Capital Raise for Tesla. Tesla is in control of when "reasonable" occurs. Those owners that paid for the FSD option, regardless of when they paid are still not receiving it.
Well, the beta period itself should give you about a decade to decide
Did you see Avengers: Endgame? That kind of looked real also. (full disclosure, I did not see the movie)
My Model S already has L2 red light recognition. Any time my wife is in the car with me, she will let me know if a light is red.
Similar threads
