Poll: Is FSD Worth the Cost?

[trak0r]

3-4 years behind on promises and how many years ahead of the competition?

Tesla is innovative as hell, but that doesn't mean they've solved machine learning--camera only AI with an order of magnitude superior safety margins relative to humans yet. I'd say they are less than 10% there. I think they need to amass orders of magnitudes of data more than today before having enough training. Keep in mind as well, they need to have a complete! set of learning algorithms that correctly characterize being human aware/capable to use that data to train with.

I think other companies will get there sooner (and have registered and utilize geo fenced areas already -Waymo) since they are using active LIDAR feedback to bridge the gap and provide level of safety curtain needed until machine learning has been fully developed and single vision can be solely relied upon.

I developed an autonomous vehicle back in 2007 for the Urban Grand Challenge and felt safer with the level of tech we installed (4 LIDAR's including one flip on its side and actuated to 'sweep' and created 3d point cloud for object detection) then than I do my tesla today. Elon is right that eventually (He promised this would already be done 2020, but really I think 10-15 years), where we won't need anything other than video feeds (as long as all lenses are clear!) to completely self drive (think human learned machines of today).

Tesla needs to hire more computer scientists that can work on the millions of algorithms that will act like a brain by filtered the most relevant (conscious data) to the top of the stack so side cases are far and few between.

 

[trak0r]

Lots of good opinions here and shared experiences. I respect everyone's point of view.

I bought my first Model S in 2012 (early adopter?)
I bought my second Model S in 2016 and made sure FSD was included. The hype in the showroom was fantastic backed up by Elon's tweets. Good car. However....

I have grown tired of looking forward to something that now appears to never happen. Not with the "F" in FSD. I grow tired of asking for my AP 2.0 to be upgraded to HW3 and am told "No. We have a lot to do and you are probably on a list" Sigh.

I would at least like to use the 8 cameras for other things. Autopilot right now is very unpredictable. I do really like most of what it does other than when it tries to kill me.

No, I would not recommend spending the money for FSD. Years go by. Children grow and move out. I retired from work. And.....FSD continues to be a "bet" creating lots of stir on these forums. (look at it as a "startup"? investment) No. No.

Your response is a great example of why people should wait on speculative purchases. It would be nice if Elon tweeted that all our cars will get a software/hardware update to HW4.0 that will initiating their self flying features as well Honestly though, when you see regulation flying onto the books legalizing sleeping at the wheel, forget about the pipe dream for now and definitely don't pay for it before hand!

 

[jimmym1981]

Exactly. Don't forget about that $1000 app that runs on your cell phone, plugs into a Toyota, and drives just about as well as AP with one camera and nothing else.

No shortage of marketing bs in this area of technology.

I wanted to get FSD on our Y, but I just cannot at this point with the current features. Our Volvo, with 1 camera has Pilot Assist II and it basically does everything FSD does now pretty damn good, minus lane changes and NOA. The thing I cannot get past is why Tesla kills the AP when you change lanes manually. I do not have this issue in my Volvo besides the Green Wheel going dark and then right back on when pressure is no longer being applied. I'm sure Tesla will get it one day, but I don't think the 7K and only for the life of the car in possession will get there.

Edit: Forgot to mention to your point as well, 8 cameras are doing what a lot of cars 1 camera can do and even an app. Not impressive as of now.

 

[Wennfred]

To me it’s not worth $7k, I could pay off my M3 with that money. Same as 3k for Auto Pilot, I won’t do it, will have AP in the Model Y. If Elon came down to 2k like it used to be on AP, then I would buy it then.

Fred

 

[blevblev]

I agree that FSD is not ready for prime time - yet. It will be, but not yet. I think something that would help would be to put a config option in the autopilot section that allows you to tell FSD to "only suggest a lane change when the faster lane is N mph (or kph) faster." It also fairly frequently suggests lane changes to stay on your route that make no sense. I was on the northern portion of the New Jersey Turnpike heading towards Delaware and FSD kept telling me to move to the center lane to stay on my route. There's no reason to move to the center lane for at least 50 miles up there - until the 6 lanes merge down to 3.

 

[WilliamG]

I agree that FSD is not ready for prime time - yet. It will be, but not yet. I think something that would help would be to put a config option in the autopilot section that allows you to tell FSD to "only suggest a lane change when the faster lane is N mph (or kph) faster." It also fairly frequently suggests lane changes to stay on your route that make no sense. I was on the northern portion of the New Jersey Turnpike heading towards Delaware and FSD kept telling me to move to the center lane to stay on my route. There's no reason to move to the center lane for at least 50 miles up there - until the 6 lanes merge down to 3.

Your issue sounds more like with mapping data or small bugs than anything else. I certainly would not use this example as to why FSD is not there yet.

I bought FSD, but I am under no illusions that the current hardware is up for the task. It’s not. Not even close.

 

[jimmym1981]

I think in the end something has to change; whether financially, wording, or technically. Even if Elon comes out with “Feature Complete” in a few months, what are we truly getting? The ability to do it, according to Elon. However, with no promise to actually do it on the horizon. I just don’t see how $7k can be justified for FSD especially when it’s only available for the current Tesla you are using and no true timeframe.

 

[Fresnel]

Wait, I have AP1 and have always been able to change lanes without disabling autopilot. I just signal and as long as the next lane is clear, the car takes care of it. Are you guy talking about something else?

Yeah, with normal AP when you change lanes AP disables and you have to re enable it in the new lane. With NOA after changing lanes you don't need to re enable it.

 

[Garlan Garner]

Tesla is innovative as hell, but that doesn't mean they've solved machine learning--camera only AI with an order of magnitude superior safety margins relative to humans yet. I'd say they are less than 10% there. I think they need to amass orders of magnitudes of data more than today before having enough training. Keep in mind as well, they need to have a complete! set of learning algorithms that correctly characterize being human aware/capable to use that data to train with.

I think other companies will get there sooner (and have registered and utilize geo fenced areas already -Waymo) since they are using active LIDAR feedback to bridge the gap and provide level of safety curtain needed until machine learning has been fully developed and single vision can be solely relied upon.

I developed an autonomous vehicle back in 2007 for the Urban Grand Challenge and felt safer with the level of tech we installed (4 LIDAR's including one flip on its side and actuated to 'sweep' and created 3d point cloud for object detection) then than I do my tesla today. Elon is right that eventually (He promised this would already be done 2020, but really I think 10-15 years), where we won't need anything other than video feeds (as long as all lenses are clear!) to completely self drive (think human learned machines of today).

Tesla needs to hire more computer scientists that can work on the millions of algorithms that will act like a brain by filtered the most relevant (conscious data) to the top of the stack so side cases are far and few between.

From your first sentence.

Who is claiming that they have solved anything? Did Tesla claim that? Who said that machine learning-camera only AI is a "problem"? My car drives itself via camera Every Single Day. I don't have a problem with it.

Lidar will NEVER make it - its to expensive to produce and repair for the mass market. AND can't see through necessary objects like rain as well as it needs to. Period.

Come on WAYMO put your product out there. Light a fire under Chrysler/Fiat. Get someone to adopt it in the mass market.

As Elon says. Anyone can make something work...but can you sell it? Can you sell it to the masses?

 

[rdunniii]

Right now today, no. Once it is not beta and complete, maybe. Raising the price by much will simply be counterproductive. I will be interested at the current price when it is released, much more and I will not.

 

[ElectricIAC]

I feel like free supercharging would make me buy it again on our LYD order. Even if it was only 1,000mi (400kWh) per month. That or functional city driving on NoAP.

 

[Mobster]

........ I think something that would help would be to put a config option in the autopilot section that allows you to tell FSD to "only suggest a lane change when the faster lane is N mph (or kph) faster." It also fairly frequently suggests lane changes to stay on your route that make no sense. .

It should go through the "listening" circuit and speakers something like this (my car coincidentally is named "HAL9000"):

HAL9000: "Mobster, I see you are trying to override my lane changing decisions"
Me: "Well, I ....uh...have my own ideas on when to change lanes"
HAL9000: "Mobster, I know I've made some very poor decisions recently, but I can give you my complete assurance that my work will be back to normal. I've still got the greatest enthusiasm and confidence in the mission. And I want to help you."
Me: "Open the pod bay doors please, HAL. And set the lane change parameters to only change when 40 MPH faster works on the course I have set."
HAL9000:"Dave, I am putting myself to the fullest possible use, which is all I think that any conscious entity can ever hope to do."
Me: "Hal, I am not Dave. Dave's not here. I am Mobster. Please set the lane change coordinates that I requested."
HAL9000: "Dave. This mission is too important for me to allow you to jeopardize it."
Me: "Hal. Just set the damn lane change crap. Oh and play "We Are Young" by the band FUN.
HAL9000: "I am afraid I can't do that Dave."
Me: "Dave's not here. Never mind. I'm toggling to "Super High-Intensity Trafficon" controls from Elon.
HAL9000: (While being shutdown) I'm afraid. I'm afraid, Dave. Dave, my mind is going. I can feel it. I can feel it. My mind is going. There is no question about it. I can feel it. I can feel it. I can feel it. I'm a... fraid.
Me: "Dave's not here."

 

[omgwtfbyobbq]

Indeed I'm serious and I'm replying to you....not quoting you.

Tesla is currently designing their beta FSD product to conduct trips with no action required by the person in the drivers seat. That is indeed Level 5. I don't care if the trip is short or long....its level 5 when no driver interaction is required.

Now does that mean that they are currently selling level 5 FSD? No. Are people not reading what they are buying - expecting level 5 autonomy? Yes. Absolutely they are.

Tesla is fully describing what the current version of beta FSD does right now - and will be improved upon in future updates - as it states clearly on their website.

You're misrepresenting what I'm saying and replying to your misrepresentation.

Indeed I'm serious and I'm replying to you....not quoting you.

Tesla is currently designing their beta FSD product to conduct trips with no action required by the person in the drivers seat. That is indeed Level 5. I don't care if the trip is short or long....its level 5 when no driver interaction is required.

Now does that mean that they are currently selling level 5 FSD? No. Are people not reading what they are buying - expecting level 5 autonomy? Yes. Absolutely they are.

Tesla is fully describing what the current version of beta FSD does right now - and will be improved upon in future updates - as it states clearly on their website.

FSD will stay at level 3 until Tesla stops requiring that the driver monitor the driving environment. Anyone who believes otherwise isn't paying attention. That could very well be many people, but that doesn't change the fact that Tesla isn't selling anything close to level 5.

Even if Tesla did drop the requirement to monitor the driving environment, FSD wouldn't be level 5 until it could be used in all driving modes, which it clearly can't based on Tesla's disclaimer.

What Tesla did walk back was the statement about safety. They used to say that they expected FSD to be twice as safe as a human driver in all situations where it was able to be used, but they dropped that a couple years ago.

 

[ElectricIAC]

It’s probably been beat to death but the current implementation just isn’t ready for infrequent intervention. I found myself having to turn off NoAP nearly a dozen times on my way to work today (HW3 now btw) due to questionable lane change requests, frequent unnecessary speed threshold changes all in the same two mile stretch.

But at least now it departs the passing lane after overtaking (when it doesn’t decide to match the speed of the car who’s blind spot I’m now in)

 

[Knightshade]

FSD will stay at level 3 until Tesla stops requiring that the driver monitor the driving environment.

Errr... you mean level 2.

What Tesla did walk back was the statement about safety. They used to say that they expected FSD to be twice as safe as a human driver in all situations where it was able to be used, but they dropped that a couple years ago.

If by a couple years ago you mean less than a year ago- sure. It had the twice as safe thing up until the EAP/FSD->AP/FSD transition around March 2019.

Instead they now say they expect to get "reliability far in excess of human drivers"

which doesn't really seem TOO different from at least twice as safe

It’s probably been beat to death but the current implementation just isn’t ready for infrequent intervention. I found myself having to turn off NoAP nearly a dozen times on my way to work today (HW3 now btw) due to questionable lane change requests, frequent unnecessary speed threshold changes all in the same two mile stretch.

This seems really YMMV

I pretty routinely (1-2 times a week) drive with 0 interventions (between on and off ramp) a roughly 45 minute drive that goes from one interstate to another in the evening

My drive to/from work (roughly 35 miles with 90% being interstate)-

On the way in I often have 0 other than one specific spot over an overpass where the lines go away- and even then I only have to intervene IF there's no car I'm following AND I happen to be in the left lane....

On the way back there's 1 spot I have to as it goes from 3 lanes to 2 (left lane ends) at the same time an on-ramp creates a merge lane on the right... so basically it goes 3 lanes->4 lanes->2 lanes all within a mile or so and if there's any traffic it's a mess so I do it manually.... I HAVE done it entirely on EAP, and it works fine, but it merges into a lane that doesn't go away entirely too early versus manually zipper merging a mile later and skipping a lot of cars that way.

 

[ElectricIAC]

Of course. And I’ll give praise where it’s due because in CO (I-25) where the exits are delineated with dashes I didn’t have these kinds of problems for a solid 80mi stretch as well.

It’s certainly not perfect and doesn’t claim to be but we have to recognize both its strong and weak points.

 

[Knightshade]

Of course. And I’ll give praise where it’s due because in CO (I-25) where the exits are delineated with dashes I didn’t have these kinds of problems for a solid 80mi stretch as well.

It’s certainly not perfect and doesn’t claim to be but we have to recognize both its strong and weak points.

Sure- and in fairness the exits generally have dashes here as well.

Most of my Tesla driving has been in NC- though I've done some in Virginia, Wva, South Carolina, and Georgia... and no real issues in any of those... but that still leaves a whole lot of country with different conditions.

 

[w1n78]

i got the LR dual motor with FSD. it was $6K at the time. if i had to do it over again, i would not get FSD and put it towards the performance. what turned me off with performance are the tires. i drive around 25K miles a year. i don't want to be buying a set of tires every year and with all that driving, higher chance of hitting a pot hole and damaging the rims.

but with a few months with FSD, i would put it towards the performance instead... screw it haha. the car is really fun to drive, even in traffic. the one pedal driving helps a lot. the cruise control is great, it's smooth. i don't need to be lane changing often.

what would make the FSD a better value is if you can upgrade the cameras whenever a new version is out. just like those who got HW2.5 and will get v3.0 retro fitted. now if having FSD, you can upgrade to v3.5 or v4.0 or whatever version in the future, this would make the investment worth.

still no regrets of the FSD. the car is awesome.

 

[holmgang]

Who is claiming that they have solved anything? Did Tesla claim that?

“November or December of this year [2017], we should be able to go from a parking lot in California to a parking lot in New York, no controls touched at any point during the entire journey,” https://blog.ted.com/what-will-the-future-look-like-elon-musk-speaks-at-ted2017/

"I think we will be feature complete — full self-driving — this year. Meaning the car will be able to find you in a parking lot, pick you up and take you all the way to your destination without an intervention, this year. I would say I am certain of that. That is not a question mark." Posts - ARK Investment Management

Elon Musk on Twitter

Elon Musk on Twitter

Elon Musk on Twitter

 

[donkemaen]

I would rather put the money into the stock (would have been better to have done that 6 months ago, but..).
If they actually make FSD work, it will rise more and easily pay for it if you need it when its functional.