I used it quite a bit during my free month, and I had far too many instances that convinced me to never pay for the feature, at least where it's at right now. There was one instance when the car slammed on the brakes to come to a complete stop at a yellow light. The speed limit was 45mph, and it scared me half to death because had someone been behind me they would have rammed right into me. There were other times when it would try to take an off-ramp that was also an on-ramp for other cars and it would get confused and disengage FSD. There were times when it would mistakenly get into exit lanes it wasn't supposed to be in, and it would constantly get into lanes that didn't need to be in. When it worked it worked well, but there's more than enough there to convince me to never pay $8,000 or $99 a month.
This is accurate. You can tell by watching the speed. It goes from "AUTO: MAX" to the normal speed controls we used to have on v11. The transition is super obvious, weird he missed it.
I feel like fsd is just missing out in that transition, it just feels a bit too obvious. I’d like them to make the transition better and also when arriving at a destination dropping you off a little better than it does already. It’s only going to get better from here and to be honest, it’s not that bad and is a huge improvement over last year
I used it on rural roads mainly and think a lot of my miles where on the v11 stack but it could never tell. Might try it again when the entire stack is v12. Creeping while traffic was coming was not good and def. confused other drivers. Turning off a gravel road to a highway was hit or miss, it would turn into the shoulder, the left lane, different every time.
I think it would be worth it if you could summon your car from anywhere. like if you were stranded somewhere in another car then you could summon your car and your car would come to your location without anyone in it.
This is exactly the feature that would make me buy a Tesla with FSD. First this needs to work in parking lots and garages; the end game for this feature is being able to summon it from anywhere to drive to me anywhere (as long as I'm by a street or road, of course; I have no felt need to have it go offroading to fetch me!)
@@Wordsmiths pretty good chance that will happen eventually. How much is FSD going to cost then? I think if you can afford it, it’s going to turn out to be a great bargain to lock the price in at the current level.
As a non-Tesla owner who's been on the fence of buying a Tesla, FSD has been enticing. After having used OpenPilot for the past month and seeing your review of FSD, I'm less inclined to buy a Tesla now. With OpenPilot I'm essentially a passenger in the driver seat for 99.99% of interstate and highway driving. I don't have to touch the wheel ever, I can still accelerate under any condition and it will keep OpenPilot engaged, for lane changes I simply hit my turn signal and nudge the steering wheel. With other third-party forks of OpenPilot I can take it a step or two farther and remove the steering wheel nudge requirement for lane changes and I can customize certain options as well. There are parts it lags behind FSD but from the sounds of it, interstate/highway driving is far more relaxing.
Tesla Fsd has those features (albeit a lot more expensive), highway driving features have been apart of the package for years now, city driving is relatively new, but you could just make fsd only work on highways or you can even just ouch the turn signal for the car to change lanes, just fsd is more advanced and less need for the driver to confirm stuff. Openpilot looks really good and the price is a lot better than what Tesla offers.
I’ve had a good experience. A few mistakes including repeatedly blowing on-ramp red lights. But overall it was relaxing and I felt like a passenger with less impatience, frustration or distraction. I could see paying for it if I had my old commute but I’m working from home so it’s hard to justify. Like you said, we bought a Tesla to save on gas so piling up new expenses is questionable.
If I have to supervise, then I might as well drive. When Tesla accepts liability and doesn’t make me pay attention and I can trust it, then I’ll pay and use it
_If I have to supervise, then I might as well drive._ Certainly your choice; but doing so deprives you and others around you of the astonishing safety record of FSD. _When Tesla accepts liability_ Tesla will accept liability right now; all you have to do is purchase your collision & comprehensive insurance from Tesla (in the states that offer it), rather than State Farm, GEICO, AAA, USAA, or whomever you're purchasing it from now. And this will often save you money; my quote for C&C insurance on my 2024 Model Y is about $40/month *less* than my current insurer, USAA. I can easily envision a scenario in which FSD receives approval for SAE Level 3 or Level 4 ADS in one or more states; but Tesla will only activate it if you purchase your C&C insurance from them. In fact, if I were Tesla, that's exactly what I'd do.
@@billspencer5589 What do you base that assertion on? You're already covering that liability in your C&C insurance. The only thing Tesla accepting liability, if ever, is shift some of the coverage from your insurance company to Tesla, which will also shift some of the cost.
@@BigBen621 I was trying to say that the FSD will be very expensive once it’s fully autonomous so I think it’s wise to consider that spending $8000 for it now given that that will be active for the life of the car could be a very good bargain
@@bspencersf Sorry, I misunderstood. Whether to purchase outright or subscribe is a real quandary. I've leased Teslas twice, and both times I included FSD in the lease. This added about $200/month to the lease. Of course now I'm wishing I hadn't included it in the lease, because now I could be subscribing for $99/month, saving me $100/month! Anyway, whatever decision you make, I hope it turns out well for you!
Just got free fsd a week ago. A major flaw on the freeway is a like to keep it at 70mph for rage efficiency and stay in the middle lane. I don’t occupy the right lane because semi trucks and I left lane for passing cars. I like that fsd will pass cars going slow in the middle lane but I don’t like that it doesn’t put me back in the middle lane. 70-75 is still considered slow in the passing lane so I’d rather it put me back. Takes off way to quickly at intersections and thus will wear out the tires a lot sooner. Stop way too late at a red light and I assume it’s using the friction breaks which is a no no for me. I would rather use regen. Doesn’t take turns wide enough and I always feel like it’s going to curb rash my wheels
You can use the blinker to tell it to change lanes back to the middle. I do wish however you could make slight route adjustments without kicking it out of FSD
@@bspencersf yah I knew that which is still convenient but I wish it could just put me back in the middle lane automatically without intervention. Knowing this now I would never consider buying advanced autopilot even though it’s not a thing anymore. Unless they include that in the future
@@mikeli1126 I haven’t tried that yet so yah I’m pretty sure that will fix the issue I just know I don’t like using chill mode on the freeway because it does everything slowly and keeps way too big of a gap. At this point it’s probably just a me problem but still a problem in my opinion
One thing to keep in mind for first time FSD users is the rate of improvement over the years. its not perfect, but if you compare it to even 1 year ago, you start to see a path to complete autonomy. I also think it takes more than a month to get comfortable with its limitations. You know when you’ll need to intervene. Right now its at a point where it takes the stress out of driving in stop and go traffic. Once nag is gone, it’ll be a huge step forward. I dont know when Robotaxis will be widespread, but I see a path to FSD being very useful very soon
I think you’re saying it takes awhile to learn to use FSD.. Which to me means it’s not nearly actually FULLY SELF DRIVING, which by definition doesn’t need me to learn anything but how to turn it on.
Agree with most of your impressions. Finished my trial and signed up for 1 month because of a long road trip.....i can see it used for long trips to justify the cost but not for daily use as of yet....it's better than a party trick to be fair but not worth the cost for day to day drive... I also agree the B pillar cam is a miatake, cam should have been on the A-PILLAR and would resolve most of the stop sign and left turn creep issues.... When all is said done this might become a FREE feature for all TESLAs.....and yes when you can sleep or lay back, the value will significantly increase
Bear in mind that when it really is autonomous, it will likely be very expensive. Consider buying now @ $8k for THE LIFE OF THE CAR if you want it for personal use
Just received the trial last week and so far, it’s not ready for prime time yet. FSD, for me, can’t make a proper left turn when there’s two left turn lanes. I’m on the outside lane and it always makes a left in the inside lane. Luckily, there are no cars to my left. It has also made some scary lane changes on the freeway when there’s a car coming fast in my blind spot. I turned it off 😂. I do like enhanced autopilot though.
We have been letting FSD do 95% of our miles for years. Yes we need to occasionally intervene, but it is rare for 12.3.6. Driving manually is way more stressful than supervising FSD.
Seems like the Y has at least some benefit at stop signs at least where there are relatively low barriers because it’s a bit taller. Not sure how much a difference it would make though.
The rural 2-lane road unnecessary swerve into a turn-or-die lane issue I experienced about 20 times extremely reliably and predictably over the course of 500 miles driving across Colorado on v11 back in January (I even included the running count on my voice feedback to Tesla). I think it has something to do with programming to stay out of passing lanes (I.e. the rightmost lane) or to avoid the rightmost lane in other cases. Spot-on analysis that matches my experience exactly. I have MY RWD with hardware 4. That and the stop sign and aggressiveness issues you identified (along with the infuriating inability to leave “minimize lane changes” on and the tendency for it to not change lanes early enough for exits even with that setting off) are really the challenges for them. There needs to be a mode chiller than chill (maybe “mom’s in the car mode”) - I can’t use any of the driver assistance features with passengers in the car due to exactly what you described.
I don't have the most recent update yet. Mine rides to closer to the lane marker. On curves or one lane roads it's very uncomfortable, I find my self tensing up.
I’m curious, how. Much would you personally pay on a monthly basis for FSD once it is say truly a level 5 autonomy, and you can do whatever you want with your time in the car
I’ve got a route where v11 did the exact same thing you described with the turn lanes. I think mine was even more egregious because the next turn on route was a left but it veered into a right turn lane at 60mph. Same hard 90 degree turn and I’ll say this, brakes on teslas are extremely good.
My trial period ended today and I had about the same issues as you described. I also had a chance to test it extensively as I was in the middle of a long road trip when the trial period began and did around 3000 miles with FSD on. As you mentioned, FSD takes way too long to go at stop signs. Other than that, my main issue is with consistency. FSD would do certain maneuvers well the vast majority of the time but would screw up once in a while (especially on left turns), which means that you always need to be very alert, which partly defeats the purpose. If I would keep it (I won't for now because of the price), it would be for highway driving and in large cities that I'm not familiar with, not so much in rural areas or familiar surroundings. Having said that, as it improves over time, I can see myself subscribing in the future for long road trips.
The problem with your approach is that the better it gets, the more expensive it’s going to get. It’s a gamble to count on it getting better but I think it’s a pretty good chance it will and if you pay the full price of $8000 now You think it’s quite a bargain when it actually becomes fully autonomous
@@bspencersf I see your point but my intent, even if FSD greatly improves, would be to subscribe during months where I drive a lot rather than purchase it. Strangely, in Canada, monthly subscription is the same as in the US at $99 but purchasing FSD here is $11,000 rather than $8,000. Living in a climate where the system is pretty useless during winter, subscribing when needed would be the better approach for me, even if the monthly subscription cost increases over time.
Many of the issues i had in v11 have been resolved. But i echo your findings with latest v12. Still a dramatic rate of improved. Maybe 12.5 will be the tipping point with the full stack implementation.
Drew you made a real interesting point about why does Tesla expect customers to pay $8k or $100 per month subscription which negates any fuel savings you would achieve. Just doesn't make sense with the current state of FSD, I had some of the same issues you had while testing. Personally if Tesla wants more uptake of current FSD versions they need to drop the subscription down to $10-$20, maybe $30 max, the lower the better. It's better to make revenue by volume versus barely any revenue from limited uptake, just my opinion.
Do you think it can be lvl 3 now. Meaning you have to be sitting there, it’s driving but you can take over if needed but no nags or keep your eyes on the road requirement. Like the Mercedes system.
Level 2 requires the fallback-ready user to take over immediately when required, while Level 3 allows a delay before the fallback-ready user has to take over, typically 10 seconds. This is trivially easy when all vehicles are moving in the same direction at
@@michaelsmithers4900 _Not sure why MB’s computers would be faster than Teslas_ It's not an issue of MB's computers being faster. Both Level 2 and Level 3 require a human driver to take over when there's a system failure or the system encounters a condition it can't handle; but Level 2 requires the driver to take over immediately, while Level 3 allows a delay before the driver takes over, typically 10 seconds, during which the system must continue driving. It's just that it's easy to delay the human taking over for ten seconds when all cars are driving in the same direction, on a limited-access divided highway, at 40 MPH, in daytime, on dry pavement; which are the conditions MB Drive Pilot is restricted to. But it's hugely more difficult to assure safety for ten seconds when cars may be driving in the opposite direction a few feet away on a two-lane road, at 85 MPH, at night, in the rain, which FSD allows; in fact, I'm not sure this is possible at all.
The shit with Tesla this year is worrisome. Love my model 3 but the massive layoffs (esp. the supercharging team), selling/dumping of Tesla stocks by top leadership within tesla, extreme over-hype of FSD and what feels like a sudden shift of focus to AI and full self driving everything has me holding my breath as far as what direction the company will go in. I don't have Tesla stock so my concern stems purely from a consumer/owner standpoint. My experience has been great in the 2yrs I have owned the car. Hope the company can really back up all the hype behind the semi, robo taxi, FSD, etc cuz right now shit is lookin SHHHAAKKYYY. My wife had expressed interest in buying a Tesla in the future for herself but now I'm like hmmmmmm let's see what direction shit goes in lol.
On rural two lane highways it’s always a little too close to the center line for my comfort. Most people stay pretty close to the right lane marker on those highways. Also yes it starts from a red light way too fast, like a street racer. I would check the Energy App and it’d tell me to accelerate slower to improve efficiency, which I thought was hilarious.
To be fair for the highway drives, it still uses the old V11 stack lol, thought you’d know this man come on. V12.5 should make it all single stack again. 12.4 and 12.5 should be huge and hopefully by August 8th we get close to Robotaxi level. 😄
Why would a company that expects everyone to let their car drive them to work, also make “sports” versions of their cars and planning dedicated sports cars? Why would people shell out extra money for a sport version just to let a computer drive them to work? Some people love to drive
Are you referring to the Performance versions of the Model 3 and Y, and Plaid versions of the Model S and X? If so, just because you don't care for driving in traffic jams and would rather have FSD do that for you, doesn't mean you can't still have fun carving canyons in the Sierras on weekends. Probably more than any other car in the world, you get the best of both worlds with a Tesla. The Tesla Model S Plaid has performance comparable to supercars costs several times as much, while still being a 5-passenger family sedan when that's needed. What other car does that?
What’s so interesting to me about this is how different peoples experiences are. The car doesn’t drive exactly like I would. There are many times when it’s more cautious than I am at intersection. It’s not my job to worry about people behind me-if they are impatient. I let the car do what it’s going to do and don’t feel guilty about making someone take an extra 30 seconds for FSD to be sure it’s safe. I hope and assume that when the situations happened to you you are sending the message back to Tesla so they can look at it and improve it. I also think you are making an assumption about the cameras that may not be true. The front windshield cameras have a very wide angle of view, and are used more than the side cameras in the situations you’re talking about. I too have seen it strangely get in the left or right lanes when inappropriate but I don’t panic, even at highway speeds I’m confident the car is not gonna run off the road…. Also what very few people seem to think about when considering the price is that the price for the subscription or outright purchase is going to go way up when it does actually achieve full autonomy. You likely won’t be able to afford it unless you’re using the car to generate income. I guess it’s gambling but I think there’s a pretty good chance that if you’re going to keep your car at least five years, paying $8000 for FSD at this point and having that stay with the car for the life of the car the value of the car will increase way more than the $8000 you pay for it. I paid 12,000 for it summer of 2022 and still feel like it’s worth it. 😮To Each His Own!
It's assuming that people keep saying when FSD becomes fully autonomous it will be more expensive despite the fact it's more competent now at $8000 than it was at $15k awhile back. Plus isn't Tesla planning to license out FSD to other automakers? What if that company charges a cheaper FSD purchase price or subscription? Come on people get real with this it will get more expensive so buy now statement.
@@BrianNedry yeah… Maybe… If Tesla does license FSD to another manufacturer, what vehicle would they be putting it on? Are they going to glue a bunch of cameras to the fenders? By the time somebody else gets a vehicle working with the FSD software Tesla will have had a monopoly for several years I think. I said it was a gamble… I’m willing to take it but certainly understand how many are not willing to. I have already gotten way more than $12,000 benefit out of it in the two years I’ve owned my model Y. Again everyone’s different and if everybody is thinking about this then great that was all I was trying to point out
@@BrianNedry yeah… Maybe… If Tesla does license FSD to another manufacturer, what vehicle would they be putting it on? Are they going to glue a bunch of cameras to the fenders? By the time somebody else gets a vehicle working with the FSD software Tesla will have had a monopoly for several years I think. I said it was a gamble… I’m willing to take it but certainly understand how many are not willing to. I have already gotten way more than $12,000 benefit out of it in the two years I’ve owned my model Y. Again everyone’s different and if everybody is thinking about this then great that was all I was trying to point out
Honestly, I'm happy with the advanced autopilot. FSD will mature when it matures, I don't care how long it takes. I'm not holding my breath. I don't miss FSD, I'm not longing for it or upset that it isn't great yet, that's for sure. I wouldn't use it much anyway!
My take on vision only approach is, if you are going vision only you need every inch of that car covered in cameras... at least as you say, better positioned. There is another problem with the FSD, if you use it ALOT.. it's going to wear out your brakes, because it doesn't use rolling stop, it actually hits the physical brakes. Over time that's going to cause the need for brake pads quicker than someone using rolling stops.
@@billspencer5589 I'm so underappreciated in my time. It feels good to see that someone appreciates my high intellect and new approaches to solving problems.
Robotaxi? Monkeyballs....would you trust this system to drive your loved ones without them being able to drive if necessary? This thing is still YEARS out from being close to safe and elon seems to be putting all the chips in on it
I think this is what most enthusiastic fans miss out on. They see something with lots of tech that is impressive (though far from finished) and they scratch their heads that the stock is not skyrocketing... That's the divide between dreams and reality, and now that we are seeing how the EV adoption rate is struggling more than many of those fans anticipated, it should be clear that the technology challenge to make FSD a reality is far from being the only issue, not even the most difficult to crack. It will take years before this truly explodes and becomes real. Even then, I have serious doubts about its monetisation potential.
Doubts about monetization? I really don't get this take. Transport costs today = cost of the vehicle + cost of labor. Transportation as a human driven/run service is *already* profitable today. Remove the human part from each individual vehicle (even if you need say 1 person per 5 vehicles for logistics/operations) that's still a 5x reduction in costs. How can it not be profitable?
@@filipehenrique3158 really? Please do the same calculation about owning an EV over a similarly priced ICE or hybrid car. Outside the US it is extremely simple to show how an EV is simply cheaper and more convenient, right after purchase and even more so over years of ownership... Are people buying EVs like crazy? No. Why? Because it's anything but a rational discussion. FUD, orchestrated attack on EVs from legacy auto, ideology, envy and resentment, resistance to change, ignorance... There are tons of reasons why people are not transitioning and buying ICE cars instead, when the rational and responsible thing to do would be to buy an EV. We should learn from this, especially because this next step involves ending lots of jobs, and I presume there will be a lot more resistance. It will not be simple nor quick, therefore, not easily profitable.
I've been using FSD for 2,500 miles, and it isn't what Musk was hawking in 2020. There's no way that the current software and hardware can be used for actual "autonomous" self-driving. I'd use it for free or as a data collector, but I'll never pay for it. Tesla Vision will be replaced by Lidar eventually. FSD on 12.3.6 has cut off too many people, and I've made it as docile as I possibly can. I never know when FSD will decide to change lanes, and I'm convinced that at least part of the reason for that is the translation of the data from the side cameras which are wide-angle, All the usual Muskateers are trying to convince everyone it's "almost there", but it isn't almost what was promised 4 years ago. No one has mentioned liability and insurance, but I guess no one should be concerned about the minutia. I agree 100% with your evaluation!
Lidar is still light and it has twice as far to travel. I would rather see cameras that can see in IR. And maybe changing the mounting position or additional cameras.
@@danharold3087the speed of light will never be a hindrance. The speed of sound is not even a fraction of 1 percent the speed of light and we use sound based mapping and sensors all the time and feel vertically instant. So even if it had to loop around 100 times it would still be just as fast as a camera because of the processing speed itself. 1solid state lidar sensor would be the equivalent of an ultra slow mode camera as long as long as the processor could handle the data. I strongly believe that if they added them today within a year they could reach level 4. The main issues with fsd right now is that the cameras can get confused by perspective and cause phantom breaking this would not be a problem if it had lidar and radar. What Tesla needs to do is add cameras on the front bumper in the center and at both of the front pillars in addition to 2 lidar sensors on the front corners and one in the back.
@@Anderson-hi9qw Sorry I should have been more explicit. Distance traveled makes a difference when it is hard to see due to snow, rain, and blowing sand.
I don’t understand why people boy subscribe to FSD on roadtrips. Basic AP is good enough on highways. FSD is used on streets during daily commutes. Makes no sense to me
Auto lane change (meaning I put the blinker and it goes - not the aggro weaving) is actually really handy in heavier traffic in the sense that I can only look in one direction at a time but the cameras can see more cars at once. There’s a section of road in Seattle where a major highway merges from the left and you need to get across 5 lanes in a mile while at the same time an onramp from the right has people trying to get across 3 lanes in the opposite direction to avoid three upcoming exit only lanes that can be chaotic when there aren’t a lot of “slots” for cars and you have a constant risk of cars coming from behind AND from two lanes over trying to take the spots from the opposite side. I wish that would be brought down to free autopilot since it’s the only part of it that I’m willing to pay for. I actually find FSD’s weaving and inability to permanent-set “minimize lane changes” really annoying and disconcerting.
Once we get to version 32 sometime in 2028 _THEN_ it will be fairly decent. Little improvements each time, but still very far away from actual full self driving.
Авто/Мото
I used it quite a bit during my free month, and I had far too many instances that convinced me to never pay for the feature, at least where it's at right now.
There was one instance when the car slammed on the brakes to come to a complete stop at a yellow light. The speed limit was 45mph, and it scared me half to death because had someone been behind me they would have rammed right into me.
There were other times when it would try to take an off-ramp that was also an on-ramp for other cars and it would get confused and disengage FSD.
There were times when it would mistakenly get into exit lanes it wasn't supposed to be in, and it would constantly get into lanes that didn't need to be in.
When it worked it worked well, but there's more than enough there to convince me to never pay $8,000 or $99 a month.
About the most fair and balanced review I have heard. Having had fsd 12.x for 2 months and using it everyday my experience is almost identical
I believe on highways and freeways it still uses old v11 stack so maybe that's why you're seeing those issues. Streets is all v12
This is accurate. You can tell by watching the speed. It goes from "AUTO: MAX" to the normal speed controls we used to have on v11. The transition is super obvious, weird he missed it.
I feel like fsd is just missing out in that transition, it just feels a bit too obvious. I’d like them to make the transition better and also when arriving at a destination dropping you off a little better than it does already. It’s only going to get better from here and to be honest, it’s not that bad and is a huge improvement over last year
Party Trick is a great way to describe this. I will remember that.
I send Tesla a response every time it asks me why did I turn off FSD.
I used it on rural roads mainly and think a lot of my miles where on the v11 stack but it could never tell. Might try it again when the entire stack is v12.
Creeping while traffic was coming was not good and def. confused other drivers.
Turning off a gravel road to a highway was hit or miss, it would turn into the shoulder, the left lane, different every time.
I think it would be worth it if you could summon your car from anywhere. like if you were stranded somewhere in another car then you could summon your car and your car would come to your location without anyone in it.
This is exactly the feature that would make me buy a Tesla with FSD. First this needs to work in parking lots and garages; the end game for this feature is being able to summon it from anywhere to drive to me anywhere (as long as I'm by a street or road, of course; I have no felt need to have it go offroading to fetch me!)
@@Wordsmiths pretty good chance that will happen eventually. How much is FSD going to cost then? I think if you can afford it, it’s going to turn out to be a great bargain to lock the price in at the current level.
I don't think AI is at the point where it can be trusted to drive outside of specific parameters.
I cannot understand why they do not add a front bumper camera
As a non-Tesla owner who's been on the fence of buying a Tesla, FSD has been enticing. After having used OpenPilot for the past month and seeing your review of FSD, I'm less inclined to buy a Tesla now. With OpenPilot I'm essentially a passenger in the driver seat for 99.99% of interstate and highway driving. I don't have to touch the wheel ever, I can still accelerate under any condition and it will keep OpenPilot engaged, for lane changes I simply hit my turn signal and nudge the steering wheel. With other third-party forks of OpenPilot I can take it a step or two farther and remove the steering wheel nudge requirement for lane changes and I can customize certain options as well. There are parts it lags behind FSD but from the sounds of it, interstate/highway driving is far more relaxing.
Really looking forward to trying OpenPilot on my Aptera!
Tesla Fsd has those features (albeit a lot more expensive), highway driving features have been apart of the package for years now, city driving is relatively new, but you could just make fsd only work on highways or you can even just ouch the turn signal for the car to change lanes, just fsd is more advanced and less need for the driver to confirm stuff. Openpilot looks really good and the price is a lot better than what Tesla offers.
I’ve had a good experience. A few mistakes including repeatedly blowing on-ramp red lights. But overall it was relaxing and I felt like a passenger with less impatience, frustration or distraction. I could see paying for it if I had my old commute but I’m working from home so it’s hard to justify. Like you said, we bought a Tesla to save on gas so piling up new expenses is questionable.
Nice to see a channel talking honestly and practically about Tesla.
If I have to supervise, then I might as well drive. When Tesla accepts liability and doesn’t make me pay attention and I can trust it, then I’ll pay and use it
_If I have to supervise, then I might as well drive._
Certainly your choice; but doing so deprives you and others around you of the astonishing safety record of FSD.
_When Tesla accepts liability_
Tesla will accept liability right now; all you have to do is purchase your collision & comprehensive insurance from Tesla (in the states that offer it), rather than State Farm, GEICO, AAA, USAA, or whomever you're purchasing it from now. And this will often save you money; my quote for C&C insurance on my 2024 Model Y is about $40/month *less* than my current insurer, USAA.
I can easily envision a scenario in which FSD receives approval for SAE Level 3 or Level 4 ADS in one or more states; but Tesla will only activate it if you purchase your C&C insurance from them. In fact, if I were Tesla, that's exactly what I'd do.
Probably won't be able to afford it then. It will likely be VEY expensive.
@@billspencer5589 What do you base that assertion on? You're already covering that liability in your C&C insurance. The only thing Tesla accepting liability, if ever, is shift some of the coverage from your insurance company to Tesla, which will also shift some of the cost.
@@BigBen621 I was trying to say that the FSD will be very expensive once it’s fully autonomous so I think it’s wise to consider that spending $8000 for it now given that that will be active for the life of the car could be a very good bargain
@@bspencersf Sorry, I misunderstood. Whether to purchase outright or subscribe is a real quandary. I've leased Teslas twice, and both times I included FSD in the lease. This added about $200/month to the lease. Of course now I'm wishing I hadn't included it in the lease, because now I could be subscribing for $99/month, saving me $100/month! Anyway, whatever decision you make, I hope it turns out well for you!
Just got free fsd a week ago. A major flaw on the freeway is a like to keep it at 70mph for rage efficiency and stay in the middle lane. I don’t occupy the right lane because semi trucks and I left lane for passing cars. I like that fsd will pass cars going slow in the middle lane but I don’t like that it doesn’t put me back in the middle lane. 70-75 is still considered slow in the passing lane so I’d rather it put me back.
Takes off way to quickly at intersections and thus will wear out the tires a lot sooner. Stop way too late at a red light and I assume it’s using the friction breaks which is a no no for me. I would rather use regen. Doesn’t take turns wide enough and I always feel like it’s going to curb rash my wheels
You can use the blinker to tell it to change lanes back to the middle. I do wish however you could make slight route adjustments without kicking it out of FSD
@@bspencersf yah I knew that which is still convenient but I wish it could just put me back in the middle lane automatically without intervention. Knowing this now I would never consider buying advanced autopilot even though it’s not a thing anymore. Unless they include that in the future
putting the car in chill mode while using FSD will tell the car to accelerate less fast from a stop
@@mikeli1126 I haven’t tried that yet so yah I’m pretty sure that will fix the issue I just know I don’t like using chill mode on the freeway because it does everything slowly and keeps way too big of a gap. At this point it’s probably just a me problem but still a problem in my opinion
One thing to keep in mind for first time FSD users is the rate of improvement over the years. its not perfect, but if you compare it to even 1 year ago, you start to see a path to complete autonomy. I also think it takes more than a month to get comfortable with its limitations. You know when you’ll need to intervene. Right now its at a point where it takes the stress out of driving in stop and go traffic. Once nag is gone, it’ll be a huge step forward. I dont know when Robotaxis will be widespread, but I see a path to FSD being very useful very soon
I think you’re saying it takes awhile to learn to use FSD.. Which to me means it’s not nearly actually FULLY SELF DRIVING, which by definition doesn’t need me to learn anything but how to turn it on.
MSD mostly self driving
Agree with most of your impressions. Finished my trial and signed up for 1 month because of a long road trip.....i can see it used for long trips to justify the cost but not for daily use as of yet....it's better than a party trick to be fair but not worth the cost for day to day drive...
I also agree the B pillar cam is a miatake, cam should have been on the A-PILLAR and would resolve most of the stop sign and left turn creep issues....
When all is said done this might become a FREE feature for all TESLAs.....and yes when you can sleep or lay back, the value will significantly increase
Bear in mind that when it really is autonomous, it will likely be very expensive. Consider buying now @ $8k for THE LIFE OF THE CAR if you want it for personal use
@@billspencer5589'"If" it becomes autonomous.
Great video I haven't seen anybody cover truly what it is about like you have good job 👍
Just received the trial last week and so far, it’s not ready for prime time yet. FSD, for me, can’t make a proper left turn when there’s two left turn lanes. I’m on the outside lane and it always makes a left in the inside lane. Luckily, there are no cars to my left. It has also made some scary lane changes on the freeway when there’s a car coming fast in my blind spot. I turned it off 😂. I do like enhanced autopilot though.
We have been letting FSD do 95% of our miles for years. Yes we need to occasionally intervene, but it is rare for 12.3.6. Driving manually is way more stressful than supervising FSD.
Seems like the Y has at least some benefit at stop signs at least where there are relatively low barriers because it’s a bit taller. Not sure how much a difference it would make though.
The rural 2-lane road unnecessary swerve into a turn-or-die lane issue I experienced about 20 times extremely reliably and predictably over the course of 500 miles driving across Colorado on v11 back in January (I even included the running count on my voice feedback to Tesla). I think it has something to do with programming to stay out of passing lanes (I.e. the rightmost lane) or to avoid the rightmost lane in other cases. Spot-on analysis that matches my experience exactly. I have MY RWD with hardware 4. That and the stop sign and aggressiveness issues you identified (along with the infuriating inability to leave “minimize lane changes” on and the tendency for it to not change lanes early enough for exits even with that setting off) are really the challenges for them. There needs to be a mode chiller than chill (maybe “mom’s in the car mode”) - I can’t use any of the driver assistance features with passengers in the car due to exactly what you described.
I don't have the most recent update yet. Mine rides to closer to the lane marker. On curves or one lane roads it's very uncomfortable, I find my self tensing up.
I’m curious, how. Much would you personally pay on a monthly basis for FSD once it is say truly a level 5 autonomy, and you can do whatever you want with your time in the car
Good analysis, I live in a rural area also and agree it does better with stop lights than stop signs.
I’ve got a route where v11 did the exact same thing you described with the turn lanes. I think mine was even more egregious because the next turn on route was a left but it veered into a right turn lane at 60mph. Same hard 90 degree turn and I’ll say this, brakes on teslas are extremely good.
My trial period ended today and I had about the same issues as you described. I also had a chance to test it extensively as I was in the middle of a long road trip when the trial period began and did around 3000 miles with FSD on. As you mentioned, FSD takes way too long to go at stop signs. Other than that, my main issue is with consistency. FSD would do certain maneuvers well the vast majority of the time but would screw up once in a while (especially on left turns), which means that you always need to be very alert, which partly defeats the purpose. If I would keep it (I won't for now because of the price), it would be for highway driving and in large cities that I'm not familiar with, not so much in rural areas or familiar surroundings. Having said that, as it improves over time, I can see myself subscribing in the future for long road trips.
The problem with your approach is that the better it gets, the more expensive it’s going to get. It’s a gamble to count on it getting better but I think it’s a pretty good chance it will and if you pay the full price of $8000 now You think it’s quite a bargain when it actually becomes fully autonomous
@@bspencersf I see your point but my intent, even if FSD greatly improves, would be to subscribe during months where I drive a lot rather than purchase it. Strangely, in Canada, monthly subscription is the same as in the US at $99 but purchasing FSD here is $11,000 rather than $8,000. Living in a climate where the system is pretty useless during winter, subscribing when needed would be the better approach for me, even if the monthly subscription cost increases over time.
@@yzd263 yeah, I really think it would be great if you could pay by mile...
Many of the issues i had in v11 have been resolved. But i echo your findings with latest v12. Still a dramatic rate of improved. Maybe 12.5 will be the tipping point with the full stack implementation.
Drew you made a real interesting point about why does Tesla expect customers to pay $8k or $100 per month subscription which negates any fuel savings you would achieve. Just doesn't make sense with the current state of FSD, I had some of the same issues you had while testing. Personally if Tesla wants more uptake of current FSD versions they need to drop the subscription down to $10-$20, maybe $30 max, the lower the better. It's better to make revenue by volume versus barely any revenue from limited uptake, just my opinion.
6:00 IMO there should be side facing camera's on the front fenders for this very reason.
Do you think it can be lvl 3 now. Meaning you have to be sitting there, it’s driving but you can take over if needed but no nags or keep your eyes on the road requirement. Like the Mercedes system.
Level 2 requires the fallback-ready user to take over immediately when required, while Level 3 allows a delay before the fallback-ready user has to take over, typically 10 seconds. This is trivially easy when all vehicles are moving in the same direction at
Hardware wise, yes I think it could do level 3 on the freeway. Legally I’m unsure, NHTSA is already biased against Tesla. Don’t see how that’ll change
For me it’s a safety issue. Every time I drive I see drivers weaving in and out of lanes distracted. FSD is just safer IMHO
@@BigBen621Not sure why MB’s computers would be faster than Tesla’s. Tesla is generally considered to have high end software and hardware.
@@michaelsmithers4900 _Not sure why MB’s computers would be faster than Teslas_
It's not an issue of MB's computers being faster. Both Level 2 and Level 3 require a human driver to take over when there's a system failure or the system encounters a condition it can't handle; but Level 2 requires the driver to take over immediately, while Level 3 allows a delay before the driver takes over, typically 10 seconds, during which the system must continue driving.
It's just that it's easy to delay the human taking over for ten seconds when all cars are driving in the same direction, on a limited-access divided highway, at 40 MPH, in daytime, on dry pavement; which are the conditions MB Drive Pilot is restricted to. But it's hugely more difficult to assure safety for ten seconds when cars may be driving in the opposite direction a few feet away on a two-lane road, at 85 MPH, at night, in the rain, which FSD allows; in fact, I'm not sure this is possible at all.
The 90 degree turns especially on the highways is an FSD/Enhanced auto pilot weakness I've noticed.
The shit with Tesla this year is worrisome. Love my model 3 but the massive layoffs (esp. the supercharging team), selling/dumping of Tesla stocks by top leadership within tesla, extreme over-hype of FSD and what feels like a sudden shift of focus to AI and full self driving everything has me holding my breath as far as what direction the company will go in.
I don't have Tesla stock so my concern stems purely from a consumer/owner standpoint. My experience has been great in the 2yrs I have owned the car. Hope the company can really back up all the hype behind the semi, robo taxi, FSD, etc cuz right now shit is lookin SHHHAAKKYYY.
My wife had expressed interest in buying a Tesla in the future for herself but now I'm like hmmmmmm let's see what direction shit goes in lol.
What data is that 2% based on?
On rural two lane highways it’s always a little too close to the center line for my comfort. Most people stay pretty close to the right lane marker on those highways.
Also yes it starts from a red light way too fast, like a street racer. I would check the Energy App and it’d tell me to accelerate slower to improve efficiency, which I thought was hilarious.
_it starts from a red light way too fast_
Musk has stated explicitly that this will be correct in V12.4, coming out any day now.
Should be renamed to "FSD urban & highway only" (not rural), or "FSD Unrural" lol
To be fair for the highway drives, it still uses the old V11 stack lol, thought you’d know this man come on. V12.5 should make it all single stack again. 12.4 and 12.5 should be huge and hopefully by August 8th we get close to Robotaxi level. 😄
I’m aware, just reviewing it as it is today, and many of my issues were not on the highway
Why would a company that expects everyone to let their car drive them to work, also make “sports” versions of their cars and planning dedicated sports cars?
Why would people shell out extra money for a sport version just to let a computer drive them to work?
Some people love to drive
Are you referring to the Performance versions of the Model 3 and Y, and Plaid versions of the Model S and X? If so, just because you don't care for driving in traffic jams and would rather have FSD do that for you, doesn't mean you can't still have fun carving canyons in the Sierras on weekends. Probably more than any other car in the world, you get the best of both worlds with a Tesla. The Tesla Model S Plaid has performance comparable to supercars costs several times as much, while still being a 5-passenger family sedan when that's needed. What other car does that?
Yeah it stops too long at stop signs
Been using FSD 12 and by no means is this software perfect but you can see where this is going.
The worst part about some RUclips channels is we can do our own research 😂😂😂
What’s so interesting to me about this is how different peoples experiences are. The car doesn’t drive exactly like I would. There are many times when it’s more cautious than I am at intersection. It’s not my job to worry about people behind me-if they are impatient. I let the car do what it’s going to do and don’t feel guilty about making someone take an extra 30 seconds for FSD to be sure it’s safe.
I hope and assume that when the situations happened to you you are sending the message back to Tesla so they can look at it and improve it. I also think you are making an assumption about the cameras that may not be true. The front windshield cameras have a very wide angle of view, and are used more than the side cameras in the situations you’re talking about. I too have seen it strangely get in the left or right lanes when inappropriate but I don’t panic, even at highway speeds I’m confident the car is not gonna run off the road….
Also what very few people seem to think about when considering the price is that the price for the subscription or outright purchase is going to go way up when it does actually achieve full autonomy. You likely won’t be able to afford it unless you’re using the car to generate income. I guess it’s gambling but I think there’s a pretty good chance that if you’re going to keep your car at least five years, paying $8000 for FSD at this point and having that stay with the car for the life of the car the value of the car will increase way more than the $8000 you pay for it. I paid 12,000 for it summer of 2022 and still feel like it’s worth it. 😮To Each His Own!
Yeah Elon said the price would only go up now it’s been lowered twice by substantial margins. Who’s to say it won’t go down again?
It's assuming that people keep saying when FSD becomes fully autonomous it will be more expensive despite the fact it's more competent now at $8000 than it was at $15k awhile back. Plus isn't Tesla planning to license out FSD to other automakers? What if that company charges a cheaper FSD purchase price or subscription? Come on people get real with this it will get more expensive so buy now statement.
@@BrianNedry yeah… Maybe… If Tesla does license FSD to another manufacturer, what vehicle would they be putting it on? Are they going to glue a bunch of cameras to the fenders? By the time somebody else gets a vehicle working with the FSD software Tesla will have had a monopoly for several years I think. I said it was a gamble… I’m willing to take it but certainly understand how many are not willing to. I have already gotten way more than $12,000 benefit out of it in the two years I’ve owned my model Y. Again everyone’s different and if everybody is thinking about this then great that was all I was trying to point out
@@BrianNedry yeah… Maybe… If Tesla does license FSD to another manufacturer, what vehicle would they be putting it on? Are they going to glue a bunch of cameras to the fenders? By the time somebody else gets a vehicle working with the FSD software Tesla will have had a monopoly for several years I think. I said it was a gamble… I’m willing to take it but certainly understand how many are not willing to. I have already gotten way more than $12,000 benefit out of it in the two years I’ve owned my model Y. Again everyone’s different and if everybody is thinking about this then great that was all I was trying to point out
Honestly, I'm happy with the advanced autopilot. FSD will mature when it matures, I don't care how long it takes. I'm not holding my breath. I don't miss FSD, I'm not longing for it or upset that it isn't great yet, that's for sure. I wouldn't use it much anyway!
Mine expired today lol
I used the trial last month. It wasn't that great. I'll just use it for long roadtrips.
I prefer to be alone when I use FSD
My take on vision only approach is, if you are going vision only you need every inch of that car covered in cameras... at least as you say, better positioned.
There is another problem with the FSD, if you use it ALOT.. it's going to wear out your brakes, because it doesn't use rolling stop, it actually hits the physical brakes.
Over time that's going to cause the need for brake pads quicker than someone using rolling stops.
Yes! You should definitely apply for a job with Tesla. Don't know why they haven't thought of this approach!
@@billspencer5589 I'm so underappreciated in my time. It feels good to see that someone appreciates my high intellect and new approaches to solving problems.
Robotaxi? Monkeyballs....would you trust this system to drive your loved ones without them being able to drive if necessary? This thing is still YEARS out from being close to safe and elon seems to be putting all the chips in on it
I agree. Hope I'm wrong, but so far I've been right
I think this is what most enthusiastic fans miss out on. They see something with lots of tech that is impressive (though far from finished) and they scratch their heads that the stock is not skyrocketing... That's the divide between dreams and reality, and now that we are seeing how the EV adoption rate is struggling more than many of those fans anticipated, it should be clear that the technology challenge to make FSD a reality is far from being the only issue, not even the most difficult to crack.
It will take years before this truly explodes and becomes real. Even then, I have serious doubts about its monetisation potential.
Doubts about monetization? I really don't get this take.
Transport costs today = cost of the vehicle + cost of labor.
Transportation as a human driven/run service is *already* profitable today.
Remove the human part from each individual vehicle (even if you need say 1 person per 5 vehicles for logistics/operations) that's still a 5x reduction in costs. How can it not be profitable?
@@filipehenrique3158 really?
Please do the same calculation about owning an EV over a similarly priced ICE or hybrid car. Outside the US it is extremely simple to show how an EV is simply cheaper and more convenient, right after purchase and even more so over years of ownership...
Are people buying EVs like crazy? No. Why? Because it's anything but a rational discussion.
FUD, orchestrated attack on EVs from legacy auto, ideology, envy and resentment, resistance to change, ignorance... There are tons of reasons why people are not transitioning and buying ICE cars instead, when the rational and responsible thing to do would be to buy an EV.
We should learn from this, especially because this next step involves ending lots of jobs, and I presume there will be a lot more resistance. It will not be simple nor quick, therefore, not easily profitable.
Good analysis ...and yes the Tesla cars need a couple of cameras in front of the car to be better in 360 vision then a human as you said ...correct .
Great review, thanks! HEY! Remember when the Model 3 was going to be the robotaxi?
I've been using FSD for 2,500 miles, and it isn't what Musk was hawking in 2020. There's no way that the current software and hardware can be used for actual "autonomous" self-driving. I'd use it for free or as a data collector, but I'll never pay for it. Tesla Vision will be replaced by Lidar eventually. FSD on 12.3.6 has cut off too many people, and I've made it as docile as I possibly can. I never know when FSD will decide to change lanes, and I'm convinced that at least part of the reason for that is the translation of the data from the side cameras which are wide-angle, All the usual Muskateers are trying to convince everyone it's "almost there", but it isn't almost what was promised 4 years ago. No one has mentioned liability and insurance, but I guess no one should be concerned about the minutia. I agree 100% with your evaluation!
Lidar is still light and it has twice as far to travel. I would rather see cameras that can see in IR. And maybe changing the mounting position or additional cameras.
@@danharold3087 Definitely needs more cameras, mounted up near the roof at the corners of the windshield, or something like that.
Lidar doesn’t solve any of the problems with FSD
@@danharold3087the speed of light will never be a hindrance. The speed of sound is not even a fraction of 1 percent the speed of light and we use sound based mapping and sensors all the time and feel vertically instant. So even if it had to loop around 100 times it would still be just as fast as a camera because of the processing speed itself. 1solid state lidar sensor would be the equivalent of an ultra slow mode camera as long as long as the processor could handle the data. I strongly believe that if they added them today within a year they could reach level 4. The main issues with fsd right now is that the cameras can get confused by perspective and cause phantom breaking this would not be a problem if it had lidar and radar. What Tesla needs to do is add cameras on the front bumper in the center and at both of the front pillars in addition to 2 lidar sensors on the front corners and one in the back.
@@Anderson-hi9qw Sorry I should have been more explicit. Distance traveled makes a difference when it is hard to see due to snow, rain, and blowing sand.
I don’t understand why people boy subscribe to FSD on roadtrips. Basic AP is good enough on highways. FSD is used on streets during daily commutes. Makes no sense to me
If I was going to visit LA I would get it in a New York minute.
Auto lane change (meaning I put the blinker and it goes - not the aggro weaving) is actually really handy in heavier traffic in the sense that I can only look in one direction at a time but the cameras can see more cars at once. There’s a section of road in Seattle where a major highway merges from the left and you need to get across 5 lanes in a mile while at the same time an onramp from the right has people trying to get across 3 lanes in the opposite direction to avoid three upcoming exit only lanes that can be chaotic when there aren’t a lot of “slots” for cars and you have a constant risk of cars coming from behind AND from two lanes over trying to take the spots from the opposite side. I wish that would be brought down to free autopilot since it’s the only part of it that I’m willing to pay for. I actually find FSD’s weaving and inability to permanent-set “minimize lane changes” really annoying and disconcerting.
It is still a beta far years away from being on level four or five of autonomy
Suicide lane!!!!
Once we get to version 32 sometime in 2028 _THEN_ it will be fairly decent. Little improvements each time, but still very far away from actual full self driving.
Drama
Drama queen
not very interesting and a little whiny