Will Tesla be able to deliver FSD with HW3.0 and current Model 3 sensor suite, ever?

[Mo City]

As long as there are no glare, rain, dirt or snow camera obstructions I think they will get there. I think the only long term issue will be camera obstructions.

This is what I consider the upside of the current sensor suite and would apply only to some segments of some roads. There ain't no way L4 or L5 will happen on all roads with today's vehicles.

In the very unlikely event robotaxis happen, pickup and dropoff points will be limited. That could still be enough to send Uber down in flames and utterly destroy the taxi business as we know it.

 

[smatthew]

I agree what tesla did with changing the advertised features of fsd is shady. I didn’t scrutinize the wording of the features when I opted to order my car with fsd back in October 2019. If I had, I wouldn’t have bought it.

Now having owned the car and reading some of the posts in this thread, I don’t see how the car as currently equipped could ever safely make a right turn from a stop sign or stop light onto a main road where cars might be traveling 50 mph perpendicular to the car. There is no camera or radar looking to either side to tell the car it would be safe to turn onto the road.

What about the B pillar forward facing cameras?

 

[birdsquared]

I will be adding FSD to my Model 3 in the near future (definitely plan to buy it before the price goes up - and hopefully Tesla will advise that the price is actually going up as they did the last time it went up (before I bought my car), or is more precise about the timing than the tweet - "likely on July 1").

What convinced me that significant improvements with current hardware are possible is the video showing the 'self driving' - and more importantly, the slowed down version with annotations. Everything showed on the display seems to me to be able to be measured by the existing array of detectors, and the software component didn't strike me as implausible either. I have no illusions that this will indeed be 'robotaxi' level 'self-driving', nor is that something I desire in the short term (and indeed, the huge delay that BC had with just allowing ride sharing shows the impact of regulators, and I will be very surprised if BC even allows autonomous cars on the road before 2030, let alone allowing them to pick up people who may be unable to drive).

However, the degree to which my car can drive itself on the highway without my input right now, with Autopilot, and the additional features FSD offers right now, suggest strongly to me that it is entirely feasible that my car will be able to do "most" of the driving between locations, and if I need to do first mile (km) and last, that to me is worth the price being asked, and if in the near future it can do even more - that is (choose your food metaphor - gravy, frosting, cherry).

 

[notrhj]

Now I guess 99% of the forum would disagree with me. I’m curious - is there anyone else who agrees with my prediction?

I’m long on Tesla stock
It already paid for the car.

Happy with the progress of FSD so far
It getting better every release.

And if it requires a new model to bring it to pass I can sell this one to all the nay sayers who never wanted FSD in the first place.

 

[SergeyUser]

I will be adding FSD to my Model 3 in the near future (definitely plan to buy it before the price goes up - and hopefully Tesla will advise that the price is actually going up as they did the last time it went up (before I bought my car), or is more precise about the timing than the tweet - "likely on July 1").

What convinced me that significant improvements with current hardware are possible is the video showing the 'self driving' - and more importantly, the slowed down version with annotations. Everything showed on the display seems to me to be able to be measured by the existing array of detectors, and the software component didn't strike me as implausible either. I have no illusions that this will indeed be 'robotaxi' level 'self-driving', nor is that something I desire in the short term (and indeed, the huge delay that BC had with just allowing ride sharing shows the impact of regulators, and I will be very surprised if BC even allows autonomous cars on the road before 2030, let alone allowing them to pick up people who may be unable to drive).

However, the degree to which my car can drive itself on the highway without my input right now, with Autopilot, and the additional features FSD offers right now, suggest strongly to me that it is entirely feasible that my car will be able to do "most" of the driving between locations, and if I need to do first mile (km) and last, that to me is worth the price being asked, and if in the near future it can do even more - that is (choose your food metaphor - gravy, frosting, cherry).

Which video you are referring to? What is being demoed by Tesla last year is not going to happen anytime soon for current owners at least. Tesla is great on highways, this is true. And I mean true highways without intersections. The same goes with heavy traffic. Everything else - not so much.

 

[SergeyUser]

I’m long on Tesla stock
It already paid for the car.

Happy with the progress of FSD so far
It getting better every release.

And if it requires a new model to bring it to pass I can sell this one to all the nay sayers who never wanted FSD in the first place.

Being happy with what you have does not make my prediction incorrect. I'm also happy to some extent with what I have. I am however re-aligning my expectations in regards to what I think Tesla will or will not deliver.

 

[birdsquared]

This is the video I refer to:

 

[CertLive]

First for the people jumping on her saying this is negative, actually this is probably the post of 2020 so far from a bunch of people who obviously own the car and can start to separate fantasy from reality. Comparing it to the statements made of the capabilities before, during and after sale. The reality is the Telsa model 3 does indeed have probably the most amount of sensors covering 360 degrees than any other car on the road. Each type has an advantage and a disadvantage depending on the situation. The cars ability to become a legal autonomous car 100% has to prove it is better than a human driver in all tasks to reach the holy grail. Now the thing is computers do not necessarily need to see like us to be able to make a better choice. But it sure does have to have the collective years of driving experience humans have developed over 100's of years well accept BMW drivers they never seem to be able to contain themselves being cremated by a Tesla in all situations lol ahhhh the torque... The issues are and it is fair to ask despite having the most advanced sensor suit is it enough to replace a human?

A human does not have a 360 degree view at all times a Tesla does. But a human can adapt to dramatic changes in weather a Tesla struggles because of camera blockage and I am sure radar gets real fuzzy in snow. You see I think FSD would have to be on nice sunny / overcast days only. This is why I always wondered why the car does not have radar / camera sensors in the top 4 corners of the roof. Raised up to be able to see over. Because in reality in order for FSD to work you need to be making visual choices using a neural net. Can you replicate the human instinct inside a Tesla? Otherwise just how is it going to deal with a car that is not quite parked where it should be but doesn't fit its data profile and so the car is constantly asking for advice on city roads. The video well I have not seen any more of that lately. Does it work on busy roads?

So far predictions are off from CEO & HQ. I am hoping for the city driving on UK streets to be pulled out by the end of this year. So far I am driving a car that finally on this patch can finally hand over to AP smooth without breaking on release of the accelerator. It doesn't like and will break for oncoming large lorries on 2 lane roads, coffee was spilled the other day. There is a lot to learn.

 

[SergeyUser]

This is the video I refer to:

Yes, the recent FSD demo. You may notice the car is making turns on intersections. This is great. Current FSD is not even close to trying this on a wider population of users, so just take this into account.

 

[Knightshade]

I don’t see how the car as currently equipped could ever safely make a right turn from a stop sign or stop light onto a main road where cars might be traveling 50 mph perpendicular to the car. There is no camera or radar looking to either side to tell the car it would be safe to turn onto the road.

It absolutely has side-looking cameras- plenty of video (including in this thread) showing that it sees oncoming cars in the situation you describe (up to roughly 240 meters away, so about 3 seconds in advance at ~50 mph)

The reality is the Telsa model 3 does indeed have probably the most amount of sensors covering 360 degrees than any other car on the road

Naah- there's lots of cars with far more sensors... they're mostly on development vehicles a consumer can't buy though (Waymo and friends)

Tesla is operating on the assumption they don't NEED any of those extra sensors either placement OR type- and the jury remains out if they're correct or not.

A human does not have a 360 degree view at all times a Tesla does. But a human can adapt to dramatic changes in weather a Tesla struggles because of camera blockage and I am sure radar gets real fuzzy in snow.

See this is an interesting example.... because I've certainly seen TONS of humans continue to drive when visiblity due to weather was such they really should not have been (and indeed accident rates are reliably higher in such conditions)

Presumably an L4 car will just pull over and stop in those situations, when a human would have kept driving but probably should not have.

Is that a failure of the car system... or the human?

 

[NOLADriver]

This is called geofencing. Everyone does that except for Tesla.

But its not. That's not remotely the scenario I laid out.

Geofencing has zero to do with routes being accessible or not. Just because its within a geofenced area doesnt mean it can drive on every street. I am talking about specific routes and streets being approved as navigation friendly.

 

[SergeyUser]

But its not. That's not remotely the scenario I laid out.

Geofencing has zero to do with routes being accessible or not. Just because its within a geofenced area doesnt mean it can drive on every street. I am talking about specific routes and streets being approved as navigation friendly.

geofencing is scalable version of your idea. it means that whatever path is within a certain area makes it valid (with probably additional ways to mark something as impossible, but again part of geofencing). this is just better way to do what you are saying. Anything like you suggested (make it path-specific) will not scale at all.

 

[ColonelRHS]

I will take the range over self-driving any day. I do like my self-driving feature. But if Tesla was to offer me a car with the million-mile battery( not based on a million miles) and a range of close to 500 and above in lieu of self-driving, it would be an easy choice for me.

 

[rogernaples]

I feel Tesla cheated me. When I ordered my M3 with FSD in April 2018, the sales rep told me that FSD was ready "now" and that the introduction was awaiting the laws catching up. That was my first experience with Tesla and I didn't yet know that the company and Musk are pathological liars. Love the car, hate the company.

 

[Givn2fly]

The more I see how slow Tesla is at delivering new FSD features, the more I think it will never happen. Assuming current HW3.0 and sensor suite.

I mean - Ok, stop signs. Great. In 5 years development timeline? How long it will be before Tesla releases unprotected left turns on intersections? Another 5 years?

Bottom line - my prediction is Tesla will never be able to release FSD, even in USA, on HW3.0 and current sensor suite.

Now I guess 99% of the forum would disagree with me. I’m curious - is there anyone else who agrees with my prediction?

So far my autopilot doesn't work when it is sunny out (7 months of the year), if it is winter (5 months of the year!), if the cameras have a spot of dirt on them or if a mouse farts in the next county over! I feel some decent autonomous will happen one day but I can't ever see robotaxis

 

[Duckling]

Now I guess 99% of the forum would disagree with me. I’m curious - is there anyone else who agrees with my prediction?

Agree totally, the sensors are too erratic in sunlight & darkness, complain they are dirty when they are not. Combine that with all of the random edge cases such as complex junctions, road works etc. I just cannot see the AI being able to complete this in anything like the near future.

Tesla have suggested that the software will be feature complete at the end of last year but we now know its being rewritten and feature complete does not equate in anyway to working.

I suspect I may regret my investment in FSD on my UK model 3, I suspect Tesla will change the parameters of the term Full Self Driving to match something deliverable and we'll all feel cheated. They recent withdrew supercharging rights base don new T&Cs issued after the date which I purchased my Model 3 so they have form.

The cost of FSD might not be so bad if I could transfer it to another Tesla vehicle such as a Model Y when they are available in the UK. I would not feel so bad about the investment if we could and I'm keen to see where the journey ends.

Gary

 

[SAronian]

FYI - Since activating Start/Stop, every drive I take concludes with 6+ Gigabytes of data being uploaded upon arrival at home. Seems like this is the data aggregation we've been anticipating for the next phase of FSD.

 

[ThingForMaxJoy]

I think there will be another version of the processor available before they achieve it with HW3 so probably not but hard to say how they will handle that.

They had it right in the beginning by offering EAP and FSD separate. They should have stuck to that. I think most of the people buying FSD back then didn't expect anything anytime soon. There are useful features if you don't get FSD but aren't worth $7k. Plus they don't get to count the revenue anyway. They are still counting less than the original EAP. It should be $4k for the current features and they should stop selling FSD. Then charge for new features as they become available. I think about two people would pay for the stoplight/stopsign update right now and yet they are talking about raising the price. There are better ways to make money that promises you aren't going to keep. It's a great product. Sell what you have.

Despite the negativity, the saving grace is they have done major rewrites of the NN and we haven't seen that yet. There's hope.

 

[Knightshade]

I feel Tesla cheated me. When I ordered my M3 with FSD in April 2018, the sales rep told me that FSD was ready "now" and that the introduction was awaiting the laws catching up. That was my first experience with Tesla and I didn't yet know that the company and Musk are pathological liars. Love the car, hate the company.

Was it your first experience with any car dealer of any kind?

because I've bought a lot of cars over the years- of a lot of different brands- and never had one of them I didn't catch in multiple lies about their product.

It's the buyers responsibility to do their own research and actually know what they're buying.

 

[AdamMacDon]

There are three big hardware hurdles I see Tesla facing: the lack of side visibility at long range, the view distance of the low res cameras, and the current road infrastructure not supporting autonomous driving in any way.

The lack of side visibility really worries me, currently it seems mostly the job of the forward facing wide angle lens, perhaps supported by B pillar cameras that seem really aimed down for road line viewing.

Currently, the camera footage looks like it has the quality of a ten year old webcam. There is no way it can compete in clarity with the vision of someone with 20/20 vision. When I'm driving I'm identifying hazards such as police speed traps, debris on the road, potholes, erratic drivers, etc from a long distance, sometimes up to a kilometer or more. I believe autopilot has a stated range of about a quarter that, 250 meters. At only 120kmph (around 75mph), you cover 33 meters a second. Only about 7.5 seconds to identify and react to a problem (assuming the problem isn't coming head on at the same speed, in which case, halve that). The radar can help in a forwards facing issue, but for monitoring blind spots, it isn't so useful.

Finally, and this is a big one, our roads and driving system as a whole are around the idea of human drivers. Stoplights could have something like infared flashers to send signals to autonomous cars, but they don't and probably won't for decades. Similar tech is used in my city to allow emergency vehicles to communicate with a light ahead to make all 4 directions red and stop traffic. It's not new or revolutionary, but it is expensive to add in and considering no cars use it right now, it would be a waste. Heck, forget that fancy tech, in many parts of Canada they can't even keep lines painted most of the year because of snow.

Another thing to consider, almost everyone in society is okay with 40,000 deaths or so in America because of driving accidents. Now look what happens when autopilot has a single accident with the driver misusing it. International news. This mindset will also take decades to overcome and allow common sense legislation regarding level 5 autonomy.

To sum it up, I don't expect robotaxi's in 2020