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

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

This site may earn commission on affiliate links.
It is amazing to read many of these responses because they show, what I believe, is an unrealistic understanding of what a full self driving system will be capable of doing.
In terms of aircraft, small to medium size planes avoid flying in extreme weather with or without autopilot systems. Drivers should do the same. It is ridiculous to believe that full driving is going to allow you to drive thru a blizzard or hurricane. Sensible drivers know to stay off the roads in such conditions. But what about 'normal' driving conditions. Humans have trouble navigating day to day driving conditions as it is. An affordable FSD system cannot be expected to do that much better than a human. It is a fantasy to believe otherwise. Under present road conditions and distracted drivers, such capability would require onboard super computers, a suite of sensors that are unaffordable and unattractive and even then would result in jerky, slow driving responses to a busy commuter route.
Under near perfect conditions, with well marked roadways, intersections and other obvious obstacles, full self driving using Tesla's vision system could successfully navigate without driver input. But such a world does not, nor will it ever exist in a free society.
To answer the original question posed in this post, if we are realistic about our expectations for the definition of full self driving than I believe in ideal driving conditions it is already possible. The question becomes when will the obstacles to FSD be removed to allow it to work reliably.
Think about what Tesla has already publicly disclosed and accomplished at a beta level, sign and signal recognition, vehicle and pedestrian identification, obstacle i.d., navigation to a destinations, autopark, traffic awareness, self steering and braking and growing AI data based to side in split second decision making. Did I miss anything? IMHO

I think it is realistic to expect for a car with FSD to make turns on intersections, avoid potholes on the road, and do this in a city. This is definitely not happening right now.
 
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.
Actually I this is my first new car. I have bought used cars from dealerships and do not recall being deceived.
 
So...robo-taxi I is just shill talk?

Aspirational forward-looking statement would be nearer the legal term :)

And unlike just delivering an L2 version of FSD, actual robotaxis WILL require regulatory approval in a lot of places where it doesn't currently exist so even if they're ever able to do it technically (which they HOPE to, but haven't as of today), that'll also delay that bit.
 
  • Funny
Reactions: jsmay311
FSD will never work with the existing sensor suite, at least not under all circumstances. During good driving conditions, sure. Night time? Sure. But the cameras simply can't look into the sun and still see the road. There are times when a human who is very familiar with the road has to slow to a crawl to navigate into direct sunshine, through snow, fog or rain, through flooded intersections, down muddy roads and driveways, squeeze between concrete construction barriers while straddling a lane marker. There's way more going on at real time speeds in a human brain than can be coded into silicon. For starters, the car would have to be able to remember situations from when it had a clear view of obstacles. Just last weekend I hit the brakes hard to avoid an impending disaster based upon only HEARING a tire explode. The car doesn't hear at all and didn't even flinch. It would have run directly into a whole pile of debris that visuals alone didn't provide enough time to compensate for.

All of this is not to say FSD features aren't worth pursuing or that they won't be useful for 97.38%+ of driving cases, just that the human will always be a necessity for fringe cases that are not all that uncommon.

Anyway, that's my $.02 worth on the subject.
 
FSD will never work with the existing sensor suite, at least not under all circumstances. During good driving conditions, sure. Night time? Sure. But the cameras simply can't look into the sun and still see the road. There are times when a human who is very familiar with the road has to slow to a crawl to navigate into direct sunshine, through snow, fog or rain, through flooded intersections, down muddy roads and driveways, squeeze between concrete construction barriers while straddling a lane marker. There's way more going on at real time speeds in a human brain than can be coded into silicon. For starters, the car would have to be able to remember situations from when it had a clear view of obstacles. Just last weekend I hit the brakes hard to avoid an impending disaster based upon only HEARING a tire explode. The car doesn't hear at all and didn't even flinch. It would have run directly into a whole pile of debris that visuals alone didn't provide enough time to compensate for.

All of this is not to say FSD features aren't worth pursuing or that they won't be useful for 97.38%+ of driving cases, just that the human will always be a necessity for fringe cases that are not all that uncommon.

Anyway, that's my $.02 worth on the subject.

But on the other hand it sees much better in foggy conditions due to the radar.
 
  • Informative
Reactions: pilotSteve
But on the other hand it sees much better in foggy conditions due to the radar.

"Sees" other objects, maybe. But not lane lines, street signs, signals, nor anything else necessary to actually, you know, drive.

I live in an area prone to dense fog and autopilot has become VERY conservative with recent releases, not available at all in even moderately foggy conditions.

TACC does still work fine, hence the radar advantage, but that's not anywhere near enough for "FSD".
 
As I read through the comments I'm happy to see that most agree with what some of us have been saying for years now. We have our doubts as to if the sensors installed and cameras will be able to achieve FSD in the city. Because Tesla has yet to demonstrate this, even with a prototype vehicle. I sure hope I am wrong and just waiting for Tesla to prove me wrong. But as the competition releases more and more videos it's hard to believe our camera's are capable of achieving the same on city streets.
 
I don't want FSD anyway as I love driving in many circumstances. I just want hands free freeway and highway driving, but with a driver still being ultimately responsible for the vehicle. In my opinion, AP is already good enough to remove the nag sequence. Just let me relax while driving. My hand will still be on the wheel, but don't force me to keep torque on the wheel.

There's nothing much as fun as carving corners on a winding highway. Keep your FSD tentacles off. :D
 
  • Funny
Reactions: SergeyUser
It is amazing to read many of these responses because they show, what I believe, is an unrealistic understanding of what a full self driving system will be capable of doing.
In terms of aircraft, small to medium size planes avoid flying in extreme weather with or without autopilot systems. Drivers should do the same. It is ridiculous to believe that full driving is going to allow you to drive thru a blizzard or hurricane. Sensible drivers know to stay off the roads in such conditions. But what about 'normal' driving conditions. Humans have trouble navigating day to day driving conditions as it is. An affordable FSD system cannot be expected to do that much better than a human. It is a fantasy to believe otherwise. Under present road conditions and distracted drivers, such capability would require onboard super computers, a suite of sensors that are unaffordable and unattractive and even then would result in jerky, slow driving responses to a busy commuter route.
Under near perfect conditions, with well marked roadways, intersections and other obvious obstacles, full self driving using Tesla's vision system could successfully navigate without driver input. But such a world does not, nor will it ever exist in a free society.
To answer the original question posed in this post, if we are realistic about our expectations for the definition of full self driving than I believe in ideal driving conditions it is already possible. The question becomes when will the obstacles to FSD be removed to allow it to work reliably.
Think about what Tesla has already publicly disclosed and accomplished at a beta level, sign and signal recognition, vehicle and pedestrian identification, obstacle i.d., navigation to a destinations, autopark, traffic awareness, self steering and braking and growing AI data based to side in split second decision making. Did I miss anything? IMHO

A little bit dramatic there do you not think? A plane can simply just fly over or around a weather system in a lot of cases. Weather is valid because one moment you can be out and its fine the other moment you are in the back seat playing cards and the car is screaming at you to take over. The thing is this is about what was said it could do when you buy and if it will ever do what was said it could do within a reasonable time frame. Like within the decade.
 
  • Funny
Reactions: Mobster
Explain how, as a human, you see a stop sign that is not visible due to being overgrown?

You don't of course.

So either you "know" it's there because you remember it from before it was overgrown, or you blow through the sign because you have no idea it's there.

The car has maps as a secondary assist to vision to avoid that second, far worse with a human, situation.

I see even this thread where the map is secondary backup to seeing or seeing is a backup to the map. Which is trusted? I assume maybe the most cautious piece of data is used, but that won't always work well. If map data says 45 because of map error or location error, but sign is seen as 55, which? If no stop sign is seen, but map says there is one, I assume you go with the map. Data fusion and deconfliction can get tricky real quick for all these cases.
 
Nope.

Tesla had a discrete cost for FSD and plenty of people bought cars without FSD, so their liability is the FSD license,. Because they had a price for EAP, which was delivered, the value of FSD is no greater than US$4K.

Resetting expectations wouldn't necessarily hurt their P&L as they haven't recognized much of that revenue yet. However, if they have to do a reset on expectations, they would have to give some back.

We were happy to get FSD based on the capabilities which I think Tesla can deliver this year.

But if Tesla eventually decide they overpromised on FSD's capabilities, the major hit would be reputational not immediately financial.

As long as they keep trying then they won't have to refund. They could layoff the whole team except one developer and claim to be working on it. There was never a promised date of delivery so they can never fail as long as they are working on it.
 
Last edited:
As long as they keep trying then they won't have to refund. They could layoff the whole team except one developer and claim to be working on it. There was never a promised date of delivery so they can never fail as long as they are working on it.

I think once more and more owners will stop owning their Teslas and practically getting very little as far as FSD is concerned, especially for cars which are 8 years or older, someone will eventually figure out how to make Tesla liable, regardless of the fact that there were no timelines indicated. I think they can fail in a legal sense earlier than having just one engineer working on it.
 
There was never a promised date of delivery so they can never fail as long as they are working on it.

To quote a famous cigar enthusiast: "It depends on what the definition of 'is', is."

I am not sure anything was "promised" Let me check on that.

I look at my contract/ bill of sale that says I get a car (A promise of when? Yes.) That I get FSD (no promise of when or what that means but the Telsa Sales "Consultant" says "by the end of the year - just checkout what Elon is saying"), Premium Upgrades Package (is that a promise or a slogan, sounded good particularly when the same guy that told me FSD was only a few months away described what I get in this package...ok I'm in.), Ludicrous Mode (A promise? I didn't care. I was told (again verbally) that it was really fast. Cool.) I'm not sure if these conversations and acronyms on the invoice constitute any "promise". Let's see....hmmm...there is other stuff like "Enhanced Autopilot". Wow. (I understand they stopped using this along with FSD) But I was assured I needed that to get the FSD later this year. Ok. Smart Air Suspension (how smart? Promise that it is smart?) Sub Zero (uh...not sure if there is any promise there but what the hell) All this in writing. Other stuff that could be seen (like red calipers if you bend over just right) All promises of what where and when? I dunno.

So what is the definition of "is"? I look up a definition of "promise". It's confusing but under the first definition it say "declare that something will happen". Wow. Did Elon make promises? Now that seems easy.....except.....did we get lots of promises of when something would happen or no promises of anything. My wife says that a promise is "something that assuredly will be broken." Oh. Ok. I'm still stuck with HW 2.0 and lots of fancy footwork by Tesla folks. More FSD litigation? Promise? Will that help? Priomises that the FSD price is going up? (that one will meet its due date.)
 
Last edited:
I think once more and more owners will stop owning their Teslas and practically getting very little as far as FSD is concerned, especially for cars which are 8 years or older, someone will eventually figure out how to make Tesla liable, regardless of the fact that there were no timelines indicated. I think they can fail in a legal sense earlier than having just one engineer working on it.
If Twitter & public presentations are considered to be inducing action of the consumer in the eyes of a judge, then there were timelines
Also ‘the reasonable person test’ & hence what is ‘reasonable time’ may come into play
 
  • Like
Reactions: SergeyUser
To quote a famous cigar enthusiast: "It depends on what the definition of 'is', is."

I am not sure anything was "promised" Let me check on that.

I look at my contract/ bill of sale that says I get a car (A promise of when? Yes.) That I get FSD (no promise of when or what that means but the Telsa Sales "Consultant" says "by the end of the year - just checkout what Elon is saying"), Premium Upgrades Package (is that a promise or a slogan, sounded good particularly when the same guy that told me FSD was only a few months away described what I get in this package...ok I'm in.), Ludicrous Mode (A promise? I didn't care. I was told (again verbally) that it was really fast. Cool.) I'm not sure if these conversations and acronyms on the invoice constitute any "promise". Let's see....hmmm...there is other stuff like "Enhanced Autopilot". Wow. (I understand they stopped using this along with FSD) But I was assured I needed that to get the FSD later this year. Ok. Smart Air Suspension (how smart? Promise that it is smart?) Sub Zero (uh...not sure if there is any promise there but what the hell) All this in writing. Other stuff that could be seen (like red calipers if you bend over just right) All promises of what where and when? I dunno.

So what is the definition of "is"? I look up a definition of "promise". It's confusing but under the first definition it say "declare that something will happen". Wow. Did Elon make promises? Now that seems easy.....except.....did we get lots of promises of when something would happen or no promises of anything. My wife says that a promise is "something that assuredly will be broken." Oh. Ok. I'm still stuck with HW 2.0 and lots of fancy footwork by Tesla folks. More FSD litigation? Promise? Will that help? Priomises that the FSD price is going up? (that one will meet its due date.)

At time of purchase it was clearly written that the functionality was not currently available but would be in the future and "it was not possible to know exactly when it would be available" as it was contingent on SW validation and it being approved by "regulators". So legally no delivery date was given other than infinite timeframe. This was the old language prior to I believe Feb 2019. So I still contend they left the door open that they haven't broken any purchase contract as long as they are still working on development and waiting for approval. This is different than any other options you listed which didn't say clearly it wasn't available at time of purchase.
 
At time of purchase it was clearly written that the functionality was not currently available but would be in the future and "it was not possible to know exactly when it would be available" as it was contingent on SW validation and it being approved by "regulators". So legally no delivery date was given other than infinite timeframe. This was the old language prior to I believe Feb 2019. So I still contend they left the door open that they haven't broken any purchase contract as long as they are still working on development and waiting for approval. This is different than any other options you listed which didn't say clearly it wasn't available at time of purchase.

Unless it can be proven via discovery that they intentionally made promises they can not keep.