Daniel in SD
(supervised)
I don’t think they’re under any obligation to keep future prices the same as current prices. The price on the website is the current price.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Bummer for those who paid for self drive.
- Thread starter Hitman007
- Start date
You say this as if most humans would be able to analyze the situation and make the most logical decision in the split seconds they have to react (they would not). FSD does not need to be perfect, it just needs to be better than the average human - which, unfortunately, is not very difficult to beat.
If the driver behind hits the Tesla, it's the driver behind's fault and he has only himself to blame. If you can't stop in time if the car in front of you slams on the brakes, you're following too closely.
when we bought the car, we had the option to either add FSD for $3k or in the future for $5k. that is what we agreed to. if they had said "you can buy it now for $3k or later for $15k," i (and many others) would have bought it for $3k. they have to honor the price that was in the agreement when you bought the car (which is why they were allowing people to add it for $4k even after they raised the after the fact price to $5k).
now, anyone who orders now that it's no longer an option...that's a different story entirely.
So rather than speculate it would be best for someone experienced at reading contracts to determine if Tesla is obligated to sell FSD at the “add later” published price at time of purchase?
The Model X had a add tow kit after delivery option. However the tow kit is out of stock forever so essentially it’s not offered anymore.
Also, what about inflation? $3000 in 2016 money is not the same as $3000 in 2020? Even if it doesn’t move THAT much it’s still a consideration.
Tesla might have something up their sleeves that they want to charge more than $5000 for AP3 post install.
The moves are confusing but if Tesla is acting rationally - it’s some combination of
- reduce financial liability
- reduce expenses
- make more money later
The Model X had a add tow kit after delivery option. However the tow kit is out of stock forever so essentially it’s not offered anymore.
Also, what about inflation? $3000 in 2016 money is not the same as $3000 in 2020? Even if it doesn’t move THAT much it’s still a consideration.
Tesla might have something up their sleeves that they want to charge more than $5000 for AP3 post install.
The moves are confusing but if Tesla is acting rationally - it’s some combination of
- reduce financial liability
- reduce expenses
- make more money later
All those things have a human who has ultimate control. Yes airliners virtually fly themselves but there is always a pilot to takeover when things go wrong. Should be the same with "self driving" cars.
Depends if FSD is more like the Knight Industries Two Thousand (KITT) or Knight Industries Four Thousand (The TV movie in '91). KIFT? didn't stop for animal because, with the shielding, it will sustain no damage running it over.
insaneoctane
Well-Known Member
I think anyone who was offered FSD for $3K during configuration has very good legal standing to expect to purchase it at $4K (or $5K for some) once it is actually functional. Your configurator clearly gave you 2 choices to decide between. It said $3K now or $4K later. Folks made their decision based on that information. If Tesla wanted more flexibility then they should have written the configurator differently (ie $3K now or $4K if ordered later* *Note: Post delivery prices subject to change.). I believe that any reasonable jury would agree that a reasonable person could expect to add FSD afterwards for $4K.
That's adorable. Also wrong. I don't care how high tech and cutting edge the technology might still be after a few years, it doesn't change the fact I'll want a new car. I'll be surprised if I have more than 50,000 miles on my Model 3 when I sell it.
If this is what you say why not simply raise the price of FSD... that will pay for that hardware upgrade... if it’s designed to be retrofitable to begin with..
Tesla removing that feature off the menu cannot possibly be any good news..
I could have bought Amazon at $500 but didn’t. Prices change. I think you’re out of luck.
IMO the over reliance on vision to do everything pretty much guarantees that FSD is many years away.
Look at basic things like blind spot detection. If they used a radar it would be trivial and we would see cars coming from a greater distance. I really wonder how making an automatic lane changes on a German autobahn works out in the current implementation.
A simple case in mind are the windshield wipers. Rather than using a basic $2 sensor to trigger them Tesla used vision. It is getting better but still not on par with what they could have achieved with a $2 sensor. And they probably spent a ton of money on additional software development.
EAP is still amazing.. but IMO it is mainly thank to the radar... but there is only one radar unfortunately. The cameras do an amazing work to detect the lane granted.
Another problem is the following: there are cars stopped at a light which is after a turn. In the current implementation it would cause an accident. To me something like this needs ultra precise GPS positioning (something the car doesn’t seem to have)
Another common problem is stationary object detection. It still isn’t working IMO (unless the stationary object is a car). Current implementation seems to be If the radar detects a stationary object then use the cameras to get the object... only if the object is recognized will the car stop.
So it won’t stop for a trash bag in my case last month for instance.
The stationary object case would be trivial to address with Lidar.
Seems to me a lot of problems would be easy to address with a hardware upgrade... but the only hardware upgrade.. IMO a major hardware upgrade which would not be retroffitable. The only hardware upgrade I have seen mentioned is some kind of faster processor...
I really fail to understand how a faster processor will address the simple problems highlighted above and I don’t think the “if humans can drive with their eyes so can computers” is a valid argument in the current state of artificial intelligence technology.
Look at basic things like blind spot detection. If they used a radar it would be trivial and we would see cars coming from a greater distance. I really wonder how making an automatic lane changes on a German autobahn works out in the current implementation.
A simple case in mind are the windshield wipers. Rather than using a basic $2 sensor to trigger them Tesla used vision. It is getting better but still not on par with what they could have achieved with a $2 sensor. And they probably spent a ton of money on additional software development.
EAP is still amazing.. but IMO it is mainly thank to the radar... but there is only one radar unfortunately. The cameras do an amazing work to detect the lane granted.
Another problem is the following: there are cars stopped at a light which is after a turn. In the current implementation it would cause an accident. To me something like this needs ultra precise GPS positioning (something the car doesn’t seem to have)
Another common problem is stationary object detection. It still isn’t working IMO (unless the stationary object is a car). Current implementation seems to be If the radar detects a stationary object then use the cameras to get the object... only if the object is recognized will the car stop.
So it won’t stop for a trash bag in my case last month for instance.
The stationary object case would be trivial to address with Lidar.
Seems to me a lot of problems would be easy to address with a hardware upgrade... but the only hardware upgrade.. IMO a major hardware upgrade which would not be retroffitable. The only hardware upgrade I have seen mentioned is some kind of faster processor...
I really fail to understand how a faster processor will address the simple problems highlighted above and I don’t think the “if humans can drive with their eyes so can computers” is a valid argument in the current state of artificial intelligence technology.
DGPS is pretty good but can still be off by a meter. The differential signals are government controlled and can be shut off, or go on the blink. Many ships subscribe to different differential signals to have two chances of one working. The other notable issue, which is why visual can be better than GPS, is how confident are we that road locations are accurately documented?
^^This
Unless its in your purchase agreement (it's not) they aren't required to honor the price of anything you did NOT purchase.
The tow kit is a prime example, as MXWing stated. Wheels are cheaper if you upgrade at time of purchase....however, there is no obligation to honor that price later. The price/cost of add-ons and accessories can change at any time. Tesla isn't locked into a price indefinitely for an option someone didn't purchase.
Bet TSLA
Active Member
Nobody thinks it is. But the only disagreement is whether AI gets good enough sooner (a couple of years) or later (a decade). Tesla is betting on sooner.
@xav- I've tried to make so many of the same points and been told a million ways I'm wrong from thousands of people that very likely aren't working in software, hardware, or vision system development fields. The rain sensor is such a classic example of overcomplicating something that had already been so widely developed. It was such a mess that it was even a point during a talk about Tesla's NN development.
I'm a proponent of lidar as well as vision and radar, but that war seems to have been lost within Tesla. The value of higher resolution, active emission sensors can not be understated. Adding that value to the radar and camera systems makes the entire suite so much more robust in the common case, and leaves headroom in the worst case. But again, this war seems to have been lost long ago.
The purpose built hardware, though, does solve for several cases Tesla is no doubt running into with their vision systems. Attempting to use common GPUs as neural nets is sub optimal to say the least, so building devices that are optimized for that purpose will allow Tesla to process more scene data more quickly. Detecting objects sooner with higher reliability and tracking them with higher granularity can help in corner cases where right now the algorithm isn't complex enough or processing enough data to get necessary detail. Basically, if you can use all cameras at full resolution, you get a big benefit. But current hardware is probably barely capable of that. Now add on text recognition and you're likely beyond HW2.5's capability. I could totally be overestimating the requirements of the NN here, but I doubt it.
And that's actually an argument against the "humans can drive with their eyes" point people try to make. Humans have very complex chains of thought that include creating future scenarios. So you know to drive slowly around a corner in a neighborhood because kids could be playing. A machine learning algorithm need to be trained for nearly every scenario it could find itself in, which is basically impossible to do before the system is released to the public. And even then, well after Tesla stops using shadow mode to learn and reinforce learning data, there are going to be scenarios that the system could have had no way to be prepared for. This is why humans are so vastly superior to machines (for now) and even other animals. Robots are good at not being tired, or risky, or getting road rage though. So again, in the common case they're likely already better drivers than us. But common case operation doesn't kill tens of thousands of people a year.
I'm a proponent of lidar as well as vision and radar, but that war seems to have been lost within Tesla. The value of higher resolution, active emission sensors can not be understated. Adding that value to the radar and camera systems makes the entire suite so much more robust in the common case, and leaves headroom in the worst case. But again, this war seems to have been lost long ago.
The purpose built hardware, though, does solve for several cases Tesla is no doubt running into with their vision systems. Attempting to use common GPUs as neural nets is sub optimal to say the least, so building devices that are optimized for that purpose will allow Tesla to process more scene data more quickly. Detecting objects sooner with higher reliability and tracking them with higher granularity can help in corner cases where right now the algorithm isn't complex enough or processing enough data to get necessary detail. Basically, if you can use all cameras at full resolution, you get a big benefit. But current hardware is probably barely capable of that. Now add on text recognition and you're likely beyond HW2.5's capability. I could totally be overestimating the requirements of the NN here, but I doubt it.
And that's actually an argument against the "humans can drive with their eyes" point people try to make. Humans have very complex chains of thought that include creating future scenarios. So you know to drive slowly around a corner in a neighborhood because kids could be playing. A machine learning algorithm need to be trained for nearly every scenario it could find itself in, which is basically impossible to do before the system is released to the public. And even then, well after Tesla stops using shadow mode to learn and reinforce learning data, there are going to be scenarios that the system could have had no way to be prepared for. This is why humans are so vastly superior to machines (for now) and even other animals. Robots are good at not being tired, or risky, or getting road rage though. So again, in the common case they're likely already better drivers than us. But common case operation doesn't kill tens of thousands of people a year.
Bet TSLA
Active Member
Are you sure about that? Impaired and distracted driving in "common case operation" are responsible for a good portion of the fatalities. Just eliminating those provides a huge win for AI over humans.
Bet TSLA
Active Member
Given that we're talking about the safety and effectiveness of alternate drivers, it makes no sense to include the condition of the human driver in whether or not you are talking about "common case". Some human drivers driving down clear roads during daytime are impaired or distracted. No AIs would be. It's a clear difference that will save lives, possibly many lives.
Similar threads
