Some of the things Elon talked about last night I figure are years away. They often talk about things as if they are imminent such as having the car park for you or autopilot in general but I feel they are likely years away. Case in point, battery swap. That was talked about a long time ago and we have yet to see a battery swap station. So if one gets their 85D in December this, in my opinion, is the realistic assumption.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
I'm thinking that what they demonstrated is probably only a month or two away - and that level of autopilot is enough to handle most of the driving on a freeway or country road. That's enough to be a wonderful help and to make me really want one.
The rest of it may take a while - or may never show up with these cars. As far as I can see, the current sensor package has no way of checking left and right at a stop sign or traffic light even if they work out everything else, meaning the cars they build today will likely never be 100% autonomous.
Walter
The rest of it may take a while - or may never show up with these cars. As far as I can see, the current sensor package has no way of checking left and right at a stop sign or traffic light even if they work out everything else, meaning the cars they build today will likely never be 100% autonomous.
Walter
callmesam
Member
The difference between this and battery swap is that while both require developing the technology, this merely requires releasing software that they have already demonstrated. I'd gladly wait a year before I could summon my car from the garage, but I'd like at least the basic safety features and ACC to work soon. It would be more analogous to battery swap if the cars weren't being built with the new hardware, but all cars coming out of the factory have it.
They haven't claimed 100% autonomy is a goal in this car and have always referred to it as "autopilot". Elon has said several times that he thinks self-driving cars as we all envision it is 5-6 years away. So that's three major model revisions in the future by the historical schedule.
That's not the way I heard it. What I heard was that the vehicle is capable of 100% autonomous driving, but due to reglatory limitations and lack of redundancy that it is not active.
I also think imminent, but measured in months, not weeks. I would expect everything in place by the time the X is being delivered in volume, so by next summer or so would be my prediction.
We are so sorry that your Tesla Model S swerved to avoid the pillow in the road, which it mistook for a cat, and in the process ran itself into a ravine. That was not a scenario that we had anticipated, and we will be sure to include such scenario in our forthcoming 6.2.5.3.2 release. Thank you for contacting Tesla Motors.
or
We are so sorry that your Tesla Model S plowed right into the rear end of an 18 wheeler. Our engineers have been made aware of this bug and are diligently working to correct the issue, which will be resolved soon. Thank you for contacting Tesla Motors.
I'm sorry to say this, but given Tesla's track record with software I really don't think I would trust their system.
or
We are so sorry that your Tesla Model S plowed right into the rear end of an 18 wheeler. Our engineers have been made aware of this bug and are diligently working to correct the issue, which will be resolved soon. Thank you for contacting Tesla Motors.
I'm sorry to say this, but given Tesla's track record with software I really don't think I would trust their system.
I have seen him say that he thinks 90% is the most that will be achieved, and that 100% will remain unachievable.
Personally I dislike the idea of the car taking over from me if I am doing something deliberate. For it to differentiate deliberate from accidental may require brain sensors and I am not yet ready to be fitted with those.
This is like so many new technological innovations it will happen over time. I prefer to call it driver assist where the percentage of the car driving itself evolves over time with the improvement of sensor hardware and software.
When the car can drive itself in the middle of a snow storm will be proof fully automous driving is just around the corner.
When the car can drive itself in the middle of a snow storm will be proof fully automous driving is just around the corner.
green1
Active Member
good news, none of the features introduced take any control whatsoever away from the driver. They allow the driver to specifically give limited amounts of control to the car, but they don't take any themselves.
UberEV1
Member
I agree! And AmpedRealtor points to the same concern - autonomous control by the car is a long way away. Case in point is Air France Flight 296, which in 1988 was doing a low-level fly-by at an airshow in France to demonstrate the new Airbus A320 and its advanced computer systems. The computer misinterpreted the pilot's intentions and maintained a safety mode that contributed to the plane crashing into the forest. From the source: "The crew applied full power and the pilot attempted to climb. However, the elevators did not respond to the pilot's commands, because the A320 computer system engaged its 'alpha protection' mode (meant to prevent the aircraft entering a stall.)" and "But in this instance, the pilots involved did not hesitate to fly the aircraft below its normal minimum flying speed because the whole purpose of the flyover was to demonstrate that the aircraft's computer systems would ensure lift would always be available regardless of how the pilots handle the controls."
One might argue that a car on the road faces far more variables (inconsistent street conditions, obstacles suddenly darting in cars path, sensors misinterpreting peripheral signals, weather influences, etc.) vs. an airplane which operates with more open space and can rely on a well-established network of ground-based sensors. And even here, where autonomous flying has been evolving for decades, giving the pilot ultimate control (i.e. the last word) remains fundamental to the control philosophy (I believe - perhaps our pilot friends can confirm).
So, IMHO, 100% autonomous is not in our near future.
P.S. I think Elon had suggested self-driving from on-ramp to off-ramp, which may be more doable given better predictability of highway design, but even that will be hard across multiple Countries with multiple design standards.
lonewolf313
Member
Elon Musk has always referred the new upgraded sensors and relevant software upgrades as "Autopilot" not as a fully self driving car.
That way, a driver must still remain alert of car behaviour and his/her surroundings and be able to take back control by switching off the cruise control and allow the vehicle to regen brake or apply the foot brake should the software fail in any situation to slow down or stop the car
@AmpedRealtor provides amusing examples of situations which have not occurred because the Tesla Auto Pilot feature is not released yet. And when it is it will come with caveats saying that while in use the driver must remain alert and focused on the road and ready to take over at a moments notice if necessary.
And therein is my concern: inevitably some drivers with Auto Pilot engaged will let their attention wander because their hands and feet will not be touching a control surface (with TAAC you are still connected to the car so your attention is more likely to remain on the road). The vast majority of the time Auto Pilot will probably handle driving situation well if it is used as directed (on freeways and open highways and not in slower speed urban environments). Very occasionally something will happen that the Auto Pilot responds to incorrectly and the driver's response will be needed but it will be delayed due to inattention, perhaps just enough to result in an accident.
Of course right now without Auto Pilot accidents occur frequently because many drivers do not give their full attention to the road. Will it occur more often with Auto Pilot engaged? I think it will, but at the same time Auto Pilot may well prevent accidents that would have occurred under full human control because computers can react far more rapidly than humans can.
On balance the accident rate may be less with Auto Pilot cars, or it could be greater. We don't know yet which will be the case.
And therein is my concern: inevitably some drivers with Auto Pilot engaged will let their attention wander because their hands and feet will not be touching a control surface (with TAAC you are still connected to the car so your attention is more likely to remain on the road). The vast majority of the time Auto Pilot will probably handle driving situation well if it is used as directed (on freeways and open highways and not in slower speed urban environments). Very occasionally something will happen that the Auto Pilot responds to incorrectly and the driver's response will be needed but it will be delayed due to inattention, perhaps just enough to result in an accident.
Of course right now without Auto Pilot accidents occur frequently because many drivers do not give their full attention to the road. Will it occur more often with Auto Pilot engaged? I think it will, but at the same time Auto Pilot may well prevent accidents that would have occurred under full human control because computers can react far more rapidly than humans can.
On balance the accident rate may be less with Auto Pilot cars, or it could be greater. We don't know yet which will be the case.
Last edited:
^^ this!
Just wait... After autopilot is released owners will check emails, text, and find all sorts of other fun things to do instead of paying attention to the road. I guarantee it.
Yes, and many drivers do that now. And Auto Pilot will likely prevent some accidents that would have occurred under full human control.
So we'll just have to wait and see how it balances out.
So we'll just have to wait and see how it balances out.
Similar threads
