delete - error
totally. My brother in the UK had a big phantom braking scare in a model 3 at speed, a few years ago, said it scared the absolute bejesus out of him and felt like he had zero control over some software possibly making him a road death statistic.
I like the idea of a car that randomly jams on the brakes every time it has an existential crisis. What could go wrong?
exactly. A safely stopped car is an infinitely better scenario than a car travelling at speed that isn’t able to figure out what to do next
Use some of your compute power. Obviously, stop with range it's confident of.
Or stop dead, similar to the lidar equipped car in that video.
definitely an edge case, really it’s just an internet video showing that cameras can be ‘fooled’, which isn’t exactly new,
Look at it this way, for your €7,500 feature, would it inspire confidence in your safety ….
So in what could be a life or death situation, the car is just going to abstain from driving and hand over to the driver with little or no time to react?
It should be possible for the system to say, ermm, I don't know about this, can you take over or I'm stopping.
It was HW3 v HW4. The later version 'saw' the wall, although the quality of the wall wasn't as good as in the above test.
A bit too 'edge casey' for me anyway. Not a fan of FSD either. Can't see it working globally when you consider the kinds of roads in existence around the world. Even in the US, roads are pretty poor generally. Lots of concrete with damaged surfaces.
I’m not a fan of FSD, no interest in self driving cars of any brand but wasn’t this video debunked?
Or was it a HW3 V HW4 difference?
Some noise of the car being on AP and not FSD.
Wouldn't trust it. And clearly not a substitute for the human eye based on this Wiley Coyote test. And others.
I can't beleive that's a serious comment.
What you've quoted there is actual fact. It's the basis of Tesla's vision based fsd system. Their plan is that it will be in operation in Austin in June and there's no suggestion they're pulling back from that timeframe. Like it or not, we'll be seeing some results of this quite soon.
Just short of 8k done in my model 3 in 8 weeks.
Hasn’t missed a beat. Very comfortable.
92% of my charging has been free also which is a fair benefit.
I'm sure they might - indeed I sold a car to someone who works in the local CU a few years ago - but I don't have an account there. It's just more friction, I like the click and buy that it used to be, you know?
Credit unions in your area might have EV or green loans and they have much more reasonable interest rates.
I'd hope it would, my 222 rwd would do the 250km too with no issues! Not going to lie I'd quite like the idea of an M3P but what I do want, another MX, is the main target. It just irritates me that I can't get a CPO one without either paying 40k in cash or taking a personal loan at ridiculous interest rates because they no longer finance through tesla.
That was a good price considering EAP and the CPO warranty. The one I was watching was €28k. With about 105k on the clock.
It will do the 250km easily too. I used to get Dublin Belfast and back without stopping once I left at 100% and stuck to the limits or about 115. It averages 19.5 kWh /100km during my 2 year ownership. No preheating, no conditioning, just hop in and drive it like any other car. It would do 16.1 kWh/100 km on regular trips to Kilkenny as that’s generally a lower speed average trip.
There was one for 24k with 120k on the clock in white. I was considering it as there may come a time where I have to go to the office a bit more regularly than the current once or twice a quarter. Round trip is 250km so it's quite expensive in the outlander phev and not really feasible in the leaf. I've just sold my P38 range rover (4.6 v8 petrol) which would have taken probably the guts of 80-100 quid to do that trip).
Regarding the kms, I do so much mileage (outside of my main job which is thankfully 98% WFH) as it is that anything used would be at 100k within a year or so so I wouldn't be pushed on 100/200/300k on the clock. To be honest I prefer to buy higher mile cars as no one else wants them (so they are cheap) and any car of mine will be high miles after me anyway.
That M3P has over 100k on it, so I’d say that’s why it was priced in the late 20’s but I don’t think it was “cheap” per se.
I sold my 2020 M3P with 45k in it last August for €33k and some change so I’d say that’s where they are now. But the EAP and additional 1 year warranty would be nice on the CPO.
The only two I was interested in (a suspiciously cheap M3P and a MX100D) both vanished over the last 24 hours. I'd have already bought the X but there's no finance available through Tesla for S/X anymore (confirmed it with both Tesla and FI)
website messed up during price adjustments and finance rate changes. The MYP I was looking at has steadily been dropping from €53,100 to €49,100 last night.
What's going on here? (tesla cpo page)
Months?
totally agree and understand that.
The argument is still that somehow all those regulations can be met using a camera-only implementation.
The only advantage camera's have over other components is that all of our road infrastructure is designed to convey information to drivers using visual means. Red lights, warning signs etc…. I don't think it would be possible to successfully roll out a system without some kind of visual processing.
The UNECE regulations are quite clear on meeting operational performance goals and are agnostic on the nature of the systems used to meet those requirements. It would be most unusual at this point in the process for the working groups handling the topic to throw away all their previous work and go down the route of a prescriptive technology stack.
the difference is the speed ffs! There’s almost zero safety risk when walking in public, and the main risk is being hit by vehicles.
All of your arguments here for months now ignore the safety-first element of vehicle regulations and concentrate only on the assumption that Elon has the answers to all camera limitations.
No regulator will sign off on a camera-only system for a driverless car ( and the US as we know from the Cybertruck, has no pedestrian safety regulations to begin with 🤪).
TLDR: I’m don’t doubt camera-only can get close to a workable technological solution, but they’ll never be approved while a better road safety solution exists.
You're missing the point. The use of eyes is more than just 3D. Its our eyelids(cleaning), its eye movement, its head movement, its hand-eye coordination, its the brain behind the eyes.
3D is a very small part of it.
You are over simplifying the task at hand. One or two forward facing static cameras cannot adequately compensate for all the characteristics of human eyes and movement to cover the edge cases and its the edge cases are the problem.
What's lazy? Humans using only eyes while driving?
We build a sufficient model of environment to plot a safe course. We depend on orderly behaviour enforced by rules. We can probably handle up to two things diverging from normal expected behaviour at one time. We employ some preemptive what-if caution, slowing down when something is suspect.
I don't see why a visual system would be fundamentally incapable of reproducing this kind of capability with suitable processing.
