A Tesla car is likely to last 600,000 miles - and with next-gen batteries, for a million miles. An average gasoline car needs engine and/or transmission work that typically costs more than the car is worth at between 150,000 and 200,000 miles.

So unless you can show that a Tesla needs between 3 and 8 times as much resources to make as a regular car - this is a poor argument.

And that's even ignoring the CO2 savings over it's life. A gasoline engine is typically only 25% efficient - and the oil refinery only about 70% efficient. So overall, you're only using aout 20% of the energy from the crude oil to propel the car forward. Even the NASTIEST "brown coal" power station is 65 to 70% efficient - and the transmission grid, charging system and electric motor is about 90% efficient.

So, again, even with the NASTIEST, DIRTIEST electricity available - an electric car is about 3 times more efficient than a gasoline engined car.

Limiting births results in a rapid increase in the average age of the population. Because people retire and need more health care support as they get older, they impose an increasing load on the younger people as the population declines...which is not a good thing. In places like Japan where the birth rate is VERY low - this is becoming a crisis. Way too many young people are having to move into jobs like nursing and geriatric medicine to support the slew in the average age slew.

COVID-19 kills (predominantly) older people - so as a means for reducing population, it does have the advantage of LIGHTENING the load on younger people - that makes decreasing birth rates much more palatable - from an economic standpoint.

Haven't GNU been telling us for decades that "Linux" is just the name of the kernel? So if Microsoft replaces the Kernel with something else - then how can we call it "Linux" anymore? At besti it's GNU/UnixCloneForWIndows.

That's not going to help - or at least it's not going to help MUCH. The distinctive property of this virus is that most people who have it are asymptomatic - and even those who do go on to get sick can take a couple of weeks to show symptoms...that means no fever in the vast majority of people who are spreading it.

The "test kits" that are out there use a finger-stick blood test that takes 15 minutes - and that can identify asymptomatic carriers as well as sick people.

79,553 cases are of people with symptoms.
It's believed (eg from people getting out of those cruise ships) that MANY people who were infected are completely asymptomatic - which means that we actually have no idea of the number of people who get the virus but never show so much as a sniffle.

So 3% is the HIGHEST the fatality rate could be - maybe it's a hell of a lot less. Remember, in just a regular flu season, between 12,000 and 60,000 people die in the US alone.

That note puzzled me too.

"One" of the best implies that there are others...and there really aren't.

Waymo, Uber, Apple, etc don't have production cars on the roads and between the three of them only have a few million miles of data to train their AI's.

Tesla have a million cars on the roads - each collecting 1000 miles a month. That means that Tesla collect more data in one DAY than the other AI self driving systems have ever collected in their entire history.

AI systems NEED lots of data to train on. It's not enough to have a video snippet of a deer running in front of the car - you need one of it coming from across the other side of the road, another with it coming out from the near side - an other when the deer has a fawn in tow, another when it's in rut and is super-agressive - then all of those in snow...all of them in rain...in darkness,.in twighlight...

20 million miles of driving is nowhere CLOSE to enough to handle a fairly basic problem like that. Then you also sell these cars in Australia - where camels on the road are an issue.

Waymo say that their cars drive billions of SIMULATED miles to gather data - but how do you know that your simulation of a deer crossing the road is accurate enough?

The other systems described in the original post aren't REMOTELY close to being self-driving. Changing lanes, merging into traffic, roadworks...none of the other systems (except Tesla, Waymo, Uber and Apple) are even 1% of the way to being able to overtake or merge into traffic on arbitrary roads.

Maybe because it takes multiple minutes to stop a 2 mile long coal train - and two minutes ahead of you is too far to resolve with cameras or radar?

On a Tesla it's adjustable between 1 and 6 car lengths - but what is a "safe" distance changes when the car has sub-millisecond reaction times and a radar system that bounces radar waves off of the road under the car in front so it can see TWO cars in front. It often starts to brake when the car TWO in front of you starts to slow down and the guy immediately in front of you hasn't even noticed yet.

It kinda freaks you out a bit the first few times it does it...it's like "Why did this stupid car just slam on the br...OH! THAT'S WHY." ... It feels like the car is telepathic and knows what the guy in front is going to do before he even does it.

With those kinds of smarts - leaving a longer distance is just there to make the humans in the car feel a bit less jumpy. The car doesn't NEED that much gap.

The fact that this aftermarket system DOES need it - suggests that it's not that good.

That's why Tesla have a test facility in Delta Junction Alaska.

https://cdn.carbuzz.com/galler...

https://www.cnet.com/roadshow/...

Yeah...what were you saying about ignorance...THAT.

Without some kind of side/rear camera system, changing lanes is impossible. You absolutely need to be able to see what (if anything) is coming from behind on the adjacent lane. So unless the vehicle has pretty much 360 degree camera coverage - it's either a death trap - or it can't overtake or merge in traffic.

The Caddilac SuperCruise relies on extremely high resolution maps - like better than 10cm resolution. It only works on 200,000 miles of US highways - there are 4 million miles of roads.,,so this isn't a system that's going to be expandable to allow the car to drive itself anywhere else.

So it's not following lane markings with a camera or anything of that nature. It's blindfolded and following a map.

So what happens if the road layout changes? Well, until Cadillac come out and re-map it - you're doomed...I mean, literally, it'll kill you! How can it handle roadworks? It can't. Furthermore - if it relies on map updates - what happens with a 10 or 15 year old car? Once Cadillac realize (as they inevitably must) that maintaining high res maps is a costly activity - and with a large percentage of the SuperCruise systems in the car crusher - will they stop re-mapping these roads?

SuperCruise is a desperation measure from a company who don't have the AI capability to do it properly. It's an insane way to try to solve the problem.

So - to re-review these systems:

* Single camera systems can't handle lane changing - so *BZZZT* FAIL!!
* High resolution road mapping is a lethally stupid idea - so *BZZZT* FAIL!!
* Tesla's approach can work - and it's getting better with every software update.

I drove 2,400 freeway miles over Xmas with my Tesla Model 3 FSD. It drove flawlessly - overtaking other vehicles, merging into traffic on freeway on-ramps, taking the correct lane through freeway interchanges, etc, etc. I only had to take over as it exited the freeway (because the system currently only works on freeways) - and three other times. Once when going through a border patrol inspection place and I didn't trust the car to make it through the maze of cones and who-knows-what scanning equipment there. Once when it was being annoyingly anal when weaving through a pack of semi-trucks and I just wanted it to get out into the far left lane and floor it. And once when a car on the opposite side of the freeway spun-out and created an enormous cloud of dust that blew over my side of the road - and I didn't want to find out the hard way whether the car could handle that.

To even consider comparing those other two pieces of crap to the Tesla system is just annoying. There is no comparison because they can't do more than about 5% of what it can do.

Sure, you can theoretically do it with four points at the vertices of a tetrahedron - and 100% of the surface is visible 24/7.

But they have to be so far away - that speed-of-light constraints kill you.

The ENTIRE point of StarLink is to keep the altitudes as low as possible to minimise latency.

OK - so the study shows that drivers were doing more multitasking and were more distracted. But do we care? If the accident rate as a result of the assist features goes DOWN despite increased distraction - then we still want it. Ultimately it would be a very good thing if the automation could do a good enough job to make distracted driving perfectly OK.

Admittedly, I have not data to prove that - but this sturdy doesn't have data to proven that I'm wrong.

The IIHS (insurance company study group) awards extra safety points for cars that have these features - and I'm pretty sure they wouldn't do that if their evidence showed more accidents in cars that have it.

94% of accidents are due to human error...preventing the human from driving the car would evidently be a good idea.