That being said, it has made me re-realize that humans are really terrible drivers as well, myself included, just less terrible on average, currently.

But driving skill training can most certainly greatly increase your skills and awareness and thus immensely reduce your risk of mistakes.

Anyone who drives should take as many car control and accident avoidance clinics as they can and refresh up every few years (if not more). These should really be required by law but since they're not, you can increase your odds of living by being proactive about it.

It's weird that this is somehow controversial for car driving. For e.g. airplane pilots there is a long and strict training and skill maintenance routine, because obviously this makes for safer pilots. You don't just get to read a booklet and go flying.

Seems like there would be economies of scale if the federal government created some really realistic simulators programmed with all the common accident situations and then everybody could be required to pass a test every four years or something.

I imagine it would be so much more enjoyable and hopefully more effective than the defensive driving classes that exist now.

Seriously:

2021 Lamborghini Aventador: 2.8 seconds

2020 Tesla Model S Performance: 2.4 seconds

(Car & Driver figures)

You're more scared of rich boys in lambos than of a Tesla that "disengages in many normal-to-a-human situations with no ability to resume"?

Forget FSD - whether it's Tesla or others, an automation in the car that stops when a living being is detected in front seems like an advance, while human drivers talking in a sales meeting while wearing wireless AirPods appear to be an increasing threat to pedestrians.

It should not be legal, to do testing on public streets with unwilling participants.

Tesla should be required to be able to enable this only in closed areas where access is limited to those participating in the testing.

Let's assume Elon and Tesla are perfect, you still need the law for less perfect CEOs and companies right? This less perfect ones might do bad things.

Step 1: you get some independent persons from the government (Tesla or Google, etc pays for this so no need to cry taxes)

Step 2: you drive the cars in the area where you want to approve it for testing/production , you record all mistakes and times the car ahd to intervene

Step 3: drive all year around, all weather conditions, all traffic conditions

Step 4: you analyze the data, see how many times you had to intervene and compute the damages that would have occurred. Get some real statisticins that could put bounds around the damages.

Using the data from 4 you ask the population to vote what is the correct value they are comfortable with.

What Tesla is doing now do not match because:

1 the tests are not transparent, we don't know or we can't trust Tesla for the number of mistakes the AI is doing

2 updates invalidate previous tests and stats

3 the computer will not drive in all conditions and is differing harder parts of the road to the human so this also is screwing the stats, if you let the AI to drive the easy part and the human the hard part then for sure the AI has a good record.

Even if Tesla is perfect you need this laws for the imperfect companies that would move fast... so think about this as laws that will protect you from the evil Google car or some Chinese brand

Why should I let Tesla put my life at risk if they're not sure it can be operated safely?

The burden of proof should be collecting evidence that they are safe, not waiting for data soaked in blood before pulling back.

Any other company wouldn't be able to pull off this kind of scam like Tesla is doing.

2025: Elon announces BabyX, in vitro baby factory, claims he's cloning any baby killed by a Tesla vehicle, for free.

Yeah dude, not sure where you are from but there are special traffic rules related to children , like how to drive around schools, school buses and children, because children are doing stupid things and schools are marked with special signs so the driver knows to pay more attention in this area.

EDIT: Seems like my "data-driven decision making" advocacy really struck a nerve with the cherry-picked anecdotes and rhetoric crowd. :)

1 the count for the times drivers intervene to prevent a crash are secret, or are not counted at a crash.

2 I don't want on my village/city some american car that is safer then the american driver that included 16 year old drunk teens and people that don't even have to pass exams and medical tests. The self driving car should be compared with healthy drivers , I don't care that some state/country is incapable of stopping speeding or drunk driving.

3 updating this cars should invalidate previous approvals, as a software developer I know well how updates include bugs not only features.

Anyway I am against companies deciding from themselves how is a capable alphya testers and how they should run alpha tests.

If we want real metrics then incidents liek the ones in the video should be counted against Tesla, is MALICIOUS to show a stats of average driver using an average car versus an expensive car with a combo of human+computer and assign the good to the computer and ignore the many times the human had to save the computer ass. Fuck shit statistics and shit companies and fanboys using them. At best you can prove the computer might be a decent copilot versus no copilot.

We measure. This requires transparency on the part of Tesla, but if we frame it as "if you want legal FSD, you have to comply with transparency". We regulate like this in all other industries so there's no reason to think we can't do it here.

> We should not assume is safer by default

No one is advocating "assume it is safer by default".

> I see many problems with Tesla fanboys numbers

Right, but by framing it with metrics rather than cherry-picked incidents, we can have a national conversation about which metrics are most appropriate. If Tesla fanboys bring numbers that you take issue with, you can debate them. As an aside, it's hilarious that my earlier comment was being downvoted for advocating a debate based on data rather than cherry picked anecdotes, but I should know my audience better :)

> the count for the times drivers intervene to prevent a crash are secret, or are not counted at a crash.

See my transparency remarks above.

> I don't want on my village/city some american car that is safer then the american driver that included 16 year old drunk teens and people that don't even have to pass exams and medical tests.

Ignoring the obvious barbs at Americans, if we quantify the safety of any given FSD system, we can compare that against local safety and prohibit FSD systems which score worse than the local average such that FSD will only ever be a boon.

> updating this cars should invalidate previous approvals, as a software developer I know well how updates include bugs not only features.

Seems reasonable to me.

> Anyway I am against companies deciding from themselves how is a capable alphya testers and how they should run alpha tests.

In this case, I agree.

> Fuck shit statistics and shit companies and fanboys using them.

I love these kinds of constructive comments. Keep 'em coming.

Good fucking luck, that's going to be harder than pulling Elon's teeth out.

Telsa has regularly lied about how well the system works, and will continue to do so unless we strong arm some independent verification on them.

And good luck getting that to happen. Governments roll onto their backs for Musk like it's the second coming of christ.

FSD should not be on the road without transparency. I don't want to get in car accidents because Musk wants to beta test crappy software.

Lol, nonsense. Musk moved Tesla HQ out of California precisely because he couldn't get the local or state government to play ball (a California politician publicly Tweeted "F*ck Elon Musk"--hardly "rolling onto their backs"). Moreover, if you can't depend on governments to defy Musk and require regulatory oversight of FSD, how are you going to depend on them to outlaw FSD altogether?

> Seems reasonable to me

Year 2026: autonomous drivers have 80% fewer accidents per million miles than human drivers.

Bug fix to improve that to 90% and prevent an additional 500,000 casualties per year must wait until 2028 for NHTSA certification.

What "autonomous drivers" means in your numbers? Does it mean autopilot without any driver in the car that intensives like in TFA? Because this is the shit stats ignore the cases where the autopilot was saved from 5 crashes in 5 minutes by it's driver, in this case you could say the combo of driver+auto-co-plot is safer , also you should compare same cars and same roads , self driving software gives up on harder conditions so this inflates your stats.

TL:DR Tesla should also show the disengagements numbers so we can add those in the stats, some of those would have been crashes if the driver did not intervene.