O.t.tesla autonomy vehicle investor day

[bob heinatz]

Not a fan nor do I want to listen 2+hrs. It will only take one fatal crash and the public will turn on these cars and the courts will bankrupt the company.

[MattTuck]

In other news, the porsche Taycan is starting production in the fall. Some have called it the Tesla killer. Not sure of that, but I do think the styling is better than what I've seen out of Tesla.

Unclear whether the Taycan will sell without the charging network that tesla has, but it is probably the first real apples to apples high performance all electric competition that Tesla will face.

[FlashUNC]

Originally Posted by mistermo

With respect, he doesn't live in reality, which is why he's one of the greatest innovators of our time.

Lol.

[fkelly]

Since I don't have 2.5hrs to spend on this, can someone post a summary?

Okay, I watched in two segments. I never would have hung in there for 2+ hours if it wasn't fascinating and informative.

I was a Musk and Tesla skeptic before. Now I might even consider buying a Tesla ... though there are some practical questions related to quality control and recharging that I'd need to address.

As someone else said: Musk has quality people with PH.d's from top level institutions and years of practical experience working for him. Two of them presented the hardware and software underpinnings. They are heavily into neural networks which basically mean "training" the software by giving it millions of examples. They use their current fleet as the basis for generating the examples and the software gets better exponentially by learning from real life examples (and some human "coaching"). They are using cameras (as opposed to lidar) and de-emphasize the role of GPS since, if GPS is "wrong" it can lead to bad things happening.

Musk is picturing a world filled with robotaxis and the removal of steering wheels and accelerators within a fairly short time frame: 2 or 3 years. And full self-driving capabilities within that same time frame. We'll see soon enough ... but what you buy now will be able to be fully self driving based on software upgrades if it turns out to work.

He states that buying any other model car now is like buying a horse. "I mean if you want to buy a horse buy one" ...

It's a good investment of 2 hours to watch for yourself.

[Tickdoc]

I love my "horse".

[fkelly]

Elon makes a bunch of fantastical claims unsupported by reality.

Notably that they can solve with software what others are pursuing via a combination of hardware and software for autonomous driving.

I don't want to take on the role of a Tesla fanboy. I still have open questions. But the presentation makes it clear that both hardware and software are involved. They spent two years developing the computer (hardware) that underpins the self-driving car ... it's fully redundant within a given car and appears well thought out in terms of security ... only software that is cryptographically signed by Tesla will run on it. There are, I believe, 8 cameras plus radar in the car looking in all directions. So it's not just or even primarily software.

People are driving the cars with some of the self-driving features activated today. And while there have been some well publicized accidents, overall I think they are a lot safer than human driven cars.

[mistermo]

The self driving feature on my car has gotten appreciably and noticeably better in the 9 months I've owned it. There's serious road construction near me and the construction barrels are moved daily. The car "sees" this and can navigate through this fairly well. No way GPS could anticipate this.

I was a skeptic too, at one time. I think there's a decent chance Tesla may not survive in it's current incarnation, but there should be no denying it's disruptor and innovator status. I'm driving to OH and back tomorrow. Presently n=1 when looking for eCars that could fill that duty.

[FlashUNC]

Originally Posted by fkelly

I don't want to take on the role of a Tesla fanboy. I still have open questions. But the presentation makes it clear that both hardware and software are involved. They spent two years developing the computer (hardware) that underpins the self-driving car ... it's fully redundant within a given car and appears well thought out in terms of security ... only software that is cryptographically signed by Tesla will run on it. There are, I believe, 8 cameras plus radar in the car looking in all directions. So it's not just or even primarily software.

People are driving the cars with some of the self-driving features activated today. And while there have been some well publicized accidents, overall I think they are a lot safer than human driven cars.

A guy got decapitated watching a Harry Potter movie rather than driving and paying attention to the road and the car missed a tractor trailer in front of it turning left.

If you trust that company to deliver fully autonomous driving in a year when other automakers are saying it's years off, if ever. Sure.

[unterhausen]

the most recent death I'm aware of was because Tesla relies on cameras, but sometimes cameras just can't see a concrete embankment and there are no white lines on the road to guide the car with either. Relying on plastic barrels and road markings being there is a recipe for disaster, but that's what they are doing. I am pretty sure any successful system is going to have Lidar. I feel fairly confident that a person would have picked up on the truck being across the road too.

[fkelly]

There are long discussions, at several points in the presentation, on Lidar. Listen and judge for yourself. Musk and his staff are dead set against Lidar.

There are going to be accidents and even deaths no matter what technologies are used. The question is whether they will be fewer and less severe than continuing to rely and human drivers and what the best combination of technologies is. We won't settle that here.

[MattTuck]

Burned 10 million greenbacks in q1. PER DAY.



Even if they are making 10% profit on a $60k car, they would need to sell 1,700 cars just to make up one days cash burn.

[mistermo]

There's many who think their margin per car on the Model 3 is closer to 30%. No one knows for sure, but given the significantly reduced number of parts in each car, that's not unreasonable to assume. Their overhead costs, somewhat fixed, are the killer. At this point, volume is what they need (and are getting) to absorb the fixed OH costs.

Originally Posted by MattTuck

Burned 10 million greenbacks in q1. PER DAY.



Even if they are making 10% profit on a $60k car, they would need to sell 1,700 cars just to make up one days cash burn.

[alancw3]

interesting article on google news this morning to somewhat support elpn musk/Tesla direction on autonomous cars:

https://gizmodo.com/elon-musk-was-ri...dar-1834266742

[peanutgallery]

Did you guys learn nothing from the Terminator movies?

SkyNet is tricking Elon into all of this

[soulspinner]

fit and finish isn't great yet and they need $ now. The Tesla 3 doesn't seem to be of the quality of their other offerings, either...…….YMMV