Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[SoGA Fan Club]

Your point is mostly correct, although even the most rudimentary search would have found the correct answer:
2012 TESLA MODEL S
4 recalls on my Model S, one of which (the parking brake jamming on) actually happened to me. The air bags haven't been addressed yet, the other two (seat belt and steering) were done very conveniently when I had my car there for other things (tire change).
3 on the X, two of which are seat related, different from the recalls of the S, the other one for the parking brake issue.
None for the 3.

So, outside of the airbags, there have been around 4 or 5 correct? The parking break was the same issue on the X and S, right?

And you say you experienced the parking break issue, do you know if any of the other recalls actually had any incidents? It seems like I recall Tesla saying that the bolt issues and the X rear seat issue had no incidents. I can't recall the others.

Either way, compare that with the 1.2 million F150s Ford had to recall - where trucks actually caught on fire. And it was the NHTSA that initiated the investigation, not Ford. Were any of the recalls Tesla has issued been due to an NHTSA investigation?

 

[DaveT]

Employees sour on Tesla amid cost-cutting, layoffs - Reuters

Tesla placed 16th on LinkedIn’s annual “Top Companies 2019” list published in April, compiled from billions of actions taken by its over 600 million users that indicate job interest and demand. It held the fifth and sixth spots in 2018 and 2017, respectively.

At jobs site Glassdoor, Tesla’s overall company rating fell to 3.2 out of 5.0 stars based on reviews written in the first quarter from a high of 3.6 in 2017, according to historical data compiled by Glassdoor at Reuters’ request. The average rating of the nearly 1 million employers reviewed on the site is 3.4.

In the first quarter, Elon Musk’s CEO approval rating dropped to 52% from 90% in 2017.

Tesla’s “recommend to a friend” rating fell to 49% in the first quarter from a high of 71% two years prior, the Glassdoor data showed.

--------------------------

Most folks here will discount this article.

But it's notable that CEO approval rating dropped to 52% from 90%. That's significant.

And it's also notable that "recommend to a friend" dropped as well.

This is from employee reviews/ratings.

Not a good sign.

 

[capster]

Conversed more with a Waymo employee about why there is a need for Lidar right now. And his key points match mine. There isn't an issue with cameras being able to "see" enough information.

The problem is the analogy of "humans see enough with two eyes to drive, so should computers w/ cameras" probably fails in the present day. And it fails not because of the cameras, but because the human vision system is vastly more complex and powerful than most (or all supercomputers). I mean, if you study neuroscience, the chapters on human vision are the most complex, impressive display of neurological evolution.

Simply, why would anyone assume with high confidence that all of a sudden Tesla has enough training compute capability (and on-board inference power) to achieve enough complexity to get near enough to human vision to perform FSD satisfactorily? This is not to say it's not possible, but where is the evidence? Why in the world would you assume this is true?

As a rule of thumb, I take a jaundiced view of the critiques of a particular technical approach by those who are advocating a different approach.

Go ask your friend to white board the Boundary Contour System/Feature Contour System of Grossberg and Mingolla with the differential equations. It’s a fairly sophisticated model of a decent chunk of human vision. If he/she can’t, they don’t know what was already known and computationally modeled a couple decades ago.

I’m guessing Tesla’s folks may be a little more on the ball. Musk in particular seems pretty wide ranging and omnivorous in the knowledge he appears to have absorbed.

Tl;dr: The existence proof offered by human vision is compelling. Ignorance and lack of imagination are weak counter arguments.

Not saying it’s necessary to simulate by numerical integration in the field, btw. You can take "inspiration" from the brain and implement in a different way digitally, e.g. multi-scale oriented convolution kernels.

 

[ZeApelido]

But LIDAR doesn't do much to help with achieving FSD. It can't read signs, traffic lights, lane lines, etc. So until you solve vision you are screwed. And Tesla hasn't "all of a sudden" gotten enough power to solve vision, they have been working on it for years. (Without wasting any time trying to use LIDAR.)

It can get the position of objects / pedestrians (even pedestrian's limbs) at a higher accuracy and repeatability than cameras can (currently). A key problem with cameras is inconsistency (you know, the march of the 9's).

 

[ZeApelido]

The fact that Waymo engineers are falling for the FUD should tell you a lot about them.

I agree, but really basically everyone who isn't paying really detailed attention will, since the media is that biased. Not really their fault.

 

[ZeApelido]

Fewer resources to do what? Lidar’s output is essentially just a, typically low resolution, single channel image, where the intensity value at each pixel is depth(usually inverse depth). How does that save resources?

Accurately position objects and pedestrians. That's what he works on, and is quite confident in their accuracy to do this relative to cameras.

Like, in his opinion, it's not even close.

 

[EVNow]

The problem is the analogy of "humans see enough with two eyes to drive, so should computers w/ cameras" probably fails in the present day. And it fails not because of the cameras, but because the human vision system is vastly more complex and powerful than most (or all supercomputers). I mean, if you study neuroscience, the chapters on human vision are the most complex, impressive display of neurological evolution.

Not really. Think about it, even the really small brained animals have good vision. Infact some of them have better vision than us. It is the general intelligence that is far more complex and distinguishes humans from other mammals.

 

[Boomer19]

Employees sour on Tesla amid cost-cutting, layoffs - Reuters

Tesla placed 16th on LinkedIn’s annual “Top Companies 2019” list published in April, compiled from billions of actions taken by its over 600 million users that indicate job interest and demand. It held the fifth and sixth spots in 2018 and 2017, respectively.

At jobs site Glassdoor, Tesla’s overall company rating fell to 3.2 out of 5.0 stars based on reviews written in the first quarter from a high of 3.6 in 2017, according to historical data compiled by Glassdoor at Reuters’ request. The average rating of the nearly 1 million employers reviewed on the site is 3.4.

In the first quarter, Elon Musk’s CEO approval rating dropped to 52% from 90% in 2017.

Tesla’s “recommend to a friend” rating fell to 49% in the first quarter from a high of 71% two years prior, the Glassdoor data showed.

--------------------------

Most folks here will discount this article.

But it's notable that CEO approval rating dropped to 52% from 90%. That's significant.

And it's also notable that "recommend to a friend" dropped as well.

This is from employee reviews/ratings.

Not a good sign.

i wonder;

- how much of that data it is ‘gameable’
- what job positions led in complaints?
- is it fremont, gf, service and sales??

i can totally see a drop off given the grueling last 18months, right?

meanwhile spacex was rated tops for engineers...
again, is that specific data gameable/subject to mis-representation?

still a tremendous drop-off

 

[ZeApelido]

So?
That statement by itself is questionable given the current state of vision system, yet you listed that like a fact with no supporting evidence.

Even if it's true, what evidence you have that this "advantage" would be enough to bridge the "huge gap" from human vision system you stated in your post above?

You seem to have trouble connecting your point with your line of argument.

Sorry let me clarify, I am not saying anything about whether you need Lidar, whether Waymo is going to make it to FSD first or not. All I am saying is it's likely that the computational resources required to use cameras reliably enough for FSD may be further away than a few years.

I didn't say it wasn't possible.

I didn't say you need Lidar.

I am saying, the overconfidence here (and Elon) in the imminent arrival of Tesla FSD and robotaxis in 2 years is ridiculous.

 

[ZeApelido]

If signal-emitting vision systems were an efficient means of navigating terrestrial habitats, we'd see (pun) a lot more animals using them. Instead, analogous systems (e.g. echolocation) have evolved only a couple of times and are limited to animals like bats with pretty unique ecological niches (see attached PDF if you're interested). Evolution has already rendered a verdict on this. If Waymo engineers think they can outwit 4 billion years of iterative problem solving, more power to them.

No, animals have more efficient processing than computers and that affects the capabilities of what sensors are needed (for now).

 

[ZeApelido]

Employees sour on Tesla amid cost-cutting, layoffs - Reuters

Tesla placed 16th on LinkedIn’s annual “Top Companies 2019” list published in April, compiled from billions of actions taken by its over 600 million users that indicate job interest and demand. It held the fifth and sixth spots in 2018 and 2017, respectively.

At jobs site Glassdoor, Tesla’s overall company rating fell to 3.2 out of 5.0 stars based on reviews written in the first quarter from a high of 3.6 in 2017, according to historical data compiled by Glassdoor at Reuters’ request. The average rating of the nearly 1 million employers reviewed on the site is 3.4.

In the first quarter, Elon Musk’s CEO approval rating dropped to 52% from 90% in 2017.

Tesla’s “recommend to a friend” rating fell to 49% in the first quarter from a high of 71% two years prior, the Glassdoor data showed.

--------------------------

Most folks here will discount this article.

But it's notable that CEO approval rating dropped to 52% from 90%. That's significant.

And it's also notable that "recommend to a friend" dropped as well.

This is from employee reviews/ratings.

Not a good sign.

That CEO rating drop is amazingly bad. Might have to do with his FSD claims.

 

[Compton]

Sorry let me clarify, I am not saying anything about whether you need Lidar, whether Waymo is going to make it to FSD first or not. All I am saying is it's likely that the computational resources required to use cameras reliably enough for FSD may be further away than a few years.

I didn't say it wasn't possible.

I didn't say you need Lidar.

I am saying, the overconfidence here (and Elon) in the imminent arrival of Tesla FSD and robotaxis in 2 years is ridiculous.

So was the development of reusable rockets, and by a small start-up mind you. I will give the man and the team a chance before judging.

As the old saying goes, may the one to develop their own reusable rocket cast the first stone.

 

[mongo]

That CEO rating drop is amazingly bad. Might have to do with his FSD claims.

Or, ya know, all the personnel cuts...

 

[EVNow]

Most folks here will discount this article.

But it's notable that CEO approval rating dropped to 52% from 90%. That's significant.

And it's also notable that "recommend to a friend" dropped as well.

This is from employee reviews/ratings.

Not a good sign.

I bet it has something to do with all the layoffs and SP.

 

[Spacemanspliff]

I bet it has something to do with all the layoffs and SP.

Layoffs SP and didn’t they cut sales bonuses?*


There was a ton of stuff to drive down employee morale, this data is unsurprising.

*edit I can’t remember off the top of my head what it was exactly but they cut compensation for sales employees fairly significantly.

 

[EVNow]

No, animals have more efficient processing than computers and that affects the capabilities of what sensors are needed (for now).

Do you think Tesla computer + cameras/other senors draw more power or Waymo's computers and all the sensors ?

 

[HG Wells]

No, animals have more efficient processing than computers and that affects the capabilities of what sensors are needed (for now).

Depends on what kind of processing. When it comes to math a simple calculator has orders of magnitude more ability than a human.
Of course that's why we thought AI was easy. It isn't.

 

[trailhound]

this last year has been a fornicopia of leap first look later decisions.
I doubt FSD has any weight in the day to day employee's feeling about his (or her)CEO.
I hope we can see a more Zen like leader this next year.

 

[EVNow]

Layoffs SP and didn’t they cut sales bonuses?*


There was a ton of stuff to drive down employee morale, this data is unsurprising.

*edit I can’t remember off the top of my head what it was but they cut compensation for sales employees fairly significantly.

Yes, sales people would not have liked the cuts. But Sales employees usually move on if they don't like the place - most of them don't stick around for too long anyway.

 

[StealthP3D]

View attachment 417736

Trouble with English? Maybe you don't understand that the word "will" is future tense?

We need less trolls on this forum.