2017 Investor Roundtable:General Discussion

[elasalle]

Source: "Pilot" Model 3 Body Line Still In Development Near Detroit - DailyKanban

What does and does not constitute a “fundamental issue” with the Model 3 production ramp is open to debate, but a source tells Daily Kanban that elements of the Model 3 body line are still in development at the Michigan-based supplier Thai Summit America and not yet installed at Tesla’s Fremont facility.


Author is Articles by Edward Niedermeyer - Bloomberg View so it could be FUDballing ...

 

[MP3Mike]

Source: "Pilot" Model 3 Body Line Still In Development Near Detroit - DailyKanban

What does and does not constitute a “fundamental issue” with the Model 3 production ramp is open to debate, but a source tells Daily Kanban that elements of the Model 3 body line are still in development at the Michigan-based supplier Thai Summit America and not yet installed at Tesla’s Fremont facility.


Author is Articles by Edward Niedermeyer - Bloomberg View so it could be FUDballing ...

It sounds to me, if it is true at all, that Thai Summit America is making a second Model 3 body panel line that will be what Tesla uses to move from 5k/week to 10k/week. But I'm not sure much in that article is actually true.

 

[TNEVol]

My hope is that Tesla is forced to do this as it goes down market in its vehicles, and also runs into actual business needs with Tesla Semi with shops that already service their own. The first time Tesla is having a contract fight with a Semi fleet owner and Tesla looks at their service department absolutely not able to meet the demand, they'll see the writing on the wall, finally, and then as the poor people start trying to fix their cars (which are their only cars used daily as commuters and not Sunday church rides like for S & X), Tesla will once again have to fight it on an individual basis one on one at every service center, and they'll be forced to capitulate. I'm not at all sure why Tesla has been so blind to this problem, but I have a feeling independent service will be happening anyway. If Tesla continues to be blind to this, there's going to be quite a lot of brand damage while they wake the f up.

I do not think S&X drivers are Sunday drivers. I would bet most are daily drivers like mine. I have simply had no problems now approaching 100K miles in 4 years.

 

[dc_h]

Source: "Pilot" Model 3 Body Line Still In Development Near Detroit - DailyKanban

What does and does not constitute a “fundamental issue” with the Model 3 production ramp is open to debate, but a source tells Daily Kanban that elements of the Model 3 body line are still in development at the Michigan-based supplier Thai Summit America and not yet installed at Tesla’s Fremont facility.


Author is Articles by Edward Niedermeyer - Bloomberg View so it could be FUDballing ...

He hasn't written a positive article about Tesla yet, why start now.

 

[Todd Burch]

Source: "Pilot" Model 3 Body Line Still In Development Near Detroit - DailyKanban

What does and does not constitute a “fundamental issue” with the Model 3 production ramp is open to debate, but a source tells Daily Kanban that elements of the Model 3 body line are still in development at the Michigan-based supplier Thai Summit America and not yet installed at Tesla’s Fremont facility.


Author is Articles by Edward Niedermeyer - Bloomberg View so it could be FUDballing ...

Total FUD. And confirmed by Wall Street Journal? Haha say no more.

They also mention testing for 12-18 months, so yes probably a second line and old bag Neidermeyer is playing his usual games. That guy needs an ignore button...

 

[kenliles]

self-driving is unequivocally harder than rocket science. We've been launching rockets since the 1950s. The equations and physics are all known. The same is not true for self-driving

As a former rocket and research scientist, I disagree.
We have been launching rockets since the 1950s, but haven’t self driven them back to earth until now.
The core technologies are closely entertwined and provide working solutions to both problems in equal measure. Resource application to deploy them is just coming to fruition.

 

[ValueAnalyst]

The following are comments from my client who just purchased a 100D. Note that he was a huge Mercedes fan and had planned for his self retirement gift to fly to Germany, tour the company's factories, buy a high-end Mercedes, drive around Europe and ship it back to the states.

Then I got him into Tesla stock...

- interior is as good as Mercedes and in some respects even better

- Autopilot drives very well on the highway, needs work on side streets

- handles other drivers who don't follow rules very well

- need to get used to its driving style (for example, doesn't steer closer to the outer lane marking when passing huge semi trailers)

- takes about a week to get used to regenerative breaking, and probably will never need break pads again

- hang out with [educated] people who still don't know anything about Tesla (remember: this is the east coast)

- model x has way more tech in it than model s (which confirms Elon's comments about hubris)

- couple of small issues like not being able to open the garage door or not having the automatic door shut if not closed properly

- not having a gear box is life changing

- usual excitement around acceleration and torque

He just lost his wife so it's good to see him so excited about Tesla.

 

[Todd Burch]

As a former rocket and research scientist, I disagree.
We have been launching rockets since the 1950s, but haven’t self driven them back to earth until now.
The core technologies are closely entertwined and provide working solutions to both problems in equal measure. Resource application to deploy them is just coming to fruition.

You disagree that self driving is harder than rocket science? I would agree that safe Level-5 self driving is at least an order of magnitude harder than a rocket that can land on a barge.

 

[erthquake]

The following are comments from my client who just purchased a 100D. Note that he was a huge Mercedes fan and had planned for his self retirement gift to fly to Germany, tour the company's factories, buy a high-end Mercedes, drive around Europe and ship it back to the states.

Then I got him into Tesla stock...

- interior is as good as Mercedes and in some respects even better

- Autopilot drives very well on the highway, needs work on side streets

- handles other drivers who don't follow rules very well

- need to get used to its driving style (for example, doesn't steer closer to the outer lane marking when passing huge semi trailers)

- takes about a week to get used to regenerative breaking, and probably will never need break pads again

- hang out with [educated] people who still don't know anything about Tesla (remember: this is the east coast)

- model x has way more tech in it than model s (which confirms Elon's comments about hubris)

- couple of small issues like not being able to open the garage door or not having the automatic door shut if not closed properly

- not having a gear box is life changing

- usual excitement around acceleration and torque

He just lost his wife so it's good to see him so excited about Tesla.

Poor guy. I'd be lost without my wife. Glad to hear he's finding some joy in life.

 

[neroden]

The lines referred to in these conversation are BIW assembly lines. The factory had originally a BIW line for Model S with the nominal production capacity of about 1200 cars per week. The new BIW line, with capacity of 2500 cars/week, was added in preparation for Model X production. The final assembly line was also upgraded to a nominal capacity of 2500 cars/week (from original 1200 cars/week).

The plan as stated by Tesla back then was to ramp production of Model X bodies on the new BIW line while continuing manufacturing MS body in white on the old BIW line. This was done to minimize impact that ramp up of Model X production could have on MS production. As stated by Tesla at that time, the plan was to blend production of MS from old BIW line to the new one once the production of MX stabilizes. This was supposed to free space occupied by the old BIW line for the Model 3 production.

In reality space for production of M3 was found elsewhere at Fremont, and urgency of blending BIW production of MX and MS went away. So currently MS and MX bodies are being produced on separate lines. Last I heard from the factory tour personnel is that above blending will occur at some time in future.

I will say that "body line 1" takes up an awful lot of physical space compared to the more efficiently designed "body line 2".

I believe S was kept on line 1 partly because of the ongoing teething troubles with the X (giving them more leeway for altering the X design with body line shutdowns without disrupting the S line). If-and-when they move S to line 2 they will do so because it is ready to handle it, so the transition should be smooth. Then they'll clear out the line 1 space. They could use it for anything, really -- I doubt they've decided yet.

 

[jhm]

Hey, Elon, the Governor of Puerto Rico is tweeting you.

Ricardo Rossello on Twitter

 

[neroden]

As for the space to be freed up after elimination of the need for old MS BIW line, my speculation is that it might be used to install second production line(s) for M3, the one(s) that is (are) supposed to achieve 10k cars/week. My thinking is that the current 5k cars/week line(s) will be operating until the new 10k line(s) is (are) installed. The 10k lines will then be ramped up, and at certain point later the old 5k line(s) will be upgraded to 10k as well. My hypothesis is that as a result of this Fremont will be able to produce 500k M3 in 2018, 500k MY in 2019, and about 120k+ MS/MX.

I've actually been speculating that that second Model 3 production line will go into the parts of the factory floor currently used for parts storage. That was an *enormous* section of the factory floor, and we already know they just built giant automated parts warehouses.

 

[techmaven]

I've actually been speculating that that second Model 3 production line will go into the parts of the factory floor currently used for parts storage. That was an *enormous* section of the factory floor, and we already know they just built giant automated parts warehouses.

I agree... Niedermeyer's article seems to point to some of the pieces for the next increase in production rate in 2018, not the current one going through the initial s-curve in Fremont. His article contradicts itself several times in its desperation to paint Tesla in a poor light, as usual for him.

 

[Reciprocity]

@Reciprocity I would take a different view on most of what you said. I think it's a false equivalency to assume that LiDAR equals a hardware solution and cameras equal a software solution. Self-driving is a software problem regardless of hardware.

The issue is that using cameras for object identification (and off-axis distance) vastly complicates just that portion of the problem. For both approaches, the rest is still software.

Further, self-driving is unequivocally harder than rocket science. We've been launching rockets since the 1950s. The equations and physics are all known. The same is not true for self-driving. No one has achieved self-driving in 2017. Nothing more needs to be said.

And I would also disagree with "everything doesn't have to be perfect". Human nature dictates that people would be significantly more willing to risk injury/death if they're in control than if they're innocent passengers and a computer were to make an occasional fatal mistake. In other words, while a 50% reduction to the 35,000 U.S. driving fatalities would seem miraculous, absolutely NO ONE would tolerate 17,000 innocent deaths under computer control. How many self-driving cars driving off California cliffs do you think would be tolerated before everyone on that road avoided self-driving cars? 10,000? 5,000? Or just 1? How many deaths before everyone avoided self-driving cars all together?

Tesla's progress to date has been pretty sobering for me. And I think we should all, as ValueAnalyst appears to have done, discount any revenue from the Tesla Network until 2020 at the earliest, until Tesla proves otherwise. My .02.

So much dumb in this post. First, my point was that many competitors to Tesla have several lidar and radar as redundant systems, this brute Force hardware. To what end I ask. If Tesla's radar fails, highly unlikely and rare, It's no different then a flat tire. No need to have a complex system to auto replace the tire while the car is driving without interruption.

Second, imperfect human drivers don't drive off cliffs in large numbers or every really, even though they are busy texting, drunk, high and not focused on the road and neither will an autonomous car with decent high def 3d maps. Again, not perfect but better then humans because it has perfect focus on the road and everything around the car. No screaming kids or cellphones to distract.

And I disagree that people will not tolerate deaths if there are half as many as what they would have been. You are either not thinking straight or have not really thought about it critically. People put up with an insane amount of crap and would happily put up with less to the tune of 17000 lives per year saved. And what you conveniently leave off is what else comes along with those saved lives. One being the comfort of having your car drive you in traffic. Of not needing to own a car or being 90 or blind and having the freedom to go where you want.

No one knows where Tesla is in this process, they don't have cameras over their shoulder showing us on real time what they are doing. They have only promised one thing, coast to coast drive by the end of the year. Why the hell does any one expect monthly updates on where they are in the process? Don't be a child. Tesla is not going to show their hand until it's forced and no one else is forcing anything. It's the same reason there won't be a P130D any time soon, because no one really has an EV that had over 250 miles of range and faster then 5 seconds 0-60.

 

[kenliles]

You disagree that self driving is harder than rocket science? I would agree that safe Level-5 self driving is at least an order of magnitude harder than a rocket that can land on a barge.

Yes I do.
Level 5 self driving is an application of the science, much of which originated with rocket science.
The corollary application to rocket science is human exploration travel and settlement of space.
Level 5 self driving is a narrow application of the underlying science, and a walk in the proverbial park in comparison

 

[neroden]

And I disagree that people will not tolerate deaths if there are half as many as what they would have been.

They won't. You're simply wrong and Alketi is right about this. The psychology on this is famous, though admittedly rather disturbing. Have you ever wondered why we don't have fully automated trains everywhere (a technology perfected in the 1970s)? Have you ever wondered why the number of deaths caused by automobile drivers is tolerated at *all*?

Musk actually agrees. He's said that self-driving cars won't be accepted if they're merely twice as safe as human drivers. I believe he said 100 times as safe as human drivers was the threshold (though I may be remembering it wrong) -- this seems about right to me based on all the psych study abstracts I've read about this.

They'll get there, mostly because the average human driver is terrible, but it'll take a while.

 

[Reciprocity]

Yes I do.
Level 5 self driving is an application of the science, much of which originated with rocket science.
The corollary application to rocket science is human exploration travel and settlement of space.
Level 5 self driving is a narrow application of the underlying science, and a walk in the proverbial park in comparison

My alluding to rocket science was the fact that in the case of launching things into space, it has to be perfect because you can't pull the rocket over and be ok. You can't just stop the rocket on the middle of street and a cop car can't just push you off the road. You have one shot or you waste everything in one giant explosion. Again, the radar fails, the car can pull to the side if the road and shut down using the other sensors. You don't need a backup radar and lidar. The car doesn't explode.

 

[neroden]

Yes I do.
Level 5 self driving is an application of the science, much of which originated with rocket science.
The corollary application to rocket science is human exploration travel and settlement of space.
Level 5 self driving is a narrow application of the underlying science, and a walk in the proverbial park in comparison

On the contrary, self driving is a much much harder problem. Landing a rocket is a triviality by comparison. Do you actually know enough physics to know why? The constraints on an automobile are much stricter, and the problem specification is much harder, because the air is a hell of a lot emptier than the land around a city street.

Now, if you could land a rocket in a hurricane (which even SpaceX can't do) then I would perhaps consider that more difficult than self-driving, though it's still a different problem. The only hard part of the landing is dealing with air turbulence, really, and there's an awful lot of room to sense and adjust to it in relatively calm air conditions. There are a lot more hard parts to self-driving than that.

 

[Reciprocity]

They won't. You're simply wrong and Alketi is right about this. The psychology on this is famous, though admittedly rather disturbing. Have you ever wondered why we don't have fully automated trains everywhere (a technology perfected in the 1970s)? Have you ever wondered why the number of deaths caused by automobile drivers is tolerated at *all*?

Musk actually agrees. He's said that self-driving cars won't be accepted if they're merely twice as safe as human drivers. I believe he said 100 times as safe as human drivers was the threshold (though I may be remembering it wrong) -- this seems about right to me based on all the psych study abstracts I've read about this.

They'll get there, mostly because the average human driver is terrible, but it'll take a while.

Because you have one train operator for a 3 mile long train and there aren't 34000 people killed every year in train accidents because they are morons.

You are wrong and people already drive in Tesla's like they are autonomous even though it's dumb. Don't recall anyone saying 100x was "required" and either way, I disagree. The number I have heard is 10x though I expect it's closer to 2x.

 

[neroden]

My alluding to rocket science was the fact that in the case of launching things into space, it has to be perfect because you can't pull the rocket over and be ok.

Ever been on a country road in the winter? You can't always pull the car over and be OK either.

Again, the radar fails, the car can pull to the side if the road and shut down using the other sensors.

This is called "level 4 self driving", not "level 5 self driving". And yes, level 4 is much MUCH easier than level 5. I anticipate level 4 within a decade, and full level 5... well... never.