Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

General Discussion: 2018 Investor Roundtable

This site may earn commission on affiliate links.
Status
Not open for further replies.
Considering Elon's recent statements about robotics I think he would agree with Sandy. Elon has backed away from the fully automated factory concept in a very big way. The super factory machine was supposed to revolutionize car building, now they are throwing more people at the problem. I see this as a major setback for Tesla.
The Model 3 production line was always going to be the alien dreadnought 0.5. That means people are needed.

My impression is that they're scaling back some of their ambitions for the Model 3 production, so instead of the alien dreadnought 0.5, we might end up with the alien dreadnought 0.4. I think we can live with that.

In my view, we're talking about a minor setback, and Tesla will continue to push the limits of automation.

It may even be a good thing that Musk gets a reality check before attempting the alien dreadnought 1.0. I agree with Sandy that getting the vehicle engineering right is essential to be able to fully automate the line. And since the Model Y is still being designed, there's still time to get things right the first time around.
 
Here’s how I interpret Sandy’s assessment.

1. He likely got an earlier M3, which had some inconsistent panel gaps. Which is a fair criticism if the model he received was an earlier built. We saw these same types of problems in early MS and MX, both cars eventually got passed their earlier built quality issues and Tesla was able to ramp with more consistency. Having own an earlier model S myself, I did not have any issues with panel gaps in my car. I also have several friends who own Tesla’s, no complain from them neither. To me this is an easy issue to iron out and Tesla might have already passed the initial stages of panel gap teething issues. I’m also seeing very satisfied M3 owners and very few panel gap complaints on the M3 forum, YouTube reviews, etc. as time progresses.

2. Tesla’s battery and overall technology is far and ahead everyone else, Sandy thinks they surpassed all competitors including BMW, LG chem, Samsung, etc. hands down no contest. “Ignore Tesla’s batteries and tech. at your own peril.” This is the most important component of the interview to me.

3. Here’s where I would ding Sandy for his assessment. He admits that the suspension is one of the best he’s seen in the industry “F1 prince” was the term he used to describe Tesla’s suspension design. However, although he criticized Tesla for not using standardized components to build the suspension, yet he managed to praise Tesla for having one of the best suspensions in the industry. This sounds to me like someone who hasn’t really understood the overall design/functionality of the M3? Maybe Tesla could build it lighter, who knows, what counts is that he’s super impressed by the suspension.

4. Robotics. He may or may not be correct about robotics being “too dumb” to complete certain tasks. This might be where software can give Tesla an advantage. Although he’s competent, I think he’s the type of guy to throw judgement before learning from others. He reminds me of Bob Lutz, the guy who said “the Falcon wing is unmanufacturable.” Or “They can’t do it like that...” etc. and here’s the kicker, he hasn’t seen the M3 line yet and is guessing along the way.

5. Weight: he criticizes the weight of the M3 body, but neglects the safety rating of what this potential weight can yield in a safety crash test. I would wait until the M3 goes through safety testing before passing judgement on weight. He also admitted he doesn’t know “about crash worthiness” of the M3 build. “It’s different than anything I’ve ever seen, there’s panels and I don’t know why they’re there...” was his exact quote. To be fair, this might be a learning curve for Tesla as well. But I’m betting there’s a reason behind those panels. He thinks “maybe they hired someone who thinks they need to run the car over with a tank.” Well maybe, Elon designed it to be THE safest car, ill wait for NHTSA on this one.

6. He thinks the M3 is a “magnificent sexy” car. He seems to be very impressed with the tech, and think the Detroit boys are in trouble if they ignore the technology. “This is not the let’s get the boys to roll up their sleeves” [to beat Tesla], “this is not that type of rolling up your sleeves, we saw components we’ve never seen before” in the processor. It’s going to take Detroit a lot of time, effort and money to catch up.

In short, I think the panel gap issue is an easy fix. Robotics might not be as easy, but we’re witnessimg the M3 ramp happening now, so Tesla’s doing something right and Sandy should really visit Nummi before passing judgment. He did acknowledge that he can’t figure out how Tesla put their batteries together to perfection. So there’s clearly robotic components/software he hasn’t seen yet. This is likely going to be Tesla’s moat.


I thought the curb weight and size was similar to a BMW 3 series? It is also interesting that the Germans had taken apart a model 3 and we're impressed with the engineering.

You also must assume the car is built to support a P75D variant with sub 3 second 0-60 times and dual motors. Not that that is a huge stress but that the qualify should be comparable to a BMW M3.
 
He certainly seems credible with a lot of experience in the field, and he specifically said that he saw no increase in strength from layering of those materials over a simpler use of the appropriate material.
I'd bet from a conventional construction viewpoint his viewpoint has merit.

However, I suspect there may be some aspects of construction they simply aren't factoring in the equation. For instance, in the model S, the box-section frame rails also act as ducts to channel any flames from the battery-pack pressure ports forward away from the passenger compartment in the case of a pack fire.

In the original Model S B-pillars, they used boron steel inserts with aircraft-grade bolts for specific strength targets well beyond what was required.

Now, I know he was talking about a different section of the car construction, but those are just examples of where the construction may be accommodating additional design goals Munroe is not used to.

Of course, I also don't rule out that Tesla also just did some things for the sake of expediency in on order meet the accelerated time schedule. But FEA tools have been around for a long time... and cost/weight optimization has been at the top of Tesla's list for the Model 3 pretty early on.
 
Considering Elon's recent statements about robotics I think he would agree with Sandy. Elon has backed away from the fully automated factory concept in a very big way. The super factory machine was supposed to revolutionize car building, now they are throwing more people at the problem. I see this as a major setback for Tesla.
I wonder if that's permanent though?

I get the impression that it, like the X, was the introduction of too much new tech all at once... not necessarily that the tech itself is a bad idea given sufficient time to implement it.
 
Maybe slow down deploying extra SC's in California. We still have holes to fill like I-10 / I-20 east of the I-10/I-20 split. Pecos has been in permit stage for 166 days and Fort Stockton for 505 days! There's a few other holes that need filling too.

I hope (and suspect) that scaling back Supercharger construction significantly won't happen. It's an enabling infrastructure component that really needs to be well filled in as we start the process of getting half a million Model 3's on the road in the next couple of years.

The estimate is something like $250K per site. That means the 80-ish sites currently in the permitting/construction phase represents something like $20mil. That's really peanuts in their overall CapEx.

Not to say that indeed they don't want to control costs... however Tesla doesn't advertise... but I'd say that the Supercharger network grabs both mindshare and provides value and competitive advantage simultaneously. What's more, I think if the Supercharger network can't keep up with the rapid pace of Model 3 adoption it would have a far more negative impact.
 
Last edited:
I wonder if that's permanent though?

I get the impression that it, like the X, was the introduction of too much new tech all at once... not necessarily that the tech itself is a bad idea given sufficient time to implement it.

That's part of my frustration, they supposedly learned from their mistakes with the X, yet here they are again saying they put too much new technology into the 3 at once and apparently too much automation into the factory too soon.
 
My impression is that they're scaling back some of their ambitions for the Model 3 production, so instead of the alien dreadnought 0.5, we might end up with the alien dreadnought 0.4.
My impression is they've wasted time trying to do too much too soon. The description of the crazy conveyor system that didn't work and had to be removed for instance, and now they are pulling robots off the line and replacing them with more people. A step backwards to say the least.
 
That's part of my frustration, they supposedly learned from their mistakes with the X, yet here they are again saying they put too much new technology into the 3 at once and apparently too much automation into the factory too soon.
Tesla could have tried out their 21-70s in the Model S/X before using it in the Model 3, but other than that, I don't see what they could have done much differently. The cars are pretty much bare-bones.

Maybe they could have held off on the automation, doing it in a number of smaller steps, but my impression is that it wouldn't have made much difference. The battery production has been the bottle neck, and I don't see anyone saying it was too automated. It just didn't work quite like planned.
 
  • Like
Reactions: landis
That's part of my frustration, they supposedly learned from their mistakes with the X, yet here they are again saying they put too much new technology into the 3 at once and apparently too much automation into the factory too soon.
I tend to suspect that Elon is a CSMA/CD kind of guy... that is: when his ambition collides with reality... he backs off some bit, and then tries again, and then when he collides again, backs of a bit more, until he finally gets what works.

I also wonder if it might take few bumps to learn that such lessons cross domain boundaries... too much complexity at once in your automotive product may also mean it's not a good idea in your alien dreadnaught product.
 
Tesla could have tried out their 21-70s in the Model S/X before using it in the Model 3, but other than that, I don't see what they could have done much differently. The cars are pretty much bare-bones.

Elon was the one who said they put too much new tech in the Model 3 at once.

Maybe they could have held off on the automation, doing it in a number of smaller steps, but my impression is that it wouldn't have made much difference. The battery production has been the bottle neck, and I don't see anyone saying it was too automated. It just didn't work quite like planned.

I don't think the battery production was the only bottle neck. They are talking about using more people in Fremont instead of robots.
 
Here's some propaganda from my alma mater:

Alumna—and Google—Make Machine Learning Easy

As an "extinguished nerd" I often wonder if Elon and Hawking are right to be concerned about the spread of AI. In the words of The Shadow, my favorite radio mystery in the 40s of the last century, "who knows what evil lurks in the minds of men." All right, I know she's a woman.
 
The whole concept of "pencils down" in July 2016 was to get to volume production as quickly as possible -- the perfect is the enemy of the very, very good.

Completing the engineering design process to get to production by definition means that further design improvements could have been possible but would have come at the price of delaying production.

Focusing on results and the overall package, the Model 3 LR is almost 1000 pounds lighter than the Jaguar iPace (3814 v. 4784), has much more range, similar interior space and costs $20,000 less in the U.S. The Model 3 SR is comparable in weight to the Bolt (3549 v. 3569), but far superior in almost every respect. Jaguar I-Pace vs. Tesla Model 3 and other EVs: How they compare on paper

The Model 3 has no meaningful competition among EVs when looking at the overall package, and Munro's focus on weight seems like a red herring.

In any case, if there are ways to reduce weight without hurting performance, reliability or passenger comfort I don't see how that helps the incumbents -- it is an opportunity for Tesla to extend its advantage even further as it inevitably tweaks and modifies the Model 3, as it has done with the Model S (thousands of changes over the years Model S Changes from 2012 to 2017 | TeslaTap). Another Tesla advantage over the incumbents: constant innovation and improvement without being shackled by the model year concept.
 
Last edited:
Considering Elon's recent statements about robotics I think he would agree with Sandy. Elon has backed away from the fully automated factory concept in a very big way. The super factory machine was supposed to revolutionize car building, now they are throwing more people at the problem. I see this as a major setback for Tesla.
If you take a snapshot in time, he and Elon probably has the same conclusion regarding some robotics used on the M3 line up to now. But the difference is this guy would just use that as justification for never trying to improve using robotics ever again, while Elon will try again using the lessons he learned. Munro seems very much stuck in the past. That's my main issue with him. I have no reason to doubt his expertise, but to borrow Elon's words from the King interview, he is just another one of those guys who can only look in the rear view mirror.
 
Last edited:
I don't think the battery production was the only bottle neck. They are talking about using more people in Fremont instead of robots.
They need more people going from 4,000 3/S/X per week to 7,000 3/S/X per week. That isn't surprising.

Now maybe they are adding more people than they were planning on adding at this stage of the ramp, but I doubt we are talking about a huge difference.
 
I'd bet from a conventional construction viewpoint his viewpoint has merit.

However, I suspect there may be some aspects of construction they simply aren't factoring in the equation. For instance, in the model S, the box-section frame rails also act as ducts to channel any flames from the battery-pack pressure ports forward away from the passenger compartment in the case of a pack fire.

In the original Model S B-pillars, they used boron steel inserts with aircraft-grade bolts for specific strength targets well beyond what was required.

Now, I know he was talking about a different section of the car construction, but those are just examples of where the construction may be accommodating additional design goals Munroe is not used to.

Of course, I also don't rule out that Tesla also just did some things for the sake of expediency in on order meet the accelerated time schedule. But FEA tools have been around for a long time... and cost/weight optimization has been at the top of Tesla's list for the Model 3 pretty early on.
Not to mention Tesla probably has access to some of the best FEA tools and engineers at SpaceX who model this kind of mechanical strength/weight trade-offs on rockets. If anyone knows how to save weight it's gotta be the SpaceX guys.
 
Tesla could have tried out their 21-70s in the Model S/X before using it in the Model 3, but other than that, I don't see what they could have done much differently. The cars are pretty much bare-bones.

Maybe they could have held off on the automation, doing it in a number of smaller steps, but my impression is that it wouldn't have made much difference. The battery production has been the bottle neck, and I don't see anyone saying it was too automated. It just didn't work quite like planned.

From what I understand in both a manufacturing sense and a financial obligation to Panasonic sense, using 2170s in S/X isn’t currently feasible. Being able to migrate those models to the 2170 is a significant concern to me.
 
That's part of my frustration, they supposedly learned from their mistakes with the X, yet here they are again saying they put too much new technology into the 3 at once and apparently too much automation into the factory too soon.
I think the challenge has been finding an operations engineer who can implement his vision at Tesla. Gwynne Shotwell has been that person more at SpaceX, but no single person has been filled that role at Tesla. I hope he makes Tesla a higher priority through 2019 to get the semi, Roadster and Y in production and to drive productivity in the SX3 programs. They need productivity gains in the SX programs to free up people and space for the Y program. Elon refered to being more 3D and dense in future manufacturing. He has been able to very rapidly turn things around on the X and 3 lines after trying to let others lead. I think he is hoping to trust someone to step up, but I think he is going to have to build up the talent and leadership internally. That takes time. BFR and Boring are important too, so not sure if he can be full time for long. Hopefully part time Tesla is about 40 hours a week with burst mode at key times.
 
Up and to the right...

Via Scientific American:
T_comp_61-90.pdf.jpg


Even more.

Honestly, anyone against this company can go screw themselves.
 
Last edited by a moderator:
Status
Not open for further replies.