Musk Takes Responsibility For Excessive Automation Of Model 3 Production


Admits human are necessary and underrated.

Initially, the Tesla Model 3 was supposed to be produced in a manner never before seen in the automotive industry.

Related – Tesla Confirms Model 3 Production Exceeded 2,000 Units In Single Week

Tesla CEO Elon Musk wanted production to be so highly automated that it was basically hands-off to humans, minus maybe some final inspections and quality control.

To achieve this level of automation, new machines were envisioned, design, developed and implemented.

However, after awhile it became clear that certain parts of automotive production couldn’t be accomplished successfully by even the most advanced of robots.

Machines were ripped out and replaced by humans. That’s how Tesla was able to finally up Model 3 production after months of constant struggles.

Now, Musk is willing to admit that it was all his fault. In a reply Tweet, Musk stated:

“Yes, excessive automation at Tesla was a mistake. To be precise, my mistake. Humans are underrated.”

A win for us humans.

Here’s the Tweet in full:

Categories: Tesla

Tags: , ,

Leave a Reply

65 Comments on "Musk Takes Responsibility For Excessive Automation Of Model 3 Production"

newest oldest most voted

I’m glad that it sounds like reality is setting in over there, but I’m afraid less automation = the end of the dream of the $35k Tesla. Humans are underrated and expensive.

I was down for a first production until I dropped out, but I know many who were waiting for the SR sans PUP to make it happen.


Musk initially stated an 18% profit margin on the Model 3, then he later on raised it to 25% profit margins.

The automation does not mean the end of the 35k Model 3 by any means, it simply means lower profit margins.

Well at least initially. Just because it is difficult with robots now doesn’t mean it will be in a year or 2.


Thank you! I find your analysis to be insightful.

Yes, I was surprised when Tesla and/or Elon announced that they were aiming at a 25% profit margin for the Model 3. That’s about what it is for the Model S. With the TM3 being a lower-cost car made in greater volume, common sense suggests Tesla should be aiming at a lower profit margin, trading higher volume for lower margin, for overall higher volume of net profits.

Hopefully Tesla will settle for a more realistic profit margin on the TM3, something like the 15% they originally talked about as the goal (before the much more recent goal of 25% was announced), rather than raise the price of the car. However, if a full-blown trade war with China gets started, then Tesla might be forced to raise its prices even if it’s willing to settle for a lower profit margin.

Martin Winlow

Were I him I would not give up entirely on plans to make the line more automated. Now that he has learned a hard lesson (more hubris?) perhaps a more pragmatic approach might be to address the processes he wants to automate one at a time, diverging the flow of part-assembled vehicles in one place to develop automated processes in slow-time in parallel with a fully-functional non-automated process. That way he can guarantee vehicle manufacture times (ish) whilst developing the automation – and no-one has to sleep on the factory floor!

Tom Dually

Keep in mind that GM/Toyota built twice as many cars, with half as many people, as Tesla is achieving at NUMMI.

Their true margins will slowly approach mid-single-digits. The math simply doesn’t work if you properly allocate costs any other way.

They simply have too many people, building too few cars– in an inefficient way– with too many external costs that dealers normally shoulder. It’s a recipe for sub-standard operating margins.


Except you are missing a few major things:

1) The workers at Tesla are not all working on the assembly line. Since Tesla doesn’t have as many locations, they have a lot of workers in the factory doing other stuff like research and the like which wouldn’t have been the case in the NUMMI plant.

As for overhead, the dealer network doesn’t eat up cost. They add cost due to their overhead.


Bad Musk, Bad, BAD MUSK!


Humans are quite remarkable machines.


This! Throw enough money (aka, time) at the problem, and entire process can be automated. For example, each stage has QA robot that defers to next robot for particular problem and so on and post QA QA robot, etc. Infinite state machine will work perfectly.


“the Tesla Model 3 was supposed to be produced in a manner never before seen in the automotive industry.”

I think it was supposed to meet the best of the best of the industry. The never before seen was the alien dreadnaught 2.0 for the Model Y. But now considering they backed off redevelopment of a Y specific platform, that’s probably not happening.

There is a reason legacy automakers have multiple factories and lines to crank out the highest volume vehicles in multiple regions.

All-Purpose Guru

There never was a Y specific platform. Musk admitted later that he was out of the loop on the decision to make Y a derivative of the 3.

He also said that it made sense to build the product that way.

David Murray

Well, the issue is, which he admitted to, is that they tried to do too much new technology too soon. Using humans now does not mean they won’t be able to automate some of those processes later.

Joshua Burstyn

This is probably the fair and balanced view.

Magnus H

Yes, like every other car manufacturer.


“Who knew that car-making could be so hard?”


Sounds to me like the technology just isn’t ready yet. Fundamentally, a robot should be able to do anything a human can on an assembly line. Its just a matter of figuring it out, which is easier said than done of course.


Kind of like self driving cars on “autopilot”. Tougher than people think.


Yeah, good luck driving in the snow, robots! What’s that, no idea how to drive when you lose traction? Can’t spot black ice? Don’t know how to follow tail lights when the road is completely gone? Can’t get out and wipe off all those shiny sensors when they are caked in ice? Don’t know how to dodge a pothole?


To be fair, most humans suck at those skills as well.


To be fair there is no such thing as AI only bots that are really good guessers


As someone who programs robots, there’s a lot they can do, but there’s a lot they cannot. Whenever I have to explain to someone why their robot idea won’t work, they usually have some vision in their head that robots are just like super strong humans , but with all kinds of AI and know exactly what’s going on. In reality, the robots are dumb and they just repeat the same process over and over. They don’t deal well with change or inconsistencies, but they don’t complain about carpel tunnel or call in sick.

Is the intelligence getting better? Of course. And the 3D vision is also getting better (but vision is a whole different Pandora’s box of problems). There are jobs on the auto line that are very difficult for HUMANS to do, so expecting a robot, that doesn’t have a brain or nerve cells or infinite joints/positions, to be able to adjust to complex situations on the fly just isn’t feasible. At least not with today’s technology, as you said. Maybe when we get to Westworld level, but I have a feeling building cars won’t be our concerns then.


Thank you for that overview. Sadly, many movies and TV shows which portray robots as — just as you said — just like humans only stronger, faster, etc. have given a lot of people a basic misunderstanding of the real state of robotics.

A robot assembly arm on a production line can work tirelessly, precisely, for hours and hours without taking a break, without slowing or tiring, and at least in theory, without ever making a mistake, so long as the parts it is using are formed properly.

What a robot is utterly and completely incapable of, is spotting when something has gone seriously wrong. A human could see that at a glance, and could call for the line to be shut down. But the robot will just keep mindlessly performing the task it’s programmed for, no matter what is or isn’t happening around it.

Reminds me of the Mickey Mouse / “Sorcerer’s Apprentice” sequence from “Fantasia”.


As you said, robots are dumb and don’t handle inconsistencies and changes well. So, a lot more effort has to go into nailing down the design, and ensuring all components and upstream processes meet tight specifications at the planned production volume before the start of production.

Tesla took short cuts in those vitally important areas. The design was still being finalized after the start of production. Suppliers weren’t properly vetted. The vehicle prototypes, components and production line weren’t thoroughly tested beforehand.


Musk challenges his troops at every turn, testing their abilities to innovate and to make the impossible, possible:
A transmissionless Roadster,
A relanding first stage rocket,
Falcon doors,
Solar roof tiles,
A Hyperloop Tunnel in L.A.,
and now…A fully-automated production line; Which I think they will develop in time.
That’s entirely different than current companies, driven by the greedy Wall Street crooks, who alway take the easy path and can’t break away from their hundred year fossil fuel past of producing expensive, pollution machines.

Magnus H

Great sarcasm, thanks for the laugh! 🙂


If he’s taking responsibility, how about cranking out Tesla 3 fast enough so that even I get to enjoy the tax credit? Or how about giving $7.5K rebate in form of Tesla stock taken from his holding after tax credit expires?

Carl XVI Gustaf

I asked Elon, and he said No!


If he’s not giving up anything, what’s the point of admitting responsibility? Heck, anyone can claim responsibility for everything if there’s no repercussion.


Giving up old Sparky!?


What this doesn’t answer is how much of the additional new automation DID work that they didn’t tear out.

If they increased automation by 40% over the Model S/X line, and had to tear out half the newly automated steps, that would still mean 20% more automation. (percents for illustrative purposes only — not actual percents)

And the follow up question would be how much work will it take to actually go through those systems one by one and slowly add them back into the assembly line.

The automation of the assembly line was always planned to be an evolutionary process, where new automation was planned to continually be added.

It sounds like they tried too much too fast, but I’m not hearing that none of it worked, or that they have permanently given up.

just frank

I still think its the cray cray pack design…can robots pack serpentine hoses as well as human fingers with tactile feedback?

They can weld the body.

What exactly can’t be automated….hmmm

Sam L

I must not be the only one who is happy Musk is eating his words. Not in the way a hater wished Tesla would collapse, but how the mighty Elon finally backtracked a little and comes back just a tiny bit closer to reality.

Now imagine if Tesla admits defeat and adopt the Toyota Production System which is the gold standard of the industry.


I don’t think Tesla is going to adopt anyone else’s system. But from what we’ve seen so far, they are likely to eventually figure out one that works for them. Will there be more setbacks, false starts, and missed expectations? Assuredly. But the last few years have taught us to never count them out, and I believe they will figure out a working system and eventually get to the mass market production numbers they’re looking for.


If Toyota is so great, then they would have come out with the Model S and Model 3 and been the clear market leader, not Tesla.

I’m just the opposite in my hopes as you. Production methodology be damned, I’m hoping Toyota will grow up and dump their silly fuel cell nonsense, and follow Tesla’s lead in building long range fast charging EV’s.

Heck, it has only been nearly a decade since Tesla first brought the Model S to their first car show where every car maker could see it for themselves. And yet still Toyota is still dithering with a slow short range PHEV that only a Prius owner could love, and compliance only full electric.


Despite even mentioning ‘production methodology’ in your post, you still managed to conflate it with vehicle types? You’re responding to something the poster never said, although I will admit to agreeing with both of you. The Toyota Production System is the gold standard, and yes, Fuel Cells are uninteresting and Toyota needs to up its EV game.


I intentionally conflated the two, because process at Toyota has failed to lead to results.


Nix said:

“I’m hoping Toyota will grow up and dump their silly fuel cell nonsense, and follow Tesla’s lead in building long range fast charging EV’s.”

Hear, hear!


That is like saying Henry Ford should have dumped the assembly line and adopted the gold standard used by others to make horse carriages.

Electric cars are different than ICE cars. Which means production method needs to change to adapt to the new technology. In both the product and manufacturing.


Stupid analogy. An EV is just a vehicle with a diiferent powertrain. Most of the vehicle is the same as an ICE vehicle.

Volvo plug-in hybrids, for example, have the same components as an EV in addition the ICE bits. They are built on the same production line as their ICE counterparts.

Martin Winlow

What I find most remarkable and interesting about this story is that we got to hear about it. If this was GM or Ford etc it would never have come out. Now maybe this is just Mr Musk working his public relations thing as masterfully as ever (again)… or maybe he really *is* an honest, up-front kind of guy. My money is (literally) on the latter.


GM, Ford and the rest of the large guys have pre production facilities to test the line prior to being install at the plant. That’s how they can shutdown a plant for a month in the summer to do model changes – the line was already tested. They might make adjustments, but they know it performs it’s intended purpose. So you wouldn’t expect a story like this after the car was in production from a major auto maker.


The thing is that THEY have problems developing their assembly lines too, they just don’t talk about it publicly.

The step of moving it onto the factory floor is meaningless, because Tesla was building a new line and simply didn’t need that step. They were able to develop their line in place.


Stop making up stuff!

On the last earnings call, Musk said a new automated battery production line had been built and fully tested in Germany, and would be disassembled, shipped and installed at the Nevada facility.

That’s how it’s normally done. The first battery production line was also built offsite, but proper testing was not done prior to disassembly, shipping and installation. We all know how well that turned out.


Where’s the article of model Y for 2019 article here on insideevs? I can’t find it


Thank you i couldnt find it seem it was deleted

How wonderful! It’s great to know that good paying blue-collar jobs are not going to disappear from Tesla. It’s doubly good that this means there won’t be a wave of manufacturing companies moving to eliminate all their blue-collar workers in favor of robots. At least not yet, and not all at once. If all the blue-collar workers have their jobs eliminated, and most of them have to take low-paying service jobs — that’s what has been happening in the USA for decades now, and it’s a large part of why the middle class has been hollowed out — if all blue-collar workers are replaced by robots, then who is going to buy the cars that those factories produce? One thing is certain — it won’t be the robots! I personally think it’s also good that reality finally seems to be setting in with Elon, at least in this one area. His many comments about using basic physics techniques to plan a many-fold increase (5x or even 10x) in output of Tesla’s auto assembly lines, his claim that everything could move at (my words, not his) eye-blurring speed, a speed at which (his words, not mine) air resistance would actually be… Read more »

👏👏👏👏. You are now understanding. Thats why Ford gave out better salaries and a living wage for they can buy his car


Tesla has about 1 year technological and EV production advantage over the Germans. 2 Years ago it was more than 2 years. At this rate, they will be caught around 2020. If this happens, they will be priced like other luxury car makers, and they won’t be able to fund expansion from capital markets anymore. At that point, they will have to sell the company to a larger automaker.


Tesla has more like a 3 year technological advantage and production advantage.

That said, Tesla is not exactly sitting still on their technology and continues to improve it.

And as Tesla continues to scale up in size, the capital they work with from their profits.

By 2020, Tesla will already be as big as Subaru and Mazda in terms of car sales. By 2021/2022 Tesla will probably match BMW and Daimler.


You may be right on the tech and production advantage, but Tesla will still have another, stronger edge: no internal army in every department of legacy ICE-bound execs to slow down expansion of BEV production and support. That advantage is worth at least a couple more years’ head start in addition to the areas you’ve mentioned.

Even when it may no longer be true in a purely production-oriented sense, it will be true, at least for a few years, in customers’ heads as a brand. Buy an EV from nearly any other car maker, and you’ll still need to have a conversation with folks to establish just what you’ve bought. Buy a Tesla, and people will know exactly what you’ve bought (even if many won’t know the details, they’ll know that’s one hot car).

At the very least, Tesla is carving out a spot for itself as the Apple of transportation. Production capacity alone won’t tell the story. Are Teslabears factoring the value of that into their valuations of the company?


Tesla is more like Blackberry or Palm of transportation. Apple did not enter smartphone market until others have been in it for a few years.


I’d say that it’s Toyota which is the Blackberry of the EV revolution. Toyota entered the market early with the Prius, but never advanced to a plug-in EV until after Tesla, Nissan and Chevrolet had started selling them.

Just like Blackberry with its smart phone failing to react strongly when Apple entered that market with the iPhone, Toyota seems stuck in a rut with its first-generation EV tech. The Prius Plug-in/Prime is a very weak response to EVs like the Tesla Model S, the Chevy Bolt EV, and the Tesla Model 3. Toyota’s attempts to promote its fool cell Mirai, while whining that there is no market for BEVs, is even worse!


Tesla first brought their production intent Model S to car shows in 2009. Nearly a decade later still no Model S competitors on the market.


There were made many electric cars before Tesla. Heck, in the turn of the 20 century (1900, about 120 years ago), about 1/3 of the cars on US roads were electric. The reason that big manufacturers didn’t make a Tesla Model S competitor was not their inability, electric cars were simply not profitable. Now the cost of batteries have dropped so much that it is starting to make sense to make electric cars.

It has always been possible for legacy gasmobile makers to do what Tesla is doing. The reason they have not done so, and won’t until it’s too late for many of them, isn’t because they can’t, but because they don’t want to. Their entire business model and all their infrastructure is invested in making gasmobiles. Eastman Kodak could have taken the lead in the digital camera revolution. They actually invented the tech! But they didn’t develop it because it would have competed with their core profit makers. Ditto for the legacy auto makers, including the Germans. They don’t want to make EVs compelling enough to compete with their core profit makers; their best-selling gasmobiles. That’s not going to change until the average car buyer starts indicating a preference for EVs over gasmobiles. When that happens, the Germans — and everybody else — will be rushing to play catch-up with Tesla, and all the other upstart EV manufacturers which pop up between now and then. Quite probably by then, BYD and possibly other Chinese EV passenger car manufacturers will have started selling in the American market, too. P.S.– If you think that Tesla is only 1-2 years ahead of makers of… Read more »

Fake news.

Tesla is a tech company.

Our internal code name for the factory, the machine that builds the machine, is the alien dreadnought,” Musk said on the call. “[When] our factory looks like an alien dreadnought, then we know it’s probably right.”

Winning. Go Elon!

Chapter 11 is the goal.

Get Real

LMAO, admitted Tesla shorter and non-EV driving Trollftf once again shows up (at least under his Trollftf username, instead of one of his other fake usernames).

Meanwhile, Tesla now has the best selling Western EV on the market even in its early stage of production ramp.


Look at the fabois spin into the ground. Lol
Humans = expensive. So what this means is ADIOS $35k Model 3. If they weren’t gonna be profitable with robots completely building them, forget about it now.
One more promise ($35k 3) that will very soon be broken from Elon!

Get Real

And…Mental MadBro continues his sleezy trolling of all things Tesla since he is terrified he might lose his GM stealership job if Tesla’s business model on distribution prevails.

Magnus H

Subaru is too small to develop their own EV technology, maybe Tesla is too?


Maybe…just maybe…some of the other car companies know something. Nah….

Get Real

Yeah Tom, other car companies know how to make LICE.

Thanh Lim

As they say, premature optimization is the root of all evil…