Hacker News Reader

Comment by eli_gottlieb | Hacker News Reader

>By 2018, the Gigafactory, which is less than a third complete, will double the world’s production capacity for lithium-ion batteries and employ 6,500 full-time Reno-based workers, according to a new hiring forecast from Tesla.

Unless those workers will be putting in thousands of hours per week (impossible, I know), this implies some truly prodigious productivity statistics: that by using the best available manufacturing techniques, they can double the world's battery production on 260,000 hours/week of labor.

View on HN

> Unless those workers will be putting in thousands of hours per week (impossible, I know), this implies some truly prodigious productivity statistics: that by using the best available manufacturing techniques, they can double the world's battery production on 260,000 hours/week of labor.

Tesla acquired a renowned German manufacturing automation firm [1]. I wouldn't be surprised if the Gigafactory is heavily automated as well.

[1] https://electrek.co/2016/11/08/tesla-acquires-german-enginee...

Part of the manufacturing has been built by Panasonic, in partnership with Tesla, based on their existing expertise on how to produce batteries at scale.

Yes, Tesla is leaning very heavily on Panosonic as a manufacturing partner at both the Gigafactory and the Solar City plant in Buffalo, New York.

I've heard somewhere (possibly in a BBC TV report) that batteries are passed through a hole in the floor from Tesla to Panasonic enclaves because neither company wants the other to know fully about the processes required to finish the final product. Hence the joint venture on one site.

edit: ok, the article[0] doesn't mention secrecy as the motive but yes a hole is involved. And the motive for a single site is to reduce transportation costs.

[0]http://www.bbc.co.uk/news/technology-36893104

Very impressive to see all those jokes about "fully automated luxury blah blah blah" starting to come true!

Well, the "fully automated" part increasingly coming true; we'll see how much the luxury spreads.

"“Over time, the majority of our engineering effort is going into engineering the factory as a product itself,” said Tesla’s CEO Elon Musk …“It’s the machine that builds the machine. If we take creative engineers and apply them to design the factory, they make five to ten times more headway than improving the product itself.”"

https://qz.com/745278/teslas-entire-future-depends-on-the-gi...

-- 16.08.03

This. Why manufacturing needs to be done here in the USA. Those who control manufacturing control all the things. When you control the factory you figure out how to make the factory smarter. Tesla could have just said "Hey, we need batteries. Let's make them in China... they can make everything!" But instead they decided to be the masters of their own destiny.

Elon is very smart, indeed.

It is smart to be involved in the manufacturing process. There's a spectrum between building a factory and outsourcing completely: having an active say in how your products are made makes sense. Paying for and owning every step of the way might not. Especially if you don't have the capabilities and bandwidth of Elon Musk -

(can we admire running SpaceX & Tesla and being impressive at SolarCity, OpenAI concurrently, or is that fanboyism?)

It definitively put stuff into perspective.

6500 jobs is a big number in human scale. That's more people that you know.

But that's 0.05% of the manufacturing jobs in the US alone. 0.05% of the manufacturing power of the US fitted with the latest tech match the output of the whole Lithium-ion battery sector.

Obviously that's 1 specific field and product and you can't extrapolate to everything else, but really you can't help but think that even maintaining current employment in manufacturing will require staggering increase in manufacturing output of the US. If Trump is even a bit serious about trying (and has the congress support for it), it would require a colossal disruption of the market either through regulation, taxation, investment, ...

If I remember correctly, Tesla are supposedly unusually keen on automation. Possibly more so than the economics actually justify, even - or at least, there were some doubts amongst industry experts as to whether the ways in which they automated tasks actually made sense, given the cost and speed of the robots they were using compared to a human.

"or at least, there were some doubts amongst industry experts as to whether the ways in which they automated tasks actually made sense, given the cost and speed of the robots they were using compared to a human."

That makes sense, considering that Tesla's approach seems to be to get the tech set up first, and then make it economical with R&D breakthroughs later.

It's all a front set up to provide a plausible reason for building the infrastructure that our upcoming AI robotic overlords requested from Elon.

6500 people need to live somewhere. They need to eat. They need haircuts. They need medical services. And all those other people need the same services too.

I could keep going but you get the idea. Jobs create more jobs which create more jobs.

The point is that it's a fraction of the number of employees needed to accomplish the same tasks 50 years ago or even 10 years ago.

What's the point of making that statement? 'Cotton production requires a fraction of the number of employees needed to accomplish the same tasks 150 years ago.'

The bottom line is this: It's an advanced economy sector industry creating jobs in the US, where none existed before. Not in spite of, but because it's a highly automated factory, it can be built in the US. The more "great American" companies that build advanced factories, the better for the US economy.

Frankly, I'm most impressed that this project seems to be on schedule. Gives me a little encouragement that we can execute on mega projects like this.

The point is that the jobs that were lost in manufacturing will never come back. Even if everything is shifted back from China/Mexico to the US, very few of the jobs will come back. The old world as many Trump supporters hope for will not return and we have to think of a way to go forward.

The great thing would be if the increased output from this automation matched what USA currently produces in China. Then it could gain independence from China, and China could finally give those goods to their internal markets. And employment standards could be respected in both countries (however products built in USA will still be significantly more expensive, matching hopefully higher salaries).

The net result is both countries would be poorer, we didn't start international trade for no reason, there are strong economic and politcal benefits to trade.

Maybe the rich will get poorer. I suspect income inequality would decrease however. Median income would increase.

What mechanism do you think will decrease inequality?

Related; the factory is said to be 1.3M square meters when completed. So each employee will have 200 square meters for himself and his robot coworkers. Except of course that they are probably not at work at the same time. So maybe 600 square meters.

So if the workers are equally distributed they never see each other except at the entrance and during lunch. Nice.

Square meter numbers sound bigger than they are because they're, well, squared. 600 square meters is less than 25 meters square.

I refer to it occasionally since I toured it once upon a time but General Motors built a new assembly factory near Lansing, Michigan that was the most advanced facility in the world at the time.

It's 320,000 sq. meters and employs about 3,800 people across three shifts, so call it 1,300/shift to make the math easy. That's about 1 employee per 250 sq. meters. Except that the machines and robots that build the modern world are enormous. They take up a huge percentage of the area of the factory.

The work that the humans actually performed was typically in small teams of 4-5 people as they wound cabling through the frame, ensured everything was aligned properly, performed a few welds that their 5-axis welders couldn't efficiently reach, etc. It was fascinating to see where the robotic investment wasn't worth it at the time.

I find the fact that some of these jobs are still not effective to be done by robots fascinating. At what point will we automate those jobs as well? Is it just the cost/benefit ratio is wrong currently? I cannot believe that it's strictly impossible to make a robot that can do all those task.

At the time, most of the human workers were doing things that required significant dexterity beyond the ability of the robots.

Taking a wiring harness;

https://i.imgur.com/H3XVf92.png

Routing it through the firewall and then making a dozen connections is simple for a worker, but extremely complex for a robot.

That factory was turning out 1 car/minute, so the $30/hr guy plugging harnesses in would add maybe $0.50 to the cost of the car? When you're selling vehicles with MSRPs over $40k (this was a large SUV assembly facility), the threshold for automating something is higher than you might expect.

SpaceX (another of Elon Musk's companies) has achieved similar feats in the realm of space flight.

According to his biography <https://www.amazon.com/Elon-Musk-SpaceX-Fantastic-Future/dp/..., Musk is particularly good at identifying supply-chain components that can be economically produced in-house. The Gigafactory looks like the same general strategy, and if it pays off as well as the SpaceX case has, they have a good chance of achieving their production goals.

What happens when he starts combining companies?

Space X launches the rocket that takes solar-powered self-driving machinery to Mars which starts building solar-powered settlements. By the time we get there, we're not fighting for basic survival, we're just moving in...

Most supply-chain components can be economically produced in-house. You just need the capital to build a 5B factory, which is a whole different kind of problem in my opinion.

Anyone know how many full-time workers are employed in battery factories currently? With that figure we could estimate the Gigafactory productivity multiplier for labor.

>that by using the best available manufacturing techniques, they can double the world's battery production on 260,000 hours/week of labor.

Mostly true, but to get a complete figure you might also factor in labor hours from upstream vendors which supply parts and/or raw materials. Increased battery production also means increased raw material extraction/processing/transportation.

Still impressive though. There's probably less than 50k workers involved in producing the world's battery supply.

I think a lot is only possible when you aren't tying you own hands with re-using/preserving already built capital equipment.

https://en.wikipedia.org/wiki/The_Innovator's_Dilemma

I'm not sure I'd class the Gigafactory as a disruption in the "Innovators dilemma" sense. I think some of the genius of Musk is that his bets actually minimize the big unknowns in a physical realm while at this point in the business dimension, the demand for Tesla's new model is known to sufficient certainty that this factory is needed to fill the capacity.

The core battery tech isn't new, but the machinery and the automation of the process is.. but that's a much smaller risk than say trying to build this factory on a "disruptive" new battery tech.

Also, Tesla is building out an energy storage business that can (according to Tesla's conference calls) eat a practically unlimited amount of battery packs with decent margins. So they also have some insurance there in case car demand should be lower than anticipated, or car production should hit delays.

What's your source on that?

I don't understand why it takes so much labor. I would have thought automation to be massive.

Well, let's put it another way. At 50 GWh, I believe the Gigafactory will be producing at least 2.5 billion individual batteries annually.

If you divide that by 6,500 employees supporting the manufacturing, it comes to about 385,000 batteries produced per employee per year.

Ok, but assembly lines, robots, automation. They don't have to necessarily employ the truck drivers/transport/distributors.

So what is the biggest labor item?

Reply on news.ycombinator.com