Manufacturing is getting more and more specialized and lean. Which is good because the barrier to entry is lower which means the innovation will only go up.
I thought a lot of base chemical production had moved to China and India? Is that not a barrier to entry for the US? We had them post WWII due to wartime production but it has slipped since. Sorry I am very interested but rather uninformed in this area.
Yes it has. More expensive in the US because the environmental standards are much higher here... but that's exactly why there should be more of this type of production here.
"Traditionally, manufacturers sell through retailers, group buying organizations and other middlemen, but times have changed." - this statement could not have been more wrong. There is a massive amount of manufacturers who sell directly to other manufacturers/companies.
That's my point. Most manufactures are stuck in these giant slow enterprise-sales processes.
When a hospital wants to buy from Armbrust, we send them to our shopify store, they can put in a PO and we ship same day. It's like bringing the internet to a world that's been using fax machines to communicate.
They are not "stuck" in anything. Manufacturing is a lot more complex than a customer placing an order online and you shipping the next day. There is production planning, storage considerations, purchasing of raw materials, and the list goes on. Manufacturers use information from their customers to increase or decrease production.
If I were to place an order for 1 million masks on your shopify store you would ship them the same day?
Given how many manufacturers are using enormous, ungodly spreadsheets (that someone who's no longer at the company created back in 2002) to manage their manufacturing flow, many are pretty clearly "stuck" in old, rather shitty ways of doing things.
I am beginning to manage physical production but come from a software background. While we don't use spreadsheets for production management (yet!), I want to explain why this is not insane.
Spreadsheets seem backward to software people because we are confident with manipulating data with more flexible tools. We see them as backward primarily because they are batch-oriented and cannot easily handle distributed writes, elegant types, reflective programming, serialization, automated updates, etc.
However, consider a small to medium scale manufacturer. They have millions of dollars invested in equipment. They have dedicated employees who know that equipment, its settings, its maintenance, and its operation. To those employees, software is a sideshow.
Chances are, for any given manufacturing order there is going to be the following: pricing the order, sending out a quote, taking a downpayment, allotting or acquiring the raw materials, preprocessing, temporary storage of parts, scheduling the machine and operator time against other orders for all unit operations necessary, post-processing, final assembly or QA, consolidating and packaging, actual shipping, settlement.
Can you see how batch-oriented processing actually makes reasonable sense for this sort of scenario? Going real time is quaint but if it doesn't buy you anything and you need to hire "$oftwar€-d€v$" to create and maintain interfaces that still require human input and have less familiarity to your very-busy-operators than a simple spreadsheet - why not use a spreadsheet? It's cheaper, faster, more familiar (read: less errors, ~no training required) and therefore from a business standpoint more effective.
Regarding small, batch-oriented manufacturing; I am sure you're right that spreadsheets are better suited for that domain. But there are plenty of moderately-sized, batch-oriented factories or continuous-flow manufacturers that are using spreadsheets as well, and I don't think the technique is nearly as well suited there.
To clarify, my problem isn't with spreadsheets as a tool in general: my problem is that spreadsheets are powerful enough to do this particular job even for larger factories, but not powerful enough to do it without quickly becoming 10,000 layers of accreted, fossilized spaghetti. If the process is illegible to everyone who manages the factory, using the spreadsheet is akin to divination by tea leaves: dump your numbers into the sheet and hope that you can interpret the right action from it.
Restated: it seems to me that spreadsheets are often taken well-past their limits as an effective tool in the management of factories, and nobody seems to notice that its happening until its too late.
Spreadsheets are the right answer for small manufacturers who don't have overqualified developers hanging around and willing to work for below market rate.
If at some point the manufacturer gets big enough to outgrow spreadsheets, there are plenty of well established upgrade paths.
I'd be very interested in what those upgrade paths are. Most of the large established manufacturers I have visited still operate on a de-facto job-shop mentality with either printed job sheets or whiteboards at each station or with each parts batch, occasionally dedicated screens or a mobile job-tracking solution. Experienced machinists I've interviewed recently have cursed the inefficacy of manufacturing management systems provided by major vendors. I read the literature and attend shows, right now the market to me looks like a dogs breakfast of largely vendor-linked but some independent solutions with all the lock-in concerns associated. I am therefore defaulting to NIH and build-from-scratch but keen to hear of others' experience!
Ha. I am aware it is complex. We can produce 1MM masks a day and it's wildly complex.
And yes, we could ship 1mm easily. We have shipped 10MM in one day. The issue would be booking the LTL (1mm is only about half a truck) which would take several days to arrange right now.
But they would be pallet wrapped and labeled that day.
Overseas manufacturing is based on arbitrage of lack of caring about human rights and ecological regulation, and a lack of a proper carbon tax on the transport.
Tariffs are a universal evil of economics, but if on-the-ground economic policy won't account for human rights and long term ecological preservation and the survival of the species, then I'll take a crude equivalence and a boost to domestic manufacturing.
They just need to be instituted gradually.
But worry not libertarians of HackerNews, that'll NEVER happen. Missing quarterly stock price targets so your options cash will always take precedence over the survival of the species.
> Overseas manufacturing is based on arbitrage of lack of caring about human rights and ecological regulation, and a lack of a proper carbon tax on the transport.
Yup. Many US made products are US assembled. The constituent components are still produced with slave labor and reckless abandon for the environment. There are some regulations around made in the usa vs assembled in the usa but they are minimal.
15 comments
[ 3.8 ms ] story [ 51.5 ms ] thread* More reliable (fewer moving parts and only one element needs calibration)
* Less expensive (just an LCD screen + Z axis)
* Much faster to print (any number of parts that fit on the build plate print get printed at the same time)
* More flexible materials (independent control over color and material properties)
Interesting times ahead..
When a hospital wants to buy from Armbrust, we send them to our shopify store, they can put in a PO and we ship same day. It's like bringing the internet to a world that's been using fax machines to communicate.
If I were to place an order for 1 million masks on your shopify store you would ship them the same day?
Spreadsheets seem backward to software people because we are confident with manipulating data with more flexible tools. We see them as backward primarily because they are batch-oriented and cannot easily handle distributed writes, elegant types, reflective programming, serialization, automated updates, etc.
However, consider a small to medium scale manufacturer. They have millions of dollars invested in equipment. They have dedicated employees who know that equipment, its settings, its maintenance, and its operation. To those employees, software is a sideshow.
Chances are, for any given manufacturing order there is going to be the following: pricing the order, sending out a quote, taking a downpayment, allotting or acquiring the raw materials, preprocessing, temporary storage of parts, scheduling the machine and operator time against other orders for all unit operations necessary, post-processing, final assembly or QA, consolidating and packaging, actual shipping, settlement.
Can you see how batch-oriented processing actually makes reasonable sense for this sort of scenario? Going real time is quaint but if it doesn't buy you anything and you need to hire "$oftwar€-d€v$" to create and maintain interfaces that still require human input and have less familiarity to your very-busy-operators than a simple spreadsheet - why not use a spreadsheet? It's cheaper, faster, more familiar (read: less errors, ~no training required) and therefore from a business standpoint more effective.
To clarify, my problem isn't with spreadsheets as a tool in general: my problem is that spreadsheets are powerful enough to do this particular job even for larger factories, but not powerful enough to do it without quickly becoming 10,000 layers of accreted, fossilized spaghetti. If the process is illegible to everyone who manages the factory, using the spreadsheet is akin to divination by tea leaves: dump your numbers into the sheet and hope that you can interpret the right action from it.
Restated: it seems to me that spreadsheets are often taken well-past their limits as an effective tool in the management of factories, and nobody seems to notice that its happening until its too late.
If at some point the manufacturer gets big enough to outgrow spreadsheets, there are plenty of well established upgrade paths.
And yes, we could ship 1mm easily. We have shipped 10MM in one day. The issue would be booking the LTL (1mm is only about half a truck) which would take several days to arrange right now.
But they would be pallet wrapped and labeled that day.
Tariffs are a universal evil of economics, but if on-the-ground economic policy won't account for human rights and long term ecological preservation and the survival of the species, then I'll take a crude equivalence and a boost to domestic manufacturing.
They just need to be instituted gradually.
But worry not libertarians of HackerNews, that'll NEVER happen. Missing quarterly stock price targets so your options cash will always take precedence over the survival of the species.
Yup. Many US made products are US assembled. The constituent components are still produced with slave labor and reckless abandon for the environment. There are some regulations around made in the usa vs assembled in the usa but they are minimal.