Ask HN: What resources to learn tech and software in manufacturing? | Hacker News Reader

Ask HN: What resources to learn tech and software in manufacturing?

73 pointsarcile50 comments

A lot of us have a history and experience in software and often online solutions where information and knowledge is often online and accessible.

I've recently joined a company that do manufacturing, involving supply chain, shop floor management, material handling, inventory, production planning, quality control, tracability etc.

I'm trying to learn as much as possible around today's innovation in these areas, what are the best technologies, best processes, what companies are in the forefront, which are up and coming, but find it a lot more difficult than for pure online software solutions.

What approaches would you take to this, and is there any sources of information that you can recommend? Ways of extracting this knowledge?

Obviously ChatGPT, perplexity and others have been asked.

View on HN

Every job in my 30-year career has been at companies that manufacture something. I have to say that what you're planning is a really big ask. I mean, you could make an entire career out of any one of those areas you mention. e.g., Supply Chain Management is a pretty popular Master's degree program. It might be better to choose one and learn what you can.

But first, why? What's your position at this company? It seems like you're trying to improve something, but perhaps first spend six months at the company and understand how and why they work before looking at areas to improve?

Agree and every topic they list is an entire discipline/career in itself!

Your seriously asking why should someone try to learn more? To live curiously, have growth, engage in lifelong learning.

This response is negative gatekeeping.

I didn’t take the OP that way. I think it’s a gentle suggestion that each topic has far more depth than might first be apparent. As someone who has a tendency to want to dive in and learn everything possible about a subject in a short period of time, I appreciate the reality check.

He literally suggests waiting 6 months. I don't think damping enthusiasm like that is helpful coaching.

Any subject worth learning has depth to it. Nothing worth knowing can be fully understood in a short period of time. That doesnt mean you cant start today.

Six months to understand and learn. If you’re completely new to MFR and Supply Chain, it will take you about that long to start understanding how to think and ask the proper things. This is also about how long it will take to build up rapport with the shop floor crew.

Six months should give you time to fully unwind the dependencies of everything. And learn the history and context of decisions.

You will be working with guys who have been welding metal longer than you’ve been alive. Guys who have machined parts longer than the internet has been around. This is a deep, deep field. Six months is pretty quick.

Did you deliberately gloss over this part?

> before looking at areas to improve?

Without knowing how a process works and why it's there, it's all but impossible to understand how to improve it.

I'm sorry I jumped on your post earlier since you clearly meant it in a positive way. I agree that listening to your colleagues and understanding then business you've joined should be your top priority.

The thing is that the original post never said they wouldn't do this. The post asked for resources and industry best practice. For all we know they would use this information with discression and practically.

You touched a nerve because I get deeply frustrated by the attitude in manufacturing where process knowledge is not readily shared like in software. The reality is that many business are not best practice.

> often online solutions where information and knowledge is often online and accessible.

I haven't had that experience at all, and my long software career has intentionally involved working in many different industries.

In my experience, most meaningful domain knowledge comes from peers, books, classes, trade journals, trade shows, etc

Online resources are almost always feeble and superficial in comparison because most domains don't have a tradition of self-educated people just rolling in and trying to slap stuff together the way software does.

It sounds like you want to take this new opportunity seriously, and I encourage you to try doing that away from your desk and browser. Instead, immerse yourself in whatever you can of those more substantial, real-world sources of knowledge -- most especially by engaging directly with your new peers and colleagues.

There are a number of great comments here, but one of the best things I've seen from "fresh eyes" is identifying entrenched waste.

In manufacturing, eliminating waste/process improvements are always being looked at and often encouraged. The vast majority of manufacturing efforts require bespoke solutions (material handling, planning, quality, etc.), that's most likely why it's hard to find "common solutions". Many get tied into common systems, but it's often a "we made it work" condition and not seamless. So walk the floor, observe what's bespoke, observe what could be common and integrated, then explore those items.

ERP software seems to be the largest contributor in this space. I would look at modern solutions as much as you can. Many companies are entrenched in SAP and Epicor (QAD, Netsuite, etc.) which try to do everything and often don't do anything really well. Start there, but there may be more obvious areas to look at in your observation. Good luck!

>There are a number of great comments here, but one of the best things I've seen from "fresh eyes" is identifying entrenched waste.

Reminds of muri, muda, and The Toyota Way book.

https://en.m.wikipedia.org/wiki/The_Toyota_Way

IIRC the Lean Startup came from that.

https://en.m.wikipedia.org/wiki/Eric_Ries

See Steve Blank.

https://en.m.wikipedia.org/wiki/Steve_Blank

Also:

https://en.m.wikipedia.org/wiki/2009%E2%80%932011_Toyota_veh...

Worked in software for manufacturing for some time. The best thing to do is learn about how processes came into manufacturing and the whys—the tech is less important. Simple concepts like introducing Kanban for inventory allowed us to go from month-long lead times to days. World Class Manufacturing is the bible (https://www.amazon.com/World-Class-Manufacturing-Richard-Sch...).

Attend an industry conference and network until your feet are swollen, voice is gone, and you have a notebook filled with everyday problems.

Someone at your company probably knows the best conferences for your particular niche. Also, talk to the sales people in your company because they spend most of their workday talking to people in manufacturing.

Here's an example of an industry conference (3D printing / additive manufacturing): https://www.rapid3devent.com/

Do you have any suggestions on how to find conferences in Europe? It sometimes appears that the US has more of these.

I'm based in the U.S., but I think this advice is universal:

- Preliminary Google search. - Ask your customers and salespeople what conferences they'll be attending. - Follow leaders in the niche or industry on LinkedIn/Twitter, and they'll post about the conferences they speak at and attend.

Thanks!

My first job was working for an ERP software provider. If you're coming from a more "pure software" background, be prepared for a much slower rate of innovation and and a more conservative mentality. IME, things like open source software is also frowned upon because the suits in charge are afraid of "anyone having access to their code". Don't expect to really change anything with any new ideas for a few years, and even then, political "soft-skills" will be the most helpful for you to learn. These organizations can be very hierarchical and traditional, even in software-oriented company. If you don't feel that way after a few months, then congratulations you're in a really rare company!

You're right, it is much more difficult to self-research, since most of the information is locked away behind proprietary systems. 99% of people in this industry learn what their company does and their niche, but thats it. The fact that you are actually motivated enough to learn about your work outside your job means you're more motivated than almost all of your coworkers (unless you work at aforementioned rare company).

For general research I would say look at the websites of your company and its competitors on their promotional materials. This will tell you what they're excited about and what they want their customers to be excited about.

If you want to succeed at your company in particular, quite simply learn how your company does things from any internal docs or processes. There are lots of acronyms so learning those and building a glossary will help a lot. Building your own wiki (I use Obsidian and its great) as you go is also helpful, but that would apply to any knowledge job.

Was super useful when I got burnt out as a developer and switched to running IT in manufacturing:

https://www.audible.com/pd/The-Goal-Audiobook/B00IFG88SM

If you are a big enough place to have them talk to your production planners, your buyers, your manufacturing engineers. They are probably going to think of you as an adversary at first wanting to 'fix things' which means 'breaking their processes' and 'putting things in the way of doing their jobs'. Lower level people will be afraid you are going to automate them out of a job. You can't been seen as just working for the carpet (people in the offices), but also for the concrete (people on the floor).

Context: worked in the space a bit (~4 years, cumulatively), mech eng with a comp-sci minor, track the space from the outside now.

As others have pointed out, this is, as far as I've seen in the last decade, a very conservative industry in terms of taking risks and trying new things. Some companies build an excessive amount of infrastructure for these operations internally, some go entirely with legacy off-the-shelf stuff (think SAP, NetSuite/Oracle, etc), and some go with a blend of the above. I've seen little traction with modern software, but this is likely a byproduct of my only having seen some small number of companies' operations.

There are a number of companies trying to innovate in the space in a variety of ways, but here are some modern manufacturing ERP companies: - https://fulcrumpro.com/ - https://tulip.co/ - https://oden.io/ - https://www.machinemetrics.com/

I can only recommend visiting any industrial trade shows. New and upcoming products and services will be shown there in plenty. Skip the sales people there and ask for an engineer. They are normally happy to talk to you openly and answer any question. I know this might be the solution involving the most work (and cost) on your part but it’s also one of the fastest ways to build your network in this space.

Was a MechE for a decade, but I do software now. Make friends with a design MechE and see if you can tag along for a single new part release. You’ll see the whole process from how the part gets designed and quoted, when it gets to IQC (incoming quality checks), how it get inventoried, added to BOM, pulled for kits during work orders, etc.

Do not take me wrong but... In PRACTICAL terms: obsolete tech, 99% of IT-related tech in manufacturing industry came from another era, something might be neat and still relevant today, but most is crap.

In theoretical terms, we are approaching the third revolution:

- the first was the factory, meaning series production

- the second was CNC, or subtractive manufacturing with generic machines able to produce generic parts

- the current, still in pioneering phase is additive manufacturing, aka 3D printing

Aside many new experiments. But that's a niche, the real world is mostly simply obsolete tech, badly kept up with strange and undocumented hacks. The part you describe (IM, PLM, ...) are not really manufacturing, just common tools from pure IT migrated elsewhere, in software we have SCMs, CI, CD, ..., in manufacturing they are named PLMs but they are just different software doing the same things. Since "programming languages" used in mechanical industry instead of languages with a VM or a compiler or an interpreter are CAD/CAE/CAM suite their SCM/PLM tend to be one of the three giant suite, Catia (Dassault/IBM), Creo (former Pro-Engineer) and NX (Siemens or someone else, I do not know actual situation, in the past was Unigraphics). Other industries are still on 2D solutions. Rarely some industries use different "less omnicomprehensive" suites. You can learn something from some published courses but most training material is on sale, not freely available, because well... Anything is a product in industryland...

Anyway at this level Catia and surrounding are definitively the most feature rich and advanced, NX is and Creo are far more limited, they have a bit of PLM and simulation but they are not meant to be "an anything tool", Dassault try even to allow modeling the Earth in their suite. Beside that any company have choose a genital lace to one of the relevant "software partner" and essentially they are tied to it like on the business side they are tied to an ERP.

Rather than trying to learn the latest innovations in these areas, focus instead on understanding the core. And you work on in a company doing this stuff, so I would say first and foremost learn from your colleagues! Talk to someone in each area, try to understand their role, and the key considerations in that role. You will get much deeper and more applicable learnings compared to anything online. EDIT: try also to visit customers and suppliers (and the customers customers), to understand how your company is part of a supply chain

In modern manufacturing, identifying and understanding bottlenecks in the process is key to knowing how to improve throughput.

One of the best resources to learn about manufacturing optimization aka the Theory of Constraints is a novel called “The Goal” by Elyahu Goldratt [1]

[1] https://en.m.wikipedia.org/wiki/The_Goal_(novel)

I work in the hardware space. Here’s what helped me get my feet wet the fastest:

1) Make friends with the technicians on the floor 2) Find ways to either shadow, observe, or actually learn to do what they do 3) Walk around the shop floor and watch (safely) and ask questions (safely)

MFR has a lot of nuance and you you won’t get a good grasp for the intuition behind being good at this until you encounter the problems and corner cases in person.

As for innovation: unless you have joined an “innovative” company, innovation will be hard to find. Iterating through manufacturing method is expensive, and each of the major methods (CNC Machining, Additive Printing, Metal Forming, Welding, and Automation) is an entire career’s worth of skill and knowledge on their own. Some of them are old and have been innovated already. Some face physical limitations in material science. Some have maxed out what’s possible under our laws of physics.

The leading edge right now centers around robotics and additive printing (especially metal printing). Try to get involved there.

Supply chain is almost purely an experience gain. You have to develop the muscle and intuition to navigate complex supply chain problems when they arise.

It’s a big field! Lots of unsolved problems. Good luck!

How much of it feels like Factorio?

Almost none. Because Factorio handwaves a lot away.

There will be a LOT of walking around though.

It is a bit expensive for what it is, and maybe predates the waning of IoT hype, but by far the best that I have come across in a single volume is "Handbook of Industry 4.0 and SMART Systems" (978-1032103433). It is readable by someone who is not involved in all aspects of manufacturing tech, but understands enough about more modern stuff, and importantly, what you would want things to look like.

Rule #15: "Consider all labor and output constrained machinery equity often bring hidden variable costs and liabilities. Thus it is generally best practice to exchange in-house production stability with purchased supply/service chains to externalize risk."

If you have to ask this question, than your survival as a company should avoid naively competing with established contract manufacturers. Startup probability of success is 1:22, and ratio of product to service company survival rate is 1:6. That means even if you are an expert and do everything right, the chances of survival of half a business cycle is 1:132... The investors are already acutely aware of these numbers if they are competent.

Best regards =)

You remind me of Chris Power, founder and CEO of industrial automation startup Hadrian, he once posted here similar ideas, now he is doing it, have a look at what he has been doing. He has narrowed his manufacturing focus to high precision parts.

Do you have some of these posts? That sounds interesting.

I consult for manufacturers and fabrication shops. In my experience there aren’t really centralized repositories of info for this stuff like there is in tech. It took me a while to get my head around it, and many years of talking to people and visiting their facilities, to realize that knowledge just doesn’t seem to be organized the same way in manufacturing as it is in tech. Most of the good stuff I’ve learned through conversations with guys running these facilities, not through some central knowledge base. I’d love to be proven wrong though if someone has a good set of resources out these...

It doesn’t help that the knowledge in this industry is 1) held by people who statistically are less online than software engineers 2) are closely guarded “trade secrets” (job security!) 3) is in one of those “is more art than science” fields where the knowledge gained is hyperspecific

I've worked in IT and software development for automotive manufacturing for over 15 years.

> Ways of extracting this knowledge?

Ask questions. Ask obvious questions. Walk the plant floor.

Be prepared for things to not make sense. The only real explanation is a long history of "They made it work"

Your machines are named Machine B, Machine A, and Machine C because A and B went in first, C was added later and there was only room on that side.

The problem is not there was no plan - the problem is there was a plan in 196x, and another plan in 197x, and .... (See "Standards" https://xkcd.com/927/ )

Learn about Lean Manufacturing and Toyota production system obviously.

Traceability - do you know what your bill of trace is? Do you really know? Can you prove that it will never be wrong? Do you audit your trace process? Do you audit your trace captures? All of them? (I worked on traceability for a few years. I will never work on a compliance app again)

Innovation in the areas you mentioned will always be about balancing constraints. The answer to "Why don't they just do X" is: try it and see. It will always work in your mind. You don't know what constraint is the hardest until you actually build it, whatever it is. The hard constraint will move around depending on exactly what you are doing - it's not consistent between companies or between lines of business or even between stages in the process.

Physical automation in manufacturing is more difficult than it seems at first, second, and third thought. Automation does not solve problems. You must solve every problem the automation will encounter.

Authors: Sidney Dekker, Clayton Christensen, W E Deming - not all about manufacturing specifically, but a very good basis for thinking about people and business.

If you're talking about manufacturing in regulated industries (e.g. Pharma) then also consider resources like Good Automated Manufacturing Practice (GAMP). There are a variety of standards that companies in this space must adhere to with respect to industrial control systems.

These aren't technical documents but they do provide relevant background to understand the landscape and might shed light as to why there is little "innovation".

Also, if you're at a place where they are implementing GMP, they probably have formal training and training records. It can be a good idea to look around the system and see what other online training outside your own discipline you have access to. A lot of the GMP documents that manufacturing people have to be retrained on annually will give you insight into why things are the way they are and what that particular company's way of handling processes looks like.

When I asked a similar question of a VC that's invested in manufacturing they recommended I tour factories and ask what their painpoints are.

For the logistic I assume there is special software around. I would say this falls under ERP systems although I am not an expert.

Quality control and traceability is likely more customized software because challenges are more unique to the product and the industry you are in.

I am not sure what you goal is. Are you a software engineer entering the manufacturing space or what's the goal?

I have been trying to do the same, but what troubles me is that I had to pick up anecdotal knowledge and then drilling down and sideways on whatever manufacturing term I came across.

I was not able to find a knowledge path that I could follow, is there a degree out there that lists its textbooks?

There are many books on MES and ERP which are good for an overview and structure. Then read documentation for the specific vendors you are working with. I have transitioned to this space over the last couple years, and mostly I use LLMs at this point.

https://config.com/. Github for hardware teams, gives some insight into the artifacts required for building physical products.

You haven't said what your role in the company is, and why you have set the goals you mentioned.

Those things will have a bearing on what you should do to achieve your goals.

Here is a 'crazy' idea:

Skim some of the source code of an open source ERP package like Odoo, depth first, breadth first, or a combination of both, module by module. And here, by modules, I mean ERP modules, like sales, purchase, inventory, etc., not Python modules, although Odoo is written in Python, so it has Python modules too, in the source code.

Due to Python, the code is relatively high level, so somewhat easy to read.

Of course you need to have a reading level knowledge of Python, but, you know, 'executable pseudocode'. ;)

Odoo is supposed to be popular and to have a good number of vendors and other things in its ecosystem.

Disclaimer: I have only read these points, not verified them for myself.

(But I do have some background in the manufacturing software domain, and in Python, both, hands on, for non-trivial periods.)

This is what I would do if I had your goal, apart from things that others have suggested, such as reading books about manufacturing and ERP.

Be warned: manufacturing is a huge domain.

Check out Enzyme, it's the only software that's been able to innovate in this space for a while

Read "the goal"

Check out https://learn.umh.app

We have a lot of free info material on the topic of IT and Industrial Automation including Shopfloor KPIs like OEE. And the project behind it, is even open source.

Disclaimer: I am the CTO

this is all old school proprietary software solutions...you are at the mercy of marketing, sales, vendor lock in, engineered obsolescence, continuous upgrades, not reusable skill sets, etc, etc...i was there for a while and am happy to have left it behind...good luck

how did you leave it behind? I am at their mercy right now and interested in switching to a stack where people actually value the tech rather than just tolerate it.

Nobody but developers care about the tech. People only care about what it can do for them.

leave behind as in "move to better pastures", "change job"

Reply on news.ycombinator.com