Future of Manufacturing

January 26, 2010 10:59 AM

Wired magazine has a long and detailed article about the future of manufacturing. Short version: the same kind of democratization that technology has effected in publishing, music, video, etc., is opening up design and manufacturing to anyone who wants to participate.

The potential for home-based fabricators, even replicators, is revolutionary.

posted by yesster (41 comments total)

The author, Chris Anderson, explains his own foray into this new kind of manufacturing with his DIY Drones.
posted by yesster at 11:02 AM on January 26, 2010 [2 favorites]

The current issue of Make Magazine is dedicated to desktop manufacturing too. And actually, I'm not to excited by the state of the art. The potential is enormous, yes. But even the most advanced of the open solutions is pretty crappy. Jaggy shot glasses and semi-functional gears seems to be the vanguard.

It only just occurred to me to put this together with another thought that's been rattling around in my brain. Another topic in this issue of Make is CNC. I see a lot of raving (around the internet in general) about CNC machines of various types. But very little of the raving is about the stuff you can make--it all seems to be about the fact that it's automated. I'm guessing this has to do with computer programmers (who are largely the people most excited about CNC) loving processes more than results.

It's easy to imagine that the crappiness of the results is due to limitations in materials and whatnot and I'm sure that's a lot of the problem. But I wonder if it is also due to a lot of the people involved caring more about doing something that They Can than doing something that They Need.

(Not that this is a bad thing. Eventually people with Needs can take the handoff from the people who Can and it transitions to a mature technology.)
posted by DU at 11:18 AM on January 26, 2010 [4 favorites]

(flagged - first link needs fixin)
posted by yesster at 11:20 AM on January 26, 2010

"Replicators" link is borked. You may want to alert le mods.
posted by Kraftmatic Adjustable Cheese at 11:20 AM on January 26, 2010

Links fixed.
posted by mathowie at 11:25 AM on January 26, 2010

Neal Gershenfeld's been working on Fab Lab stuff for years now; I remember reading a book by him that blew my mind and made me want a CNC table and a landfill. Someday I will have the wherewithal to run a free, online CAD library of stuff anyone with a computer, a cutting device, and some (s)crap can make.
posted by carsonb at 11:28 AM on January 26, 2010 [1 favorite]

The current issue of Make Magazine is dedicated to desktop manufacturing too. And actually, I'm not to excited by the state of the art.

It's an interesting issue but so far the only real practical use I can see is being able to go from .jpg to frosted cupcake with the push of a few buttons.

Still, that's pretty damn huge.
posted by bondcliff at 11:33 AM on January 26, 2010

I like that photo, the guy looks like something out of an old sci-fi movie or something. Then there's the caption:

In an age of open source, custom-fabricated, DIY product design, all you nee to conquer the world is a brilliant idea.
Photo: Dan Winters

All you nee people. All you nee.

Hopefully this DIY revolution will bring the same quality to our physical objects that the internet has brought to our standards for the written word.
posted by delmoi at 11:36 AM on January 26, 2010 [1 favorite]

Manufacturing stuff is uninteresting. It's a solved problem at all scales.

Designing stuff is still goddamn hard.
posted by GuyZero at 11:41 AM on January 26, 2010 [1 favorite]

Manufacturing stuff is uninteresting. It's a solved problem at all scales.

For some values of $.
posted by DU at 11:42 AM on January 26, 2010 [4 favorites]

I started getting interested in this a few years ago when I was working for a metal products shop that was circling the drain. I did a bunch of research and pitched the G.M. on the idea of opening up an online portal for custom short-run orders - much like what Arduino is doing for circuitboards and so forth. We never would have had enough time to get that going, but the research has stuck with me.

DU - yeah, the output isn't much to crow about yet, but hey, baby steps and whatnot. The point is that it is a rapidly accelerating process.

Part of what intrigues me is that the products we live with now have largely unnoticed artifacts of their production processes. Your phone consists of discrete parts - battery, screen, keyboard, case - that are snapped/glued together. With unitary design and manufacture, the design can be more organic, with say the battery, keyboard, and case blended together without any discrete separation. Also how time changes - if you're making 100,000 phones, you need to make them quickly in batches that can be palletized and shipped. If you're making one phone on your fabber at home, it doesn't matter if it takes 12 hours - you start it up and then go to work or sleep and don't think about it till it is done.


Manufacturing stuff is uninteresting. It's a solved problem at all scales.

Designing stuff is still goddamn hard.

But that's kind of what this is all about. When the constraints around manufacturing are changed the design possibilities change. Componentization, for example, is strictly a function of modern distributed manufacturing. Unitary construction opens a different kind of design space. Its also about changing the relation between design and manufacture in important ways. As the linked main article explains, a good designer or design team is constrained by the currently dominant large manufacturer model.
posted by yesster at 11:48 AM on January 26, 2010 [4 favorites]

Part of what intrigues me is that the products we live with now have largely unnoticed artifacts of their production processes.

Not to mention all the saltwater they have to cross. Maybe the most exciting aspect of localized, low-run production (and open-source, freely-distributed design) is not having to ship everything we own halfway around the world anymore.
posted by carsonb at 11:55 AM on January 26, 2010 [1 favorite]

Componentization, for example, is strictly a function of modern distributed manufacturing.

Well.....I disagree with this for two reasons.

1) Modularity is still a good design principle. Software (often) isn't assembled from pre-existing standardized parts, but it is still made modularly. It helps with debugging and upgrades. It also helps to be able to "chunk" part of the design and think about it only in terms of inputs and outputs. Similarly, I think it still makes sense that I might want to design the interior of a phone and let my friend design the box to put it in because that's what we are good at or what we are interested in or all we can handle at a time.

2) And from the consumer's perspective, I might want to download the case for my phone separately from the internal workings because I like how one source did one and how the other source did the other. Without standardized interfaces, this is impossible.
posted by DU at 11:58 AM on January 26, 2010

Componentization is also the easiest way to charge for IP - selling a component IC for echo cancellation is a lot easier to track and charge for than licensing the IP for the circuit design. Also, bespoke IC manufacturing is still a way off for small production runs.

Without standardized interfaces, this is impossible.

This is a double-edged sword. If you look at modern PCs there are a lot of standards but they tend to dictate the final design so anything really innovative has to break a standard or two somewhere. It's nearly impossible to buy a replacement PSU for your neato SFF PC. Even so though, you are correct that without component interfaces you have to design everything every time which makes doing stuff really, really hard.
posted by GuyZero at 12:04 PM on January 26, 2010

I'm not really knowledgeable about this kind of stuff, but recently, I really wanted to get a rep-rap type thing, because I'm working on a board-game.

I'm ending up outsourcing the prototype to bluepanther games to start with. But it would be nice to just flip a switch on , press "print" and here's my game printed out.

Would be nice, but yeah, I'd like to see how good the tech advances first.
posted by symbioid at 12:25 PM on January 26, 2010

Everything has its uses. Rapid prototyping isn't really meant to build a production model of anything; you "print" something in resin or cornstarch or whatever because you want to visualize a three dimensional object. CNC is great for large-scale production, but the time and effort programming tool paths and whatnot into the machine make it sort of a silly option for a lot of one-offs or small scale runs. Some automated machinery can make cuts without the complicated programming, but I believe they're still limited to 2 degrees of freedom (and not exactly cost-effective for home use). What you need is a 3+ DOF mill that doesn't need the complicated tool paths.

(Insert comment about how no one gets their fucking hands dirty anymore here.)
posted by backseatpilot at 12:26 PM on January 26, 2010

And now I link one of my favorite short stories. Please don't hate on it just cause Doctorow wrote it. PrintCrime
posted by ExitPursuedByBear at 12:32 PM on January 26, 2010 [1 favorite]

Rapid prototyping isn't really meant to build a production model of anything; you "print" something in resin or cornstarch or whatever because you want to visualize a three dimensional object.

In RepRap's case, our models of 3D printers are working models.

Also, lost wax casting is a favorable technology if you want to go bronze and therefore demonstrate you are doing 'real' art.

Self replication of subtractive machines, e.g. things with large routers or small dremels on them will take a bit longer, but we're working on it.

Lastly commercial RP machines are doing laser or electron-beam sintered steel powder dipped in bronze, and so on. Those workpieces are for production as well.

------------------------

I haven't read PrintCrime yet. 'Kiosk' by Bruce Sterling is good, although I'm ... having trouble finding a copy online rightnow.
posted by sebastienbailard at 12:42 PM on January 26, 2010 [1 favorite]

Someday I will have the wherewithal to run a free, online CAD library of stuff anyone with a computer, a cutting device, and some (s)crap can make.

Like, say, Thingiverse?
posted by kableh at 12:49 PM on January 26, 2010 [1 favorite]

The current issue of Make magazine is all about this.

There are still a lot of things (like welded stainless tubing stuff) that this tech can't do.

Also, like so many "technological breakthroughs", it's really about available capital and marketing.

Somewhat more interesting is the Open Source Hardware movement.
posted by warbaby at 12:58 PM on January 26, 2010

I read both the most recent issue of Wired and the recent issue of Make and part of me wants to run out and get into 3d printing head first, the other part of me thinks that the long term effect of this technology is that people who have capital to afford it could become rich, but in the long run it'll equal less jobs rather than the more that Anderson implies.

Has anyone tried the build your own CNC machine for wood yet? Supposedly it's possible to do it for $800 or so which is about half of what the smallest craftsman version would cost.
posted by drezdn at 1:16 PM on January 26, 2010

Like, say, Thingiverse?

Yep. The eternal mystery is why I've been telling people about this for years instead of actually doing it.
posted by carsonb at 1:18 PM on January 26, 2010

Techshop is a really cool idea. The hardest part about trying to DIY anything is investing in the tools (and gaining the knowledge), a well outfitted place that anyone could use could be really cool. I know I would use it until I can get my hands on a band saw and lathe.
posted by drezdn at 1:22 PM on January 26, 2010

What gets left out of a lot of the breathless discussion and shrill marketing around rapid prototyping tools is that they have knock-on effects that are really interesting. The primary effects are that people are building more things, and sharing more ideas.

This, in my opinion, is what's great about these machines, because in many cases their manufacturing ability is rather limited (most laser cutters, for example, can't cut things thicker than 1/4", can't cut metal, etc). I like seeing the little reprap shot glasses not because we need another sloppily made shot glass (far from it, ugh), but because it represents a lot of effort someone put into building something, which they might otherwise not have done.

The more people we have making things, IMHO, the better off we are.
posted by fake at 1:59 PM on January 26, 2010 [1 favorite]

For the backyard hacker, design is not the problem, fabrication is. I'd love to have a CNC milling machine and space to put it. But until then -- are there currently places that will do 2-D machining operations on sheet metal? That alone will let me build the cool things I want to make. Just simple stuff like polygons with holes.

(Like BatchPCB will make PCBs of your designs.)
posted by phliar at 2:08 PM on January 26, 2010

The more people we have making things, IMHO, the better off we are.

In general, I agree with this. I am a bit concerned about how much pointless crap people end up making out of plastic though. At least with wood or metal, if you make something and grow tired of it there's still other things that can be done with it.
posted by drezdn at 2:18 PM on January 26, 2010 [1 favorite]

drezdn, I agree wholeheartedly.

I think a lot of projects we see now are of extremely low quality. But like anything, we have to start somewhere, and some number of those people who made crappy plastic shotglasses/wrote poor quality code/baked a shitty batch of cookies will become virtuosos.

After all, for many young (and old) Americans, there is no precedent, really. Legos? Shop class? It's hard to make your own things, so the beginning/learning stage of much of this is going to be extra ugly/fumbly.
posted by fake at 2:23 PM on January 26, 2010 [1 favorite]

I am a bit concerned about how much pointless crap people end up making out of plastic though

Then design a mechanism capable of grinding up formed ABS, melting it, and spooling it back out onto printer reels. Hey presto, pointless 3D-printed crap gets recycled into new raw printing material.
posted by aramaic at 2:28 PM on January 26, 2010

Metal can be cut with water jet cutting. I've had it done and the setup charges make small batches cost prohibitive.

This is true of a lot of CNC work: the scale of the job is still expensive.

You can have your own laser cutter (a small one) for around $10,000. But people who have $10.000 to spend on tools usually have money to spend on other things, too.

I recently had some stainless welding done at a small local engineering shop. They were the only ones would touch a job this small. And the bill was so high, I'll never do that again.

I did find somebody in Hanoi who could do the work and send it through the mail. It's slow, but the cost is far less.

So the world of small makers is still bumping up against cost and scale limits that no amount of geewhiz writing by a Wired senior editor and his hobby business has overcome.

Small scale PCB fabrication is getting darned cheap, though.
posted by warbaby at 2:35 PM on January 26, 2010

It's all fun and games until *click* *SNAP* *TWANG* MY EYE!
posted by Artw at 3:33 PM on January 26, 2010

I'm not to excited by the state of the art. The potential is enormous, yes. But even the most advanced of the open solutions is pretty crappy.

The downward push on price & increased availability is happening at other levels than home-tabletop, though. Rapid prototyping used to be something you did to make *one* of something before you committed to making a tool (mold) which is expensive, but now you can make tens or maybe hundreds of something, enabling small businesses to do semi-custom work.

Tech Shop and getting your hands dirty is cool, but another possibility is a Kinko's-like "clean" 3-D print service bureau where you send them a file & go pick up your gears or knobs or cut-to-length 80/20 pieces later that you screw together back at your shop.

Small-scale/price CNC is still interesting and growing. A cabinetmaker can automate some of her work on a ShopBot and focus on hand inlays or installation or drumming up business and can buy custom made knobs from a designer with an account at a (future) metal CNC service. What you can get for $2k (Taig or Sherline) hasn't improved much lately, but mills made from increasingly Common Off The Shelf stuff in the $10-100k range is improving.

I dunno- mill tools can't get cheaper, can they? And making toolpaths for tricky shapes isn't getting any easier or more automated is it?
posted by morganw at 3:55 PM on January 26, 2010 [1 favorite]

Desktop Manufacturing means EVERYTHING will look like a geocities or myspace page.

Are we really ready for blinking unicorn lawn ornaments?
posted by empath at 4:21 PM on January 26, 2010 [1 favorite]

The potential is enormous, yes. But even the most advanced of the open solutions is pretty crappy. Jaggy shot glasses and semi-functional gears seems to be the vanguard.

The fact you can print a jaggy shotglass at all is a god damned miracle, imo.
posted by empath at 4:23 PM on January 26, 2010 [1 favorite]

Techshop is a really cool idea. The hardest part about trying to DIY anything is investing in the tools (and gaining the knowledge), a well outfitted place that anyone could use could be really cool. I know I would use it until I can get my hands on a band saw and lathe.

I need to write a blog post about this that I can link to, but: High school and vo-tech shops. Imagine the incredible surge in innovation and productivity if public schools (which already exist in all but the micro-micro-micro-tiniest communities) opened their shops to their local communities. A lot of these places are shutting down (or have already) but a lot of them are still around and/or can be started back up with a relatively small investment. Freely accessible TechShop-alikes in every town in America.
posted by DU at 5:36 PM on January 26, 2010 [3 favorites]

Very interesting article, thanks.
posted by blue shadows at 6:28 PM on January 26, 2010

OK, here's a smattering of what's available:

online Shopbot atelier 100KGarages

Ponoko, a vertically and horizontally integrated maker market

Big Blue Saw - on line waterjet cutting (stuff that lasers can't do)

Custom Laser Cutting (wood and plastic, not too thick)

What's really needed, IMHO, is open-source microfinancing for people who have good design ideas, but don't have $50.000 (which is about the threshold to get something rolling.) Something along the line of the Mondragon corporation, brought up to date.

The Arduino / Processing confabulation for physical computing is a whole 'nother can of worms that deserves its own thread. (saw Arduino mentioned above, but open source hardware/software is pretty much what started this whole show.)
posted by warbaby at 8:00 PM on January 26, 2010 [5 favorites]

Gizmodo weighs in

What I don't get: Why is it revolutionary to outsource you manufacturing to slave labor? What happens when that super-cheap labor goes away? And... How does this change the time + effort = profit dynamic?

Sure, great, we will all be able to make our own cell phones... but who will want to? Maker types for sure... but the rest of us will still buy them because we have other things we prefer to do with our time. If making things takes significant time & effort, wouldn't that mean those things will then need to provide some kind of income? So we're kind of back to the future... Those who make things will sell them... More people will prefer to buy pre-made things than make their own... I just don't see how this changes manufacturing. You still need to produce things in volume... Maybe in generally smaller volumes... but still not do-able at home DIY style.

Similar theme for these RepRap machines as cool as they are. When will the be $150 and NOT require tons of effort to program? I still can't print simple 11x17 color posters at home (nevermind anything bigger).
posted by MeatLightning at 10:14 PM on January 26, 2010 [1 favorite]

eMchineShop will do small batches of things that you design yourself in their simple CAD application.

That being said, a lot of manufacturing is not simply extruding/layering or cutting. Heat treating metal parts is one example of a vital step that is diffiult to do properly on your own. But I guess that designs will evolve to take advantage of what this technology can do. Interesting times...
posted by Harald74 at 1:56 AM on January 27, 2010

Yeah I am not really sure this is as likely as they say, or if it's a good idea. Most of those rapid prototyping machines produce parts which are expensive, inaccurate and weak. This is also not great in terms of sustainable energy and resource use.

A better solution (which already exists, just could be cheaper), is to email your designs to a manufacturing centre, which produce them with proper machines and then mail your part back to you. Like eMachineShop above. I'm pretty sure that's a better approach, I guess the problem now is that few people have the design skills or the inclination to use these services. And I doubt that people will pick up the design skills all that easily from a home kit.
posted by theyexpectresults at 2:05 AM on January 27, 2010

DU's idea is interesting, but it would really take a forward thinking school district and really good liability protection. It would be so cool though.

Personally, I wish I had access to my old High School's foundry. Then I could try to churn out my own metal shop following the Gingery plans.
posted by drezdn at 5:31 AM on January 27, 2010

To pick a different aspect of the story, it looks like the gents making the cars at the top might have the quality down pat. Do that with an electric and I'm right there. Subject to finding $50k.
posted by imperium at 2:11 PM on January 29, 2010