Why 3D printers won't go mainstream
shapeways.com
shapeways.com
Sewing machines were completely different. First, you had to get a pattern. Patterns weren't cheap and weren't easily available. There was no decent distribution mechanism for patterns, which is half the problem. Then you had to learn how to sew. Then you must collect the materials. For each garment, you would input time.
Tell me how these two models are anything alike? And how could a person from a 3d printing company not get it?
Even if you just wanted to hit print, you can currently only print plastics. To make an object of any complexity you would need to gather electronic components, perform assembly, which may require secondary skills like soldering. Given the state of the art and development roadmap homegrown manufacturers would need to sand and paint the output of the 3D printer to achieve a "commercial" finish and tht requires time and training.
At a high level I think the comparison is apt. It is fairly easy to learn to sew (as easy as learning HTML, for instance) but we still do not have a movement of homegrown tailors on the scale of amateur web designers.
How are those opposing forces?
I am surprised to read so many comments and see nobody articulate what I thought the 3D-printing idea was, which is that it will bring open source to the physical world. No more, no less.
You don't need a degree in mechanical design to use an open-source-powered 3D printer anymore than you need to be a programmer to use open source, with roughly the same capabilities and tradeoffs. Most of the rest of the problems cited in this discussion are supposed to be solved by automation; for instance, I can easily imagine how to help people stock up on raw plastic stock with arbitrary degrees of customization, that's not even remotely a challenge.
"Even if you just wanted to hit print, you can currently only print plastics."
Considering that 3D printers only barely exist today, basing your skepticism of the concept around only the progress that will be made in the next 5 to 10 years is an odd position to take. (And I feel I'm being generous when you explicitly say "today".) There is nothing fundamentally challenging about integrating electronics into a sophisticated 3D printer; it's only engineering of the type open source does every day. We could make a pretty decent pass at a generic electronic loadout today and it will only get easier over time. Already we see full Linux machines barely larger than the plug they attach to.
In fact, it's not even that 3D printers will be "like" open source, the hope is that it will be open source. 3D printers are just another hardware peripheral, hardly different in principle from 2D printers or graphics cards or audio cards or any of the other myriad bits of complicated technology already attached to computing power. Trying to separate the "3D open source community" from the greater open source community would be as futile as trying to separate the "3D graphics open source community" from the greater open source community; oh, there will be specializations and such but there's no way you could produce a criterion that cleanly separates each and every project into either "3D" or "not 3D", the way you can separate "open source" and "not open source" (for a suitably well-defined definition).
1. How are those opposing forces? The first statement would be me selecting an item out of a catalog and printing it as a designer intended, so a traditional ecommerce system, but with no delivery, just on demand manufacturing. The second would be more of an open source model, but customization could be limited through DRM.
2. I didn't say people would need to have degrees in mechanical design, just that they would have to become skilled in it. For most people this is going to take an investment of hundreds of hours.
2. I think you are underestimating the "ease" that plastics for manufacture can be lined up. I think people are lulled into a sense of ease when it comes to 3D printing because we have seen such dramatic improvement in info tech and desktop printing. In the case of software your raw materials are 1 & 0's with 2D printing you have cyan, magenta, yellow, and black ink. Moving into atoms the variety expands incredibly quickly. Materials all have quirks and the process that manipulates them further complicates the matter.
3. With electronics you are still going to face issues. First what is the "ink cartridge" analog for electronic components? How do you handle the regulatory issues? UL, FCC? I'm sure it can be done, its just not that easy.
4. 3D printing is a 20 year old industry. It faces long development cycles and requires significant and costly improvements in material science. Thanks for your generosity, but I will stand by my skepticism.
In closing I don't doubt the rise of "open source" manufacturing. I just don't think 3D printing will have major role in it. You seem to have confidence and a vision for how to make 3D printing ubiquitous. I hope you help move the state of the art forward.
Surely this is better explained by the (current) lack of overlap between the community of tailors and the community of people who are savvy about the possibilities of web-based collaboration. There are, e.g. pattern-exchange sites out there for people who like making their own clothes.
To use your HTML example, you're saying that since few early web hackers had pages dedicated to kitty cats and bad poetry, nobody interested in such things was ever going to turn up online.
I'm interested in seeing how the print materials will evolve. Current materials are great for prototyping and generating forms, but materials like starch are very brittle, and anything less than .125" thick will usually break. It would be great to print models that are not only functional and durable, but viable as end products.
A lot of attention is on 3D printing, but that's connected to a larger set of technologies that make up "digital fabrication," which includes laser-cutting and stereolithography, etc. Laser-cutting is as significant as 3D printing, I think, as you can design interesting 2D patterns as well as folded surfaces, that can be aesthetic or even structural.
For one of my projects this semester, I created a nice set of computer speakers for my macbook, out of foamcore(!). I lasercut the foam to create the housing for the speakers, and encase the circuit board that had the power and volume controls. The next step is to buy higher-end speakers and create the housing out of high-grade wood for better acoustics.
Anyway, the point is, you can make a lot of interesting things with digital fabrication.. I don't think it'll be as ubiquitous as inkjet printing, but more similar to the structure of Software Development Kits, where industrial designers will facilitate co-creating your customized products. I can imagine separate teams of designers making customized laptop chassis, medical equipment, cell phones, etc...
Once molecular manufacturing is viable, this conversation is over so there's not much point in getting too agitated about precisely when the size, cost, resolution and work envelope of 3D printers makes it more appealing to have one in your house or to have Shapeways print it and mail it to you.
Graphic 3D modeling is easy and useful and I have to guess (based on the tools and practice people get in this day and age) that generation to generation people are improving at picturing abstractions/descriptions.
Is this tool even necessary anymore? Isn't it a leftover "futuristic" device from the 20th century?
There are times when a personally customizable object is more suitable than a mass produced one, and that is what 3D printing allows.Eg, a small plastic component breaks, its better to just fab the little part up than replace the whole thing.
If you didn't read it, the author is basically saying that 3D printers will never go "mainstream" for the same reasons that most people don't own or use sewing machines and make their own clothes. But this is almost a completely spurious point and refers only to the current and near future state of technology. He is right in that no-one but nerds would be interested in current or even substantially improved 3D printers, where there's still a lot of cost, time and fiddling around to do in order to print your useless plasticky widget. But who doesn't think the technology will improve, massively?
It's a completely bogus comparison. Clothes are almost a special case. The source textile is important, the way it's designed is hugely important - more important than function, a lot of the time. Good textiles, in case you've ever tried to buy any, are not available below truly massive quantities and half the time not available at all. I once had the idea that I was going to make myself a "kewl" jacket out of black gore-tex (or similar) based on a pattern from a disassembled other jacket. I could not source any material from anywhere in quantities under 100m or so. This is $60+/m fabric or something, so my starting cost, just to get the material, was $6k+ (numbers were probably higher, I forgot). Needless to say, I didn't bother. So it's not like you can just sew anything at all - you are dependent on upstream factors a lot.
There's another factor he overlooks, too. Clothes are a huge but fairly generic category. Most people wear fairly "normal" clothes like, say, jeans and t-shirts, or suits and shirts. If you need, say, another black t-shirt, it is fairly safe to assume you can find one nearby, and these basic items can be produced on massive scale at low cost. Of course no-one is going to sew their own black t-shirt. And most people don't really wear clothes that aren't available at scale. You could make a good case that 90+% of clothing is widely available mass market, and the psychology of clothes prevents deviating from that too much.
But the 3D printing idea isn't supposed to cover the generic cases, it's supposed to cover everything.
No, I think 3D printers will be more like, well, 2D printers, and right now we're in about the 70s level of small-scale print development. Does anyone know someone who doesn't own an inkjet or something? And yet you have to feed that as well - paper and ink - and you have to tell the computer to print. Is that just for nerds?
When the technology is there, imagine buying the raw materials at a nearby shop, just like you'd buy a ream of paper. You plug the thing into your computer, just like a 2D printer. You open the "document" and hit print. Anything the thing can build - and the potential is almost unlimited, in time - will be built. How is that not going to be popular?
It's not going to be this year or next, but in a decade I'd be pretty surprised if there wasn't a much, much larger market for on-demand manufacturing (and, for that matter, book printing). It just seems inevitable. This guy needs to think more long-term.
To your point about 2D printing, think about how uniform the feedstocks are. 4 colors of ink can be sprayed on paper to recreate a rainbow of colors. 2D printing is almost platonic in its simplicity. 3D printers may work with a subset of materials in the future, but its a drop in the bucket.
Also, I see clothes as a special case as well, but one that is MORE likely to be customized. Sure you may just need a commodity item like a black Tshirt, but clothing is one of the most personalized product cateogries. Custom neck ties, new fashions, printed Tshirts, etc. Just look at the startups in the space: Threadless, Cafepress, Zazzle, Spreadshirt, etc.
For example, you rightly mention the awesome variety of plastics. But how many of them are really, when it comes down to it, essential to the functionality of the end product?
I see many, many of these different plastics as existing not because of their inherent qualities in the end product but due to other factors. For example, you mention a type of plastic that bonds well to others. That is a manufacturing concern. It could be essential for the final idealised product to glue two plastics, but it could just as easily be an artefact of the manufacturing process.
Another thing you fail to consider is the disappearance of the scale manufacturing incentive to reduce material costs. If you're making 100,000 copies of something, every percent reduction in your materials cost, and in the complexity of manufacture, becomes very significant. But if you're making 1, the materials cost is almost irrelevant, and you don't care about being able to stamp out one a second. How much could the removal of that factor influence the design?
Consider a milk bottle. In manufacturing millions of them, there is every incentive to reduce costs. So we have the cheapest possible plastic which is food-safe but strong and moldable enough to have a handle, etc. Another type of plastic for a soft drink bottle. Another for a jar of honey, another for your takeaway food container. But every single one of these could actually be functionally replaced by, say, metal. The different plastics, the use of plastic at all, is not fundamental to the function, it is a side-effect of marginal costs in scale manufacturing. This is probably a bad example but I hope it shows the point.
How many types of "building block" do we need? It is probably an unanswerable question but even so, it is surely orders of magnitude less than the question you're answering, which is "how many do we currently have".
And your point about clothes - the "customisations" you are talking about are superficial only, just what is printed on the front! The actual clothing is identical. Cafepress do not design and manufacture a unique type of t-shirt just for you. You are actually kind of proving my point ;)
Of course, we could move to metal for some product categories, but cost becomes an issue there as well, you might be willing to pay more for a one off object, but 20X as is the norm now?
Regarding the superficial nature of most clothing customization, you are correct that most offerings only allow you to change graphics, but Walmart has a tremendous business along those lines. Also, next generation startups like www.propercloth.com are allowing more customization in fit and function.
In any case I think we both have valid points and there is only so much we can accomplish on a message board. If you are ever in Boston give me a shout.
Same principle as ordering a 2D print right now. Some people print at home because it's instant and cheaper and there's no shipping cost. But there's plenty of reasons to order 2D prints online instead, including different materials, potentially higher quality, and lower up-front costs.
The 3D printing concept seems like it works even better than 2D printing as a made-to-order service; it's more dependent on materials, a fancier machine will be able to make better stuff, you probably don't need it instantly, and shipping costs will be relatively smaller.
> every incentive to reduce costs. So we have the cheapest possible
> plastic which is food-safe but strong and moldable enough to have a
> handle, etc. Another type of plastic for a soft drink bottle. Another
> for a jar of honey, another for your takeaway food container. But
> every single one of these could actually be functionally replaced by,
> say, metal. The different plastics, the use of plastic at all, is not
> fundamental to the function, it is a side-effect of marginal costs in
> scale manufacturing.
I’m no expert on manufacturing, but I think you could use the same low-grade high-density polyethylene used for the milk bottle for the soft-drink bottle, the jar of honey, and the takeaway food box as well. It would be thicker and heavier in some cases, and you wouldn’t be able to see through it as well as the PET that’s currently used for the soft-drink bottle or the crystal polystyrene used for see-through takeaway food boxes. (You could also use PET for all of those things. PS is less versatile because it’s brittle.)
I don’t think this is where 3-D printing is going to make a big impact first, though. The objects you’re describing have a large potential market for them, where they compete on material properties and low cost with other objects. You can presumably mix just-in-time filled systems in your 3-D printer to get a wider range of material properties but I think that may not be a comparative advantage for 3-D printers in the near future.
Instead, think about objects that are more customized to their surroundings, where shipping and inventory is a major part of the retail cost: drawer pulls, power strips for the number of plugs you actually have (integrated into your 3-D printed desk), planters, replacement machine parts, prosthetic body parts, screws of funny sizes, and so on.
As someone who owned and learned how to use a sewing machine at the ripe old age of 10, I can't help to find fault in this guy's argument.
Admittedly I was building Rokkaku's, Speedwing Stacks and Parafoils out of ripstop nylon and carbon fibre (well, the parafoils didn't use carbon fibre), but that's not the point.
I recently looked into 3D printing for producing components for another hobby of mine (tabletop wargaming) and the fact is that it is stupidly expensive at the moment in terms of both purchasing a machine or getting someone to print for you. I ended up just learning to sculpt and scratchbuild with polystyrene card.
Now to address some of his points.
1. when someone develops a 3D modeling application that anyone can use and many can master
2. when someone develops a post consumer recycled 3D printing material
3. when printing metal and circuits becomes cheap
It's menu's might be a bit confusing to begin with, but I had the same sort of experience when switching from office 2003 to 2007, until I got used to it that is.
I learnt modelling under 3DS max, purchased and worked on a copy of Maya 2008 and only recently have switched to Blender to work on a different project (mostly to try and prove a concept)
While Blender does have it's own nuances (eg, the menu's change based on context) it isn't any worse than the other tools in the case of workflow for model creation. Nothing overly noticeable anyway. In fact, they advocate learning and using mostly keyboard shortcuts which is handy.
However on the animation side (different kettle of fish understandably) it has a vastly superior workflow and I can produce content at a faster rate, which I would estimate to be 2x faster?
But really, the difference in price between Blender (free) and Maya (starts at $2k per license) and many of the other Autodesk (Autocad starts at $4k, 3DSmax at about $3k) the difference in price is astounding. Even Zbrush which is a fairly artist friendly tool starts at $600
Besides, Blender is open source. If you don't like something in it, change it...