I had the pleasure of hosting the first Digital Factory conference at the MIT Media Lab this week, joining my friends at Formlabs to talk about the intersection between digital design and new forms of computer-mediated fabrication (especially 3D printing). The audience was an incredibly stimulating blend of designers, engineers, academics, artists, and representatives from both brand-new startups and long-established industrial companies. In that respect, as well as in the exuberant feeling that everyone there was participating in something new, it reminded me of the Solid Conference, which I hosted with several of the same friends from MIT and elsewhere in 2014 and 2015.
At the beginning of the day I summarized the context of the digital factory in terms of software, and drew some analogies to argue that digital fabrication could radically transform not only the low-level industrial processes by which things are made, but also the economics and sociology of manufacturing. Here’s a summary of my Digital Factory keynote.
In the 1940s, computer programmers looked like this.
Gloria Ruth Gordon and Ester Gerston programming the ENIAC. US Army photo via Columbia University Computing History.
Look at the cables they’re carrying: in order to write code, they had to engage directly with the architecture of the computer, stating the math problems they wanted to solve as connections between different computation units. Formulating the computer instructions to do arithmetic was considerably harder than formulating the arithmetic itself.
Modern computing looks like this:
a = 3 + 5
print(a)
> 8
In this very simple case, writing the code to perform a calculation is barely harder than simply writing out the calculation itself.
I’ve obviously picked a facetiously simple example, but modern scripting languages like Python and JavaScript are so simple—and free, high-quality resources for learning them are so plentiful—that programming is now within the reach of anyone. Elementary school students can build and program interactive web pages. You don’t need to be a formally-trained programmer in order to fruitfully use programming as a tool to solve some problem you encounter in your business, schoolwork, or hobby.
Beyond the rise of abstract high-level programming languages, I pointed to two other transformative developments that have made it easy to experiment and innovate with software.
The first is the availability, beginning in 2006, of Amazon Web Services and its competitors. Thanks to them, as well as to high-quality open-source software, powerful computing is now available on demand with zero fixed up-front investment. A software experiment that might have cost thousands of dollars in computer hardware and software licenses just 15 years ago now costs only the experimenter’s time at the outset.
The second is the rise of the web browser as a universal platform. Web browsers run everywhere from laptops to phones to car dashboards to inexpensive embedded computers like the Raspberry Pi. A single application written with JavaScript, HTML, and CSS can run in any of these venues with no code changes necessary. Programmers who need to get the best possible performance from their applications will likely work closer to their hardware, but an experimenter can write once and launch everywhere.
You can probably tell where I’m heading with the analogy. The digital factory—especially 3D printing—abstracts the process of fabricating things, reduces up-front costs, and acts as an incredibly versatile production platform for a wide range of products.
In traditional manufacturing processes like injection molding and machining, it’s often considerably harder to design the process by which some object will be manufactured than it is to design the object itself. The digital factory abstracts the process design away—an advancement similar to the move from computing with cables and sockets to computing with easily-readable code. 3D printing inherently reduces the complexity of fabricating a design because there are no molds or tools to interfere with the workpiece. Other traditional techniques are now mediated by an abstracted software layer as well: subtractive processes like milling can be guided by computer-aided manufacturing (CAM) software that’s easy to use and that makes intelligent design recommendations. Even the design process itself is becoming computer-mediated; I’ve written about how artificial intelligence could guide many of the lower-level decisions that designers and engineers make.
Just as high-level programming languages made computing accessible to a great number of non-experts, digital manufacturing will make fabrication accessible beyond its current priesthood. Digital manufacturing may never be as accessible as JavaScript, but it’s a significant advancement if, say, an electrical engineer is able to design, prototype, fabricate, and manufacture her own enclosure without engaging a mechanical designer and design-for-manufacturing engineer. It broadens the reach of an individual with a particular vision, and brings more innovative action within her grasp.
The second element of my analogy—the reduction in up-front costs—is most commonly cited as a motivator for digital fabrication. 3D printing is (in theory) a zero-tooling manufacturing process. In processes like injection molding, the need to fabricate a mold means the first part to roll off the line costs thousands of dollars (or much, much more). With 3D printing, the per-part cost may be significantly higher on the margin, but the first part costs as little as the thousandth. Other digital-factory technologies like the machine-shop automation system developed by Plethora promise to decrease the setup costs for more industrial processes.
With low up-front cost comes cheap experimentation. Big companies and startups alike benefit from the ability to iteratively explore the solution space for some problem quickly and inexpensively.
Finally, the digital factory could become a universal platform: the web browser of fabrication. 3D printing isn’t about to displace the vast constellation of manufacturing processes available today, but its capabilities are growing rapidly and its cost is falling (the Fuse 1 from Formlabs, announced this week, produces sturdy nylon parts and costs an order of magnitude less than the previous standard), so the range of products that could conceivably be made on a 3D printer (complemented, perhaps, by some software-abstracted machining) is quickly widening. Once you’ve figured out how to operate a few key additive and subtractive processes, you’re empowered to design and fabricate a wide variety of objects.
Each of these aspects represents a revolution in accessibility in itself, and combined they’re huge. Anyone can now explore, act on ingenuity, and take a low-risk experiment that could change the world.