SUPERMECHANICAL.BLOG

• 2012-05-25

  A graduation of sorts: assembly, beta, schedule

  

  Hello, supporters —

  We’ve got another action-packed update for you. But first, let’s talk about the schedule.

  We’ve come a long way in mere months. We’ve got what we think is final hardware, but it needs to be tested in a small batch before we spend all the money to produce thousands of Twines. We have a duty to spend your pledges wisely. We’re going to ship Twine only when we think it’s great, and that means it will be longer that we wanted.

  Here’s the plan. 100 beta units will be shipping next week. We have picked some Kickstarter backers to help us with beta testing — those of you will be getting an email asking you to help us by putting Twine through its paces and reporting issues. We’ll start by testing just the hardware and gradually add features over the next several weeks.

  After two weeks of field testing, if all goes well, we’ll set in motion the wheels of full production:

  • Full case and circuit board manufacture - two weeks
  • Board assembly - two weeks
  • Software installation, QA - one week
  • Packing, shipping - one week

  Our schedule is driven by the hardware production, but this doesn’t mention the ongoing work on software development, packaging and other fronts. With so many interdependent tasks, you can see how important coordination is. (Gantter has proved invaluable for project management.)

  Assembly

  You got a sneak peek if you follow us on Facebook or Twitter, but last week we got the beta circuit boards, as you can see above.

  We took inventory of the components — all 175,697 of them — and loaded them into the car to deliver to our assembler.

  

  For assembly of Twine’s circuit board, we chose Spin PCB in New Hampshire because they’re nearby and a pleasure to work with. When you’re very particular about getting things done right on a tight schedule, that’s invaluable.

  Ever wanted to know more about automated electronics assembly, the technological wonder that drives modern industry? Here’s a tour of the process.

  

  Those printed circuit boards pictured at the top of this post first get placed in this machine, which spreads solder paste through a mask onto the boards. It’s not too different from how t-shirts are screenprinted.

  

  The paste-loaded boards are then put in the pick-and-place machine. It has what looks like an inkjet printer head, but uses suction to pick up the tiny electronic components from trays and reels, and place them on the boards.

  

  Here, reels of components are being loaded for the pick-and-place machine. The machine uses computer vision and sensing arms to place the parts precisely on the correct spots on each board.

  

  The solder paste keeps the components in place until the boards can be run through the reflow oven, which melts the solder to make a permanent connection. Yes, you can cook a pizza in it, but this oven carefully controls the temperature to avoid damaging the components.

  

  And here are the completed boards, fresh from the oven and ready to snap apart!

  On other fronts

  

  And just in time, the first injection molded inner cases have arrived. Making objects is hard, but making interfaces between objects is harder. It’s a great feeling when the circuit board snaps into the frame perfectly the first time! That’s thanks to extensive 3D prototyping.

  Before the board design was frozen, David worked to eke out every bit of power savings that he could. Now he’s chewing through firmware to make the network connection as reliable as possible. Rehmi and Arash, our engineering intern, have been guiding the circuit board production. After getting the tooling for the cases underway, John has been working on packaging and making the network setup process for Twine simple. Around this crazy schedule, we’re uprooting ourselves from Boston — the next update will be posted from sunny Austin.

  Through our updates, you can see how we’ve adjusted to meet the exceptional demand from our Kickstarter backers, and the love and sweat that we’re putting into making Twine great now and for the future. We’re among the first to undertake a nimbler and more open product development process, and for all the bumps and bruises, we love it. We’re excited about how much better it will make Twine, the product, and Supermechanical, the company. We’re glad that you’re participating with us in this.

  We can never say it enough: Thanks for your support and patience.

  — John, David and team

• 2012-05-10

  String and Stone

  Hey all!

  First of all, we have some news to share, in concert with another Kickstarter you might have heard about:

  We’re pleased to announce that Twine will talk to the Pebble smartwatch. In Twine’s web app, you’ll be able to relay word of real-world events, like a basement flooding or a door opening, to the Pebble on your wrist. Welcome to the future!

  Eric and the rest of the Pebble people are true gentlemen, and we’re happy to be working together to make your world even more connected. This idea got rolling when Nathan, a backer of both our projects, made the suggestion. And we expect that once again, you’ll come up with more great ideas that will surprise us!

  (If you’re new here, please feel free to follow us on Facebook or Twitter.)

  

  On to the work we’ve been doing lately. You may notice the sneak peak for Twine’s web app in the video above. Jay and John have refined it inside and out, and it’s built for future expansion. There’s still work to be done on the action modules (what happens after ‘THEN’), the compiler that turns rules into Twine code, and making the whole chain from web to Twine nice and strong. We’ll take an in-depth look in a future update.

  In hardware news, we’re near the end of our engineering slog. One big milestone has been checked off: we’ve ordered the production tooling for the cases! That’s one big item off John’s plate. The design is cast in stone, so to speak (actually, $20,000 worth of metal), and we’ll be getting the first injection-molded parts from Minnesota in two weeks. Which is great timing, as our first lot of circuit boards (the guts that make Twine work) are going to the assembler in New Hampshire on Monday.

  Hardware development always has surprises waiting for you, and we’ve endured prototypes (and a laptop) being stolen from our car, material samples being stolen from our doorstep, molds for a component breaking, uncountable prototypes snapping, exceeding the world supply of a part, and more overnight shipping than you could imagine. But the light at the end of the tunnel is getting brighter - thank you for your patience and companionship on this trip!

  Finally, a bit of news about Supermechanical. This summer we’re moving to Austin, Texas, which we estimate will make us 60% sunnier. It’s a great technology hub in which to grow our vision. And we’ll be hiring there! If you’re fluent in Python, embedded C, EAGLE or JavaScript, and are excited about connecting lonely objects, check out our job posts.

  — John

• 2012-04-20

  Refinement is unglamorous.

  Hello, backers — it’s been a while. We’ve been working as hard as ever, but refactoring C functions and moving lines by .005” do not lend themselves to blog posts. It’s the dog days of development.

  Remember that pile of case prototypes in the last update? We made a whole new pile, and these proved our redesign:

  

  Each model got us closer to the right fit and feel, and we’re about ready to commit to the molds.

  Many people think that 3D printing is the future of manufacturing. It’s not going to replace injection molding any time soon — the additive process produces parts that are rougher and weaker. (We keep breaking them.) But it is the present of prototyping. On a schedule as tight as ours, it’s invaluable to be able to test how parts fit together, tweak the CAD file, and get a new printed piece in two days that’s accurate to a few thousandths of an inch. You can even print in a rubbery material as I did to test Twine’s sleeve. As I mentioned on Facebook, it smells like Starburst. (Or ammonia. Whatever.)

  
  On the hardware side, we’re likewise days from freezing the final candidate. Prototyping electronics for production is not as quick. David and Rehmi lay out the electrical paths for the circuitboard, order the bare board with those paths etched in, and then once it arrives, solder all the components on by hand. Including performance testing, the process takes about two weeks each time.

  And hardware is not as easy to edit as software. One of the tools of the trade are “blue wires,” very fine wires that are patched onto a board to test changes before they’re committed to the next revision. You can see here that we only have two on the latest rev.

  

  Progress on the software front has been steady, and we’ll cover that in next week’s update, along with a new-fangled output we’re adding — one that many of you will have soon!

  Homework for backers

  We’ve updated the Spool site with your survey answers — you can now check that your shipping address is correct and make changes. (This site is to handle logistics for Kickstarter backers only, for now. If you pre-ordered a Twine from our site, we already have everything we need from you.)

  Many backers haven’t yet signed up — if you didn’t get the registration email (check your spam folder), please email us at help@supermechanical.com with the email address you used for Kickstarter, and we’ll resend your registration link. This is important to make sure we have your shipping address and order correct!

  
  Multiple tasks are starting to come together, so you can expect more frequent updates soon. Developing a great consumer electronic product from the ground up is a ridiculous amount of work — David and I appreciate every supportive comment you’ve made. Thanks for your patience. Onward!

  — John

• 2012-03-15

  The State of Twine, and a Shipping Update.

  Hello Kickstarter backers,

  For this week we’ve put together our most comprehensive update so far—we call it the State of Twine.  This update covers all aspects of Twine: the electronics and hardware that make Twine tick, the rubbery yet firm case that protects those insides, the layers of the software that keep in constant touch with thousands of Twines and communicate with a multitude of other services (email, SMS, Facebook, Twitter, HTTP, and few surprises), and finally Spool, our web application that helps you set up your Twine and keeps you up to date on its status.

  As you can tell just by the length of the last sentence, there’s clearly a lot to do.  It’s not enough to merely get all of these pieces built and tested—we believe that having an incredible user experience is the core of what Twine is about. We won’t be finished until Twine is easy, fun, and has that little bit of magic in it that makes it a Supermechanical object.

  
  
  

  Lets start our whirlwind tour with Twine’s electronic entrails, the hardware.

  

  Our third full revision is complete, and David is running it through the ringer to make sure it meets all of our performance requirements.  Meanwhile, our trusty hardware engineer, Rehmi Post, is rolling all of our improvements into the design for almighty revision four.  We’re really excited about his work there because not only is revision four very close to the production version, Rehmi’s been working extra hard to make it more compact at the same time.  So far it’s looking really nice—keep up the good work, Rehmi!

  

  The fabulous Rehmi Post.

  Production is one area where we’ve been victims of our own success. You’ve made us so successful that the project has had to scale beyond our original plan. To meet the demand we need to build about 6,000 Twines, about 30 times the number we anticipated when we launched.

  When you build 6,000 of anything, interesting new challenges pop up.  For example, we’ve ordered 24,000 battery clips, 17,000 headphone connectors, 6,000 ARM processors, and more than $100,000 worth of wireless modules.  A few times, we’ve needed more of a particular component than the entire world has in stock.  We had to call up the factory and have them make more flash memory chips—just for you guys.  That’s crazy.  

  Sometimes it takes many weeks for the factory to make those new parts, and unfortunately that can wreak havoc on a schedule. Twine is completely made in the US (no small feat), but the supply chain is long and complicated. Even US suppliers have to get some components from overseas. And another obvious thing we don’t think about in the Information Age — parts typically take several weeks to get to the US by boat!  We paid more to have our parts air freighted.

  

  We’ve done our best to minimize those delays through careful component selection and by purchasing everything as early as possible.  Here’s a picture of our current stockpile in David’s apartment.  If you were to fire up all of the ARM processors in those boxes, their combined speed would be about 672 GHz!!!

  
  
  

  All right, enough hardware — on to the case.

  It’s probably safe to say that no part of the project has seen more changes than the case design.  We originally planned to cast the electronics by hand directly into a solid block of rubber.  When we though we were going to make 200 Twines (hah!), that would have worked fine.  However, to make 6,000 Twines there is only one real way, and this is injection molding.

  

  Moving to injection molding has cost John most of his sleep for the past few months.  It requires metal tooling that costs up to $40,000 and can’t really be modified after it is produced.  Also, due to the temperatures and pressures involved in the molding process, the electronics can no longer be cast inside a solid block.  We’ve spent many, many hours trying to design a stiff protective body that still meets our goal of being soft and friendly to the touch, within the constraints of automated processes. Sounds like a contradiction, right?  We got a whole education here very quickly, both from talking to manufacturing partners and, as you can see above, from making many more prototypes.

  

  The approach we’ve ended up with has a strong ABS shell that provides mechanical support for the batteries, protects the PCB, and provides rigidity to the overall form.  Over this inner skeleton, we slip a soft Santoprene skin that dampens any shocks and gives the outside of Twine a great rubbery feel.  We can’t cast the final material by hand in our prototypes, so we’re using a urethane rubber that shares several qualities with Kraft American cheese.  Rest assured, once we’re happy with the dimensions and structural properties, we’ll re-add the texture, and raised surface features and things will start looking pretty stellar again.

  
  
  

  Software

  Twine has quite a few layers of software.  There are no less than three ARM processors on the Twine itself, and the code for each of them is a world unto itself.  This has resulted in a dramatic reduction in the amount of hair that David has left.  Well, at least he’s been writing good code.

  

  Peter, pictured above, is our lead cryptographer and the big gun that we bring in whenever we get really stuck on a coding problem.  He’s also the only eighteen-year-old we know who was accepted to graduate school at MIT.

  On the web side, we’ve launched the Spool portal.  If you’re a backer, you should have gotten an email with a special link that allows you to access the site.  Please help us test it at large scale by logging in to create your account. (If you were quick on the draw and got the maintenance page, try again now.)  Once you log in, you can verify that we have your Kickstarter order correct, and you’ll be able to track your individual Twine as it goes through the manufacturing process and finds its way home to you.

  

  Jay is a master web craftsman who’s been critical in getting the Spool web app up to speed. You’ve probably used some of the sites he’s helped build, including those of U-Haul and the Federal Communications Commission.

  
  
  

  Getting Twine in your hands

  When we sat down to write this update, we wanted to bring all of our backers up to speed on where we are with Twine.  We hope it’s evident how incredibly much work has gone into Twine so far, and that we are very close to a working Twine platform.  It really is no joke to say that we work on Twine, eat, sleep, and not much else.

  Nonetheless, there are still some significant challenges left out there, mostly in the domain of supply and manufacturing, and also in quality assurance.  We’re starting to sound like a broken record, but it is so important to us that Twine works as well as it should.  We won’t ship an unfinished product—it would only be expensive and frustrating for everyone involved.  We’re really sorry to say that we won’t be shipping in March. The scale on which you guys made Twine successful, ironically has made that goal impossible to meet.  Instead, we have taken a step back, evaluated where we are, rubbed our crystal ball, added a bit of margin, and created our best estimate for a new ship date.  That date is May 31st.

  Between now and then, we will continue to hold our noses to the grindstone, and step by step, conquer each challenge that lays in our path.  We will, of course, continue with our regular updates to keep you in the loop to share all of our victories and setbacks along the way.  Since January, Twine has come a very long way.  We’re excited to share the rest of that journey with you, so stay tuned.

  —David, John, and the entire Twine team

• 2012-03-02

  Wireless signals, new hardware, and rocket fins!

  Hello Twine backers,

  We’ve got quite a collection of new things for you in our latest Twine update.  To start it off, we did a bit of wireless range testing to optimize the wireless antenna position and case enclosure material.  (RF design and material compatibility can be tricky — think iPhone “Antennagate”.)

  

  A wireless test board in a prototype enclosure, another naked one, and our trusty Linksys E1000 access point.

  I put the access point on my kitchen counter and then ventured out the door and down the hall while my trusty assistant monitored the signal strength.

  

  The 2.4GHz spectrum in my apartment complex is a veritable traffic jam, with tens of wireless networks clamoring to talk over each other.  In spite of this, both test boards managed a range of more than 90 feet.  This is pretty impressive given the number of walls the signal had to penetrate and the large amount of interference.  I imagine that things would be even better in a quieter or less obstructed environment.

  Next up, we have a brand new hardware revision:

  

  This revision (number 2.35) is getting pretty close to the production hardware.  We’ll change the PCBs to a nice color and probably move the connectors around to better fit the case, but it feels really good to see the hardware stabilizing.

  The case design, however, has seen quite a few changes.  When we started the project, we planned to encase each Twine in a solid urethane block by hand.  Then about 4000 of you ordered Twines and that plan went completely out the window.  Since then, we’ve redesigned the case about four times trying to get the feel right while still being compatible with mass production.   We want it to feel soft to the touch and yet still be sturdy enough to protect the innards.  This combination has been quite a challenge to get right, but I think we’re getting pretty close with the latest design.  (Don’t worry — the one above is an old design and the newer ones are much prettier!)

  And then, cables.  Finding a nice four-conductor headphone-style cable has been quite a challenge, but I think we’ve finally got it licked.  The white cables above are the first samples of our custom super-flexible Twine cable.  We like them a lot.  Many thanks to Heather and the entire crew out at Tensility!

  You might remember that a while back, we put up a poll to figure out which wireless routers our backers have.  Over two thousand of you guys responded, and we read every single one of them.  Well, actually John’s brother Sanjay read every single one of them.

  Of the routers you submitted, we’ve tried to assemble as broad of a cross-section as possible.  Please don’t worry if your router isn’t on this list.  It will still work.

  Our formidable test array currently includes:

  • Actiontec MI424WR
  • Apple Airport Express and Airport Extreme
  • Belkin N600
  • D-Link DIR-655
  • Linksys E1000
  • Linksys WRT-54GL
  • Linksys WRT-54GS2
  • Netgear WNDR3400
  • Novatel MiFi 2200 Mobile Hotspot (cellular)
  • A super-cheapo router with no name that we found in a drawer

  I think you’ll agree that’s a pretty good list.

  However it could never be complete without the router that most resembles a 1957 Chevrolet Bel Air:

  

  

  The Fritz!Box.  (Special thanks to the guys at AVM for sending us one!)

  Finally, we’ve been getting some questions about delivery dates for Twine.  This week we’ve put a lot of effort into finalizing our supply chain and we should have some good information about ship dates in our next update. Stay tuned!

  -David

• 2012-02-17

  Crafting chocolate and cases

  

  We’ve made a lot of progress in development in the last few weeks. (More on that in a bit.) We also crammed in a visit to our friends at Dandelion Chocolate. Since we’re all about making stuff these days, I want to give you a peek at how they make stuff, and how they’ll use Twine.

  Todd and Alice were nice enough to show us their new bean-to-bar chocolate factory, one of only a dozen places in the US that make chocolate from beginning to end. Dandelion travels around the world to find the right cacao beans, then brings them back to San Francisco to turn into fine chocolate, adding only sugar. We tasted it at each step, and it is magic to get from that bitter bean to a uniquely tasty bar. Like us, they’re really small, and have improvised a lot of their processes. As you can imagine, the machinery needed to do this is very specialized (ever heard of a bean guillotine?) and almost as hard to get as Oompa Loompas. So they started with roasters made for coffee beans, and a grinder powered by a hand drill.

  

  Because they make small quantities with personal attention, Dandelion wants to use data to keep the quality consistent and learn from each batch of chocolate. Even the industrial machines they have don’t accurately measure critical things like temperature and viscosity. So they plan on using Twines with custom high-temperature probes to log all data into a computer, where they can not only monitor the progress (some steps can take a day), but also go back months later and learn what made that batch so good. This is the sort of thing that we would’ve never imagined Twine being used for, and we’re gratified that it’s helpful.

  

  Thanks to Dandelion for the tour and tastes! It’s amazing to realize how many steps there are to making our chocolate — finding, sorting, shelling, roasting, grinding, more grinding, mixing, tempering, molding and cooling — and seeing it all being done by one small company reminds me to savor each piece.

  
  We’re also grinding away. Assembling the first pre-production boards, stringing together the software all the way up to the web interface, and iterating through case prototypes. We’ve got some extra hands now but it’s a lot to do, and do well — especially when we got 15 times the response we were expecting! As other Kickstarters have found before us, that requires an adjustment in plans.

  For example, instead of casting each case by hand, we’re now going to produce them using injection molding. The design of an injection molded part is a complex process that requires many rounds of iteration to get all of the details exactly right. Injection molding requires very expensive steel molds with several interlocking pieces that must come apart cleanly and automatically to form a case that meets our quality standards. Before we commit tens of thousands of dollars to that, we’ve had to prototype the process. Here’s a peek:

  

  Sucking bubbles out of urethane using a vacuum pot. 

  

  Pouring the urethane into a syringe to inject into the prototype mold (in the clamp).

  

  A prototype curing in a 3D-printed acrylic mold, next to an earlier cast prototype.

  But these are good problems to have. Your amazing support is giving us the opportunity to make Twine much more than we had planned, and David and I are working double shifts to make sure you’ll love it. This project is our lives now (apologies to our significant others). We’ll be back soon with an update to cover some of the improvements.

  — John

• 2012-02-03

  Shirts, Shoes, Service

  Hello! We’ve been cranking on all cylinders, including some we didn’t know we had. Iterating on circuit board design, case design, firmware (the code that runs on Twine) and several layers of web software, while lining up production of about 20 times the number of Twines we thought we’d be making when we began this Kickstarter. That scale has also made it feasible for us to add a few new things based on your feedback — but we’ll keep them a surprise for now.

  Shirts

  Our first backer rewards, the limited-edition T-shirts for $30- and $500-level backers, are in transit. This was a whole mini-labor of love in itself — let’s give credit where it’s due. First, of course, to our generous supporters at these levels who helped make the project a success.

  

  Our good friend and Media Lab cohort Richard The meticulously hand-Illustratored a uniquely awesome design for you. Know that you are wearing aesthetic gold.

  

  Then our long-time buddies at Acme Prints cranked out stacks of top-notch shirts in an unreasonably short time. We aspire to their level of customer service — order from them to see what I mean. Acme has made pretty much all of the band shirts I wear as well as the Internet Brand T-shirt I model in our Kickstarter video.

  

  The unenviable task of fulfilling orders went to Supermechanical’s official little brother/machine, Sanjay Kestner. He operates The Lenin Closet, where you can get said Internet Brand T-shirt.

  If you’re wondering “Is that a toaster?” let us tell you why it’s our spirit animal of sorts. Henry, our research advisor, has long worked on connecting the objects in our lives to the Internet. As a thought exercise, he likes to ask people what they would do with an Internet-connected toaster. Silly, but some of the best ideas might change how we use computers — and Twine is the tool we made to help discover those ideas. When toasters can compute, computers will be toast.

  -John for Supermechanical

• 2012-01-24

  An enlightening Twine update.

  Breakout boards are simply misunderstood.

  We’ve heard all sorts of ideas about what they might do, like let you interface Twine with all sorts of other devices, connect every possible sensor to Twine at once, end football huddles, and even make possible Twime travel.  (Sorry.)

  If you guessed A: let you interface Twine with all sorts of other devices, give yourself a gold star.  That is indeed what breakout boards are for.

  This is what one looks like:

  

  Breakout boards have nice lever actuated connections for an input, power, and ground.  These connections let you easily plug all sorts of outside things into your Twine.

  They’re also pretty smart—-they have a built in identity so that Twine can remember your creation the next time you plug it in, and they also will protect your Twine against backwards connections, shorts, and most other things that could damage it.

  To give you a simple example of what’s possible, let’s use the breakout board to build a light sensor.

  

  Here are the parts we’re going to use, a light sensor, a resistor, and the breakout board. The light sensor and resistor are pretty generic items that you can get from a variety of electronics-type sources.  If you’re following along at home, I got these from Digi-Key (digikey.com).  The light sensor is part number 475-1081-ND and the resistor is part number CF14JT470RCT-ND.  You should be able to get the pair for less than a dollar.

  Let’s get to it!

  

  First you’ll notice that the light sensor has leads with two different lengths.  Stick the shorter one in the GND port, and the longer one into the IN port.  Go ahead and push down on the GND lever to keep the sensor from falling out.

  

  The two leads on the resistor are exactly the same.  Plug one into the +3.3V (power) port and the other into the IN port.  Lock down the other two levers and we’re done.  Super easy!

  

  Here’s your new light sensor.  Now your Twine can measure ambient light in your space, or maybe use an extra laser pointer to rig up a tweeting laser trip wire!

  

  Lots of things are possible using the breakout board, and we’re looking forward to sharing many more neat “recipes” in the near future.

  -David for Supermechanical 

• 2011-11-22

  Twine: Twitter/SMS/email for your objects

  We have had our heads down working like mad (because we are mad) on our new project for months, and now we’re ready to announce it.

  Twine is the simplest possible way to get your objects connected to the Web. It’s a 2.5” square that uses internal and external sensors to tell you about its surroundings via Twitter, SMS or email. It’s tightly paired with a web app that makes it easy to set rules that control Twine. This was spurred by our own acute need in our work at the MIT Media Lab - we wanted to build new kinds of interactions between the physical world and people, often remote, but would end up spending too much time on networking code and electronics instead. While we’ve designed something that non-engineers can use, it’s still geek-friendly, allowing you to hook in your own hardware inputs and software outputs.

  Please support us on Kickstarter with a pledge, reblog, or tweet. The more funds we get, the quicker we’ll be able to add features to Twine. Your rewards await you in post-PC heaven.

  Source: kickstarter.com

• 2011-10-19

  We outsourced the future.

  …product and process innovation are intertwined. So the decline of manufacturing in a region sets off a chain reaction.

  — Gary Pisano and Willy Shih, “Restoring American Competitiveness”

  As manufacturing moves from the U.S. to other countries, we miss the jobs lost. But we also lose two things that continue to hurt us in the future: expertise, and the ability to create significant new products.

  When we crystallize our manufacturing processes and ship them to the factory Over There, we lose the ability to improve them ourselves. This isn’t felt immediately, but outsourcing means you’re paying someone else to get better at it, and then they can use that expertise for their own benefit when new opportunities come along.

  But you’ll just move on to the Next Big Thing, right? Unfortunately, you’re at a disadvantage there, because even new product categories are built on expertise in existing ones. Pisano and Shih offer the example of the rechargeable battery market, which is swelling to supply the most expensive component to tomorrow’s electric cars. U.S. companies can’t compete here because battery technology has grown by leaps and bounds in Asia, ever since we farmed it out along with the rest of our consumer electronics in previous decades.

  It is now apparent that these firms have not merely outsourced production. They have actually outsourced their future.

  — Steve Denning, “The End Of The Road For Outsourcing?”

  Read Denning’s series at Forbes. He makes a powerful case that it’s just better management to consider: Are you cutting costs now, or expertise later?

  Supermechanical is using American manufacturing to build not only our products but also home-grown expertise. We believe that this investment leads to innovation through original, solid products at good prices.