Howstr is the startup I’m leaving the Air Force to found. It’s like Google for how-to knowledge. Howstr makes complicated projects simple so that people can build the same stuff when they aren’t in the same place. This is the story of where the idea came from, where it’s at now, and where it’s going.
About five years ago I stumbled across 3D printers during a serendipitous youtube search 1. Of course, I had to have my own 3D printer, so I bought a Thing-O-Matic from Makerbot.
The TOM was amazing, but soon I needed more so I made the parts for a RepRap Mendel. That was my introduction to open source hardware. The RepRap was several times better than the Makerbot because it benefited from the input of an order of magnitude more people.
After that I tried to participate in bigger open hardware projects like Wikispeed (cars) and OSE (farm equipment). Turns out it’s basically impossible to make open source principles work for big hardware projects. The volume and detail and update frequency of information necessary to make a copy of a car is an insurmountable burden. Imagine open source software on punch cards.
So I wrestled with the structure and process necessary to unleash the pent-up potential of open hardware. About a year ago I was confident that my abstract understanding would work, so I started learning to code. About six months ago I was satisfied that my prototype 2 actually did the thing it was supposed to do. About a month ago my friend Jeremy was satisfied that my beta version performed as expected and more. As I’ve developed the idea I’ve become convinced that it’s a fundamentally better approach to a lot of the things people are doing. It’s sort of the culmination of a feeling I’ve had my whole life that there has to be a better way.
Peter Thiel says that “secrets” are important discriminators; something you know is true that nobody else knows is true. Well, what I’ve learned so far is that writing is holding us back. Did you know that writing is stone-age technology 3 ? Writing only has one feature; it’s a list of symbols that hopefully mean the same thing to more than one person. The printing press and the internet merely allow us to share the lists of symbols better.
The reason writing is its own profession is that it’s really hard to take knowledge in your head, serialize it, and have someone else understand it, let alone a large audience with diverse perspectives. However, the real limitation is that it requires substantial effort on the part of every single reader to parse the symbols every single time. That’s why open source project documentation is so horrible. Nobody wants to write or read, everybody just wants to build.
If words (1 dimension) are worth 1 unit each, and pictures (2 dimensions) are worth 1,000 words, then models (3 dimensions) must be worth 1,000,000 words. So it stands to reason we get a 1K multiplier every time we integrate another dimension into the knowledge representation. That’s why so many of the most expensive activities rely on 3D files (x1,000,000), or at least multi-layer maps (+1,000). Well, the backend of Howstr is called NotionAll and it can represent n-dimensional conceptual space. I’m using two dimensions at the moment, but there are obvious ways to integrate a third, and we’ve brainstormed out to a dozen dimensions. The point is that there’s no limit and all of the dimensions (can) occupy the same relational coordinate system. So Howstr can scale to handle the complexity inherent in tangible projects.
What’s 2+2? The answer popped into your head without effort because it’s intuitive. Given a couple simple inputs, the resulting output is trivial. That’s what Howstr does. It just strings together one intuitive action after another, presenting exactly the amount of information necessary to inspire a correct intuitive action in a person. Since Howstr is “parsing” an n-dimensional model it can adjust the display to exactly what each individual person needs at any given moment. So, for the user, it’s like having MacGyver feeding them the next step. It doesn’t require effort to consume and it requires only minimal effort to create, effort which can be spread out across many different people. Eventually Howstr will be able to build its own models by watching a person do their thing.
My working vision statement for Howstr is “once and for all.” I want to eliminate the phrase “reinventing the wheel”. Also, there are a lot of downright offensive examples of people profiting by withholding information. I’ve been keeping an informal list of medical devices that range from trivially easy to replicate, to modestly complicated, and which cost several orders of magnitude more than they should. Howstr will provide the substrate for a network of knowledge that everyone can contribute to and utilize. Explain it once, and everyone will know how to do it forever. I actually want to measure Howstr’s impact by estimating “dollarn’ts.” As in, dollars that never had to be exchanged because people can satisfy their own needs, or satisfy each other’s needs more efficiently. I believe we can create abundance by pushing costs down to only what the laws of physics demand.
- It was “3 axis linear motion” because I wanted to build a robotic table that could follow you around carrying all your stuff. ↩
- It was called Platypus because it stuck together whatever random project information you threw at it. ↩
- The oldest writing discovered is around 5-6000 BC, right around the end of the Stone Age. ↩
Be First to Comment