Note that I am saying "as an idea." The Hemingwrite exists as a decent-looking prototype, and its creators seem truly dedicated to building the highest-quality final product they can manage. But I'm not sure it's possible to create the thing they promise: a complex, highly ecosystem-dependent device that "lasts generations." The most immediate point of comparison, the Model M, works because it has no mind of its own; it's a peripheral you can map onto anything that accepts keystrokes. No matter how good the hardware is, or how replaceable the battery, it's going to be tricky to base a device around cloud syncing and then make it survive the passage of different wireless standards and word processing software, especially if it somehow outlives the company that builds it. I'm not sure how much I'm willing to spend on a more convenient version of a half-dozen writing implements I already own. And I'd honestly feel a little embarrassed answering somebody's innocent "what's that?" with "Oh, it's a Hemingwrite."
The Hemingwrite recently made the slate of semi-finalists for Engadget's Insert Coin inventor's competition, and it will be appearing in my home city of New York on November 7th. Looking at it would probably be like staring into a dark mirror.