Not the OP, but in the TechCrunch Disrupt launch, founder Brynn Putnam says, "capacitive material manufactured into the pieces."
If you put capacitive material in a unique pattern on the footprint of each piece, and the rest of the piece material was conductive enough to carry your body's charge to register a touch, the shape of that touch could be unique per-piece.
There's no mention of syncing pieces, charging pieces, keeping pieces in view of a wide-angle camera, anything like that, so that's my bet. (This would also mean moving a piece using a non-conductive material would be a way to cheat by having it not get registered!)
I just shared this on LI this morning, linking back to a video showing showing related touchscreen explorations I did for a colleague in early 2013, sensing different coins by their radii as you touch them: https://www.linkedin.com/posts/vmiliano_a-vertical-triptych-...
This is nearly bang on correct. The pieces don't contain any electronics or sensors, they have a conductive pattern built into the surface using specialized materials and a manufacturing technique we developed in house. Our custom software stack processes the raw data from the device's touch sensor using embedded ML on the NPU, which detects and tracks the pieces in real time.
That said, the device can detect the pieces whether you touch them or not. Touching them absolutely does change the response, and we pass that along as a parameter to the SDK.
Your coin exploration is seriously cool, please hit me up when you're next in NYC!
> they have a conductive pattern built into the surface using specialized materials and a manufacturing technique we developed in house.
Would this be something a home 3D printer could do? I'm not a maker but I could see the value of others being able to quickly build a universe of playing pieces if that was possible.
It's possible to make your own pieces with a multi-material 3d printer (our early prototypes have been made with Bambu X1C & H2D printers), though it's pretty finicky to do so, and requires some rather expensive filament. Happy to help anyone along though!
This was my immediate thought. I have a Bamboo X1C; buying special conductive filament is fine if you're just making a few pieces.
Board is MUCH more interesting to me if I can easily program it for my own games, especially if I can share them with others. I'm thinking about TTRPGs and other army building games where each piece has unique capabilities and rules.
It would be nice if I could buy smart bases, kind of like RFID stickers. Something I can glue to the base of an existing miniature with a 10-20mm round/square/hex base.
With those tools, I would build deep accounting assistance for complex games. If it's cheap, you could put custom terrain pieces on the Board with smart properties (hard cover, soft cover, blocks line of sight, etc).
Obviously there are lots of partnership opportunities there. You could get the table top game publishers to publish Board editions of their games that automatically keep up to date with the latest rules.
If that all works, the obvious next limitation would be board size. a 24" screen can do a lot, but many games would benefit from a screen measured in feet. That would be expensive and hard to move around though, given the current thickness.
Perhaps an intermediate solution would be support for tiling Boards. Everyone brings their own, you shove them together and have a much larger playing surface. Bezel width would be annoying here.
It might be easier to 3d print normally, then separately 2d print the capacitative layer, then stick it on the bottom of the 3d printed piece: https://youtu.be/ON-6bdhQHpI
It would sound more logical to me to buy a stack of pre-made patterns (e.g. coin or cube form-factor) and glue them into a like-shaped slot in a 3d-printed playing piece. Assuming that is possible, and you'd still have to make a conductive path to the person touching the piece, but this would be much easier than printing the pattern yourself.
It is not a requirement to provide a conductive path to the person though. The patterns (glyphs as we call them) are detected and tracked regardless of whether they are being touched. However, when there is a conductive path to the person, the system detects that which provides another input vector.
This screams for compatibility with 3d printers. eg: design a piece to be "absorbed" by a LEGO brick (2x4, duh!), and design a capacitive pathway for "two buttons" a-la: https://a.co/d/f7wm3GA
3D print your goblin army, snap it to the base, touch the sword arm to attack, the shield arm to defend, etc. light up the base via capacitive to 0/1/2 inputs and you're set!
Reminds me of some ~10 years ago I found a dirt cheap game in a budget bin in some store. It contained a few plastic, pre-painted, miniatures, that had some kind of special base with bumps on (possibly in some unique pattern for each miniature?) and then you were supposed to install an app game on a tablet and place the miniatures on the tablet to move them and have their locations (badly) tracked. Seems like it may have been a similar technology, but you only had to pay for the playing tokens and use a standard tablet. Not sure if special hardware is needed for better tracking of enough tokens or if a standard tablet today could be used with the right software?
Game(s) that you were supposed to play was not very fun, in addition to the tracking not working well, as I remember it, but I may still have the miniatures somewhere. There was another game from the same company that I also bought at the same time, but that one was made to be played with a phone camera as some kind of AR game instead, moving some plastic objects on a table, that also worked about as well as the other game.
All of vintage computer emulators on the Internet Archive owe a tip of the hat to the Canon Cat emulation, because that was the computer that started me on automating the original cross-compilation infrastructure.
I ended up using Shortcuts to publish a set of screenshots from an always-on Mac every 10 minutes or so for a set of shared, family AirTags: http://vitor.io/android-airtags
I couldn't reliably get iOS to do the same, and my tests of an Android-compatible tag found it was detected much less often than an AirTag riding along with it.
Wow, that's some dedication to solve this problem! Very nice write-up. It was also good to see your thinking behind why doing this on an always-on Mac was better than possible alternatives. Thankfully M-series Macs don't draw much power at all at idle.
Nice, I'll link that later as well - the gsettings trick would have been great the first time I installed (trying to rotate the screen via tab and arrow keys was painful!)
I like your point about the anchoring effect of the GPD price. I think that influenced me exactly as you say. It's a shame about the few rough edges; this is so nearly a very compelling gadget!
I don't use it as a daily driver any more (I did for a bit while traveling, for obvious reasons), but it's been very useful as bench computer. Doesn't take up as much room as a laptop on a desk, has both a USB 2.0 and a USB 3.0 port for serial consoles and data transfers, is proper x86 so I don't have issues with running proprietary flashers on it. I think it's running 18.04 right now because I needed to flash an old Jetson board.
Bench machine makes sense. I have a very locked down MacBook for work and I don't much like macOS - a little Linux machine for my personal stuff that takes up almost no space on office days seems very appealing. I'll see how that goes!
34 diatoms can be browsed using the left and right arrows in the UI. The diatoms of Yellowstone Lake can also be viewed in a separate section by clicking the link in the lower right.
There's only a fixed amount of phone numbers, but you can gate the person behind a single number by an extension, just like businesses do.
Give out your regular number, and generate a unique extension for each business or individual.
Have an IVR, bot, or service answer the number. "Hi, you've reached blakeburch, if you have an extension number, please dial it now." Valid extensions get routed to you, invalid or expired ones go to voicemail.
Having my own extension system is interesting... I'll explore that avenue.
Minimal research says that even with restrictions on certain number combinations and area codes, there should be ~5 billion possible numbers, enough to make a virtual number system possible.
