Artsci and his crew tried to decode the video signal, but it's not in any usable or standard format to allow recording or injection of some other standard video signal. Also, even if it was a standard signal, writing usable dashcam software would be a quite a large undertaking for this project, IMHO. Sure you could record the signal to memory card but that's still adding more hardware, more software and at a minimum you'd need to build in automatic file rotation, maybe motion detection (or parking mode), settings of some sort (date/time?), a user interface (even if it's just buttons), etc. I believe Artsci is simply taking the existing V1 PC board and getting it reproduced, so no chance of re-engineering it.
Also, both the front and rear cam are extreme fish-eye view lenses, which aren't great at capturing detail beyond a couple of feet. So in addition to writing dashcam software, the onboard code would need to optically correct that perspective into a usable image compared to a regular dashcam. Or people would have to convert the saved videos after-the-fact, reducing the utility of such a device.
And then someone has to support it and update it when people start requesting more features or there are bugs (note: add USB port or Bluetooth for firmware updates!). The front camera switch is just a dumb switch, that's set-it-and-forget-it.. and really isn't meant to be a dashcam, so trying to make it into one I think isn't a great idea. Nobody is making any money on these things and several people are putting in lots of personal time and R&D to get these products out to the community just because they are awesome products. It's easy to say "oh, just add recording" but in reality, that's a huge undertaking all by itself. Leave the dashcam software to the dashcam makers.