Kevin

I would actually bring a parallel to the device driver-firmware blob split that’s common with hardware support in Linux. While the code needed to run inference with a model is straightforward and several open source versions exist already, the model itself is a bunch of tensors whose behaviour we don’t have any visibility into. Bias is less a problem of the inference code and more an issue with the data it was trained on

If I were to guess, it would be the additional pins. USB-C PD is capable of decent power transfer while also having enough data transfer capability simultaneously. USB-C docks are a good example, seeing that you can hook up a display, charger, other USB devices, ethernet, etc and have it all go through a single cable and (compact, convenient) connector. The reversibility is an added bonus