Looks fantastic, congrats on the launch. Using git for versioning and PCB development has been nearly unworkable. So nice to see a visually-oriented solution.
Vulkan (MoltenVK) does seem to be a valid option for those heavily invested or interested in porting existing OpenGL projects (I plan on evaluating it):
https://github.com/KhronosGroup/MoltenVK
Great concept. Would like to see more of this idea applied to neural network processing and configuration in general (which in my experience can sometimes be a tedious, hard-coded affair).
Great idea, however would be great to see an option which used the dataset but did not upload your data.
I've read the "FAQ Why must IoT Inspector upload the data to Princeton?" and do understand the reasons for data collection, maybe once it reaches a data saturation point this new option could be introduced (I believe there is a “Start/Pause Inspection” button but that seems to only temporarily pause collection).
Naturally, there's going to be some reluctance in using and deploying a tool that uploads data to an outside party.