Signstar design concepts - questions

Hi, reading through https://signstar.archlinux.page/architecture/design.html can someone explain how the SSS shares are distributed among holders? Do they physically meet for a beer and each takes out a print out from initial console setup? Could you perhaps pre-bake share holder's SSH keys into the image and do a one share per key as the do first login? This way they could retrieve it remotely. My second question is if confidential compute was considered instead of a dedicated physical hardware and OS? Could you replace the setup with a confidential VM that runs signstar as verifiable service?