Goffredo Marocchi@marcoarment for your fleet of Mac Minis our EIT department swears by JAMF https://www.jamf.com Apple actually did a lot of work and JAMF even more for enterprise setups including the updating features, shared configuration deployment, etc… want me to put you in contact with a local JAMF expert I know? He is mega busy, but he likes fun project so maybe he can make some time.
@marcoarment also, with 48 Mac Minis with 16 GB of RAM each, you might be a bit limited but you should be able to crawl through this with local models and LMStudio and the like https://lmstudio.ai/docs/integrations/claude-code maybe you could use the full big bad frontier LLMs like Opus to prototype the work and build scripts or prompts for a more limited LLM to use and the throw your swarm at it. Eventually they will finish :).
