yeah, i've got an idea to use a minimalistic lattice FPGA to run a monitor that gathers the data, and then get an offline machine running a reasonable (eg 32gb) LLM to analyse the logs. all low bandwidth packet protocols ie dns, icmp, igmp, and log all ip addresses and compare them against a traffic logger capturing the data that the OS on my pc can see, the differential and the patterns would likely help discover anything creepy going on. most likely these things are not often used but i'd rather not be unaware if i could find out. i'm planning to do work with FPGAs anyway in the future for developing custom processor logic so it'd just be one set aside for this task. it probably would be fairly simple to whitelist address ranges from a known clean OS without any telemetry crap allowed through to act as a filter use on a machine specifically dedicated to secure communication.
super creepy stuff, and super lame that so little is mentioned about it. even the way this video doesn't talk very loudly about what it does in the name or info section. the presenter seems more concerned about the license on minix than the blatantly obvious privacy implications.
Is ME really a concern? Maybe. Is it designed, so it can force the owner to do something? Maybe.
Was it used on a large scale to force users to behave in some way? No.
What was used like that? Apple locked bootloaders in iphones.
Even if IME has flaws or backdoors, it's not used on a large scale. Network traffic would be visible immediately.
It is a problem but way less intrusive than fully locked bootloaders. And as far as I predict, it won't probably be used against people.
