Comment by mindslight

The only thing I've found that RMS is wrong about is defining away firmware-stored-in-flash as "hardware", while getting stuck on firmware blobs loaded into coprocessors indicting whole systems. It appears to be a handwaving distinction formed from previous times - when the distinction didn't really matter, a line had to be drawn somewhere, and being able to do things like declare an entire CDROM as "free software" was a cool goal in and of itself. But these days we could use a much more vibrant definition centered around processor domains and security relationships.

For example I see zero freedom difference between my ath10k with its firmware loaded from disk by libre software, and my x520 with firmware stored in onboard flash. Neither undermines the freedom or security of my workstation user domain, and both are unfree if I get the itch to dig into modifying my network cards.