Boot Environments and You: An AI Primer

[sam-kleiner]

First off need to say thanks for one of my favorite pieces of software. It’s damn near perfect imo. It’s already saved my ass twice by making the rollbacks so easy when I broke stuff. Not only that, but it’s well documented and those docs take the time to teach about how the Linux boot process works so I’m not just copy and pasting commands without knowing wtf is going on.

So a year or so since I first tried zfs boot manager by copy and pasting commands without knowing wtf is going on, I finally got around to reading the Boot Environments and You: A Primer page. I’m comfortable using Linux day to day but haven’t ever taken the time to really learn the internals.

While reading I was confused on how the passphrase encryption worked. I threw it into Gemini (chat link below) to get an ELI5. I haven’t checked it for accuracy yet, but I figured this post would be a good place to start. If it’s accurate, I thought parts of its description might be good to add the Boot Primers and You page. Would love to see an FAQ with this kinda info that’s confusing when you first start learning about this.

If you like that idea comment your early confusions here.

The chat: https://g.co/gemini/share/8d116949e6a3

[classabbyamp]

we're not interested in contributions generated by so-called artificial intelligence

[sam-kleiner]

That was rude. Did you even read my post or the chat. I didn’t say take this and copy and paste it into the docs. I thought this was interesting and useful information to possibility add to an FAQ to help the community. Even if this chat ended up being factually incorrect and not included, the idea is still a good one. No AI isn’t great yet, but it’s really useful to summarize info. It’s a tool and you’re the old fart yelling that video killed the radio star.

Don’t be an asshole

[zdykstra]

I read the post. I'll happily engage with a human that's raised concerns or requested additional information on something. What I won't do is read through a tool regurgitating my own documentation back to me, and then figuring out if it got it subtly wrong. If there's something you'd like to see added - feel free to request it directly. But don't expect anybody here to dig through a chat log with an LLM for inspiration on how to improve our docs.

[sam-kleiner]

This is the discussions section of the repo, so I kinda thought I would start a discussion. If you don’t want community discussions disable them.

You write good software but community builders you are not.

[zdykstra]

Again - happy to have a discussion directly with you about how to safely interact with encrypted datasets and what their pitfalls are. If you come here with questions, you'll likely get answers. If you come here with a request for maintainers to review the output of an LLM / plagiarism machine for accuracy, you'll get told that we're not interested.

With all of that said, do you have a specific topic that we can help explain in greater detail? We have a long-standing PR for a basic FAQ, some of your questions might be a good fit for that general document!