Just Fucking Read the Docs

We’re living in a golden age of documentation. Markdown files, detailed wikis, man pages, vendor PDFs, config examples—all right there, one ctrl+f away.
And still, somehow, people would rather take a screenshot, toss it into ChatGPT, rub it like a fucking magic lamp and hope for the best.

Stop.

LLMs are not genies. They won’t grant your wish to fix that broken firewall rule, reverse-engineer your .yaml mess, or figure out why your SQL maintenance plan is silently wrecking everything. It’ll try, sure—but if you don’t understand what’s going on, it’s just guessing. With confidence.

Sysadmins, DevOps folks, Network admins—we’re supposed to be the ones who go a layer deeper. Who know the difference between a real fix and a Stack Overflow cargo cult.
And yet, I see more and more people praying to AI like it’s divine intervention, instead of—stay with me here—reading the fucking documentation.

Yes, some docs suck.
Sometimes on purpose—designed to nudge you into paying for training or chasing a certification.
Other times because they were written by someone who clearly never used the tool they’re writing about.
But c'est la vie. You still have to deal with it. Because guessing is not a solid strategy.

A professor once told me, "You're doing stochastic debugging"—just changing random things and hoping something works.
That phrase stuck with me. Because that's exactly what it feels like when someone pastes an error into ChatGPT and starts trying whatever sounds vaguely related.

I’m not even against the use of LLMs—in fact, this very text was structured and given sense by one of them.
But please, for the sanity of your teammates, for the holy sake of your own time:
just fucking read the docs.