The Ten Commandments of AI Usage
A humorous and critical list of ten rules for using AI responsibly, warning against the dangers of over-reliance and the loss of independent thought. Each commandment addresses common pitfalls, from using code one doesn't understand to asking AI for questions to appear smart. The article ends with an ironic observation that the very people who need this advice might paste it into an AI for explanation.
Francesco Gadaleta
Jul 12, 2026
While the ambition is correct, this image is AI generated.
AIpidemic, a condition where the patient still has a brain, technically, but has outsourced its warranty to a chatbot.
Symptoms include: asking “what should I ask in this meeting to sound smart” before the meeting, and confusing “I pasted the stack trace into Claude” with “I understand the bug.”
Let’s write the scripture this plague deserves.
The Ten Commandments of AI Usage
I. Thou shalt not use AI-generated code you do not understand.
Because three months from now, “git blame” will point straight at you, and “idk, the robot wrote it” is not a valid entry in a postmortem.
II. Thou shalt trust thy brain. It is, in fact, still working.
It’s just been sitting on the bench so long it forgot it knows how to play. Put it back in the game before it atrophies into a decorative skull ornament.
III. Thou shalt not ask AI what questions to ask in order to appear knowledgeable.
This is the managerial equivalent of hiring a ghostwriter to text your therapist. If you need a script to sound competent, the problem isn’t your vocabulary.
IV. Thou shalt not paste a teammate’s code into a chatbot and call the printout “feedback.”
Code review is a relationship, not a relay race where you’re just the guy/girl holding the baton between two AIs. If your entire managerial contribution is Ctrl+V, HR should really be reviewing you.
V. Thou shalt not confuse fluency with correctness.
The model will explain your bug beautifully, with impeccable grammar, complete confidence, and total conviction — and still be wrong. Charisma is not a compiler.
VI. Thou shalt verify before thou shippest.
“The AI said it was fine” has never once held up in an incident report, a court of law, or a marriage.
VII. Thou shalt keep at least one opinion that did not originate from a prompt.
If your takes, your code reviews, and your birthday toast all have the same faint smell of RLHF (Reinforcement Learning with Human Feedback), it might be the time for an intervention to yourself. Or at least the time you touch a keyboard without assistance.
VIII. Thou shalt not let the machine make thy decisions of consequence.
Drafting an email? Fine, delegate away. Deciding whether to fire someone, sign a contract, or diagnose a patient or submitted code? That one’s still yours. Accountability doesn’t have an API.
IX. Thou shalt remember that “prompting well” is not the same skill as “knowing the domain.”
Being excellent at asking ChatGPT/Claude about kidneys/code does not make you a nephrologist/coder. It makes you someone who’s very good at asking ChatGPT about kidneys/code.
X. Thou shalt occasionally produce a thought, a line of code, or a decision with zero AI involvement. Just to prove thou still canst.
Call it a fire drill for your own cognition. If you fail it, that’s not a productivity gap. That’s the AIpidemic, and it’s already stage two.
The tragic irony, of course, is that half the people this list is aimed at will paste it into an AI and ask it to explain why it’s funny.
Amen.