Diary of an AI Trainer

Chronicles from the frontline of AI literacy education — the absurdities, the quiet revelations, and the daily gap between what the technology promises and what it actually does when someone is watching. Written from Pandanus Reach, these dispatches document what happens when you’re the person organisations call to explain the thing that changes faster than anyone can learn it: the live demos that fail on cue, the webinars that answer every question except the right one, the Word templates that defeat the most powerful language models on Earth, and the peculiar existential vertigo of occupying one of those “new jobs” the tech industry keeps promising will emerge. When Sam Altman says confidently that AI will create exciting new roles, you do sometimes wonder — does he mean this one? Part field notes, part comedy of errors, part genuine reckoning — this is what AI adoption looks like from the chair of the person doing the training.

The Thinking That Survives
A long-form essay on the slogan that promises to retrain everyone for the new jobs, the kind of learning that promise quietly assumes, and the four hard-to-train modes of thinking the cognitive-science evidence suggests will actually carry people across the gap.
Why I Care That You Don’t Know How to Use AI
On the silence that follows certain workshop questions, the angst that follows the silence, and an inconvenient truth about other people’s AI illiteracy the author has, until now, been polite enough not to articulate.
In the Beginning Was the Prompt
On the great schism between the Hedonists and the Luddites, the founding of two opposing canons, and the particular agony of the trainer who must serve communion to both congregations on Monday morning.
Attention Is All You Need
On being asked, again, for the one-hour version of a one-day workshop, and why the author is formally retiring as your gentle AI trainer.
What We Never Get To
In which two hours of policy discussion leave eleven minutes for the actual AI content, and the author spends a subsequent weekend creating an Excel training manual set in the Ozarks. This is a normal professional response.
The Refrigerator Principle
In which the author teaches AI literacy using an example involving leftover vegetables, and the crushing realisation that the workplace equivalent of “what’s in your fridge” is a SharePoint drive nobody has opened since 2022.
The Longest Road to Copilot
In which every AI workshop scheduled for ninety minutes requires a full day, every room contains at least four shadow AI users who don’t know their organisation’s data policy, and the feedback forms all request more hands-on time with the product we spent the morning explaining we couldn’t open yet.
So You’ve Become the AI Person
Ten things nobody warned you about becoming an AI literacy trainer, including performing competence in post-graduate mathematics, straddling the gap between AI cheerleader and existential doomer, and the specific psychological horror of demonstrating Copilot live.
The Template Wars
In which the most powerful language model on Earth is defeated by a Word document containing merged cells created by an MVIT administrator in 2017, and the author discovers that the gap between AI benchmarks and AI usefulness is approximately the width of a table border.
Schrödinger’s Copilot
In which a helpful AI assistant confidently offers services it cannot provide, performs feats it claims are impossible, and answers a yes-or-no question with the structural complexity of a Senate estimates hearing.
The AI Webinar Industrial Complex
In which the author attends forty-seven webinars on AI and academic integrity, learns that detection software doesn’t work, and waits in vain for someone to answer the only question that actually matters.