Stone & Megan
Sydney · RN + ECT
What AI actually does in our jobs.
Two real professionals — a nurse and an early-childhood teacher — writing about the AI tools we use on shift. Including the ones we dropped. Including the failures.
-
The Day I Stopped Pasting Whole Files Into the AI Prompt
Should I paste the whole file into an AI coding assistant? I did for two weeks, the model invented functions, and pasting less code dropped hallucinations to near zero.
Read -
Aura Week 3: The Bug I Couldn't Even Describe to the AI
Voice transcription was 60% accurate in dev and 92% in prod. Same code. Three days of describing it to an AI taught me when AI can't help you debug.
Read -
What I Throw Away After a Week of AI-Assisted Coding
I killed four AI-built features in one week. The pattern that decides when to throw away AI-generated code, and the 30-minute spec note that now prevents the rewrite.
Read -
OWNA Templates: Using AI Alongside, Not Inside
How I use AI with OWNA templates without breaking them: a Sydney ECT's draft-outside, hand-condense workflow after two paste-induced sync errors in one month.
Read -
The Observation an AI Could Not Help Me Write
Can AI write childcare observations? The day a two-year-old broke six weeks of silence, the AI mis-tagged the whole story, and why I wrote it by hand instead.
Read -
Using AI for Parent Comms When Families Don't Read English
How I use AI for parent communication when families don't read English: a Sydney ECT's plain-language workflow, the words AI gets wrong, and what I never automate.
Read -
The clinical question I will not ask an AI (and why)
A patient's grapefruit question taught me why you shouldn't ask AI about medication interactions. The one rule a Sydney RN now follows on every shift.
Read -
Why I Draft Patient SMS Reminders by Hand, Not With AI
A Sydney RN on why I draft patient SMS reminders by hand, not with AI: voice consistency, disclosure, privacy, and the six-template library that cut my error rate.
Read -
When my content agent invented a fake author persona
A post-mortem on how my own AI content agent invented a fake author persona — a credentialed nurse byline — on a post about my work, and the two guards that stopped it.
Read -
The first time the AI coding tool lied to me about an API
An AI coding tool hallucinated a method that doesn't exist while I built Aura. It cost me 45 minutes. Here's the exact bug and the rule it taught me.
Read -
Why my 7am blog cron silently skipped for three days
A launchd job silently skipped daily for three mornings while reporting exit 0. The cause: a ThrottleInterval sitting too close to the 24h cadence with no safe margin.
Read -
Drafting EYLF Learning Stories With AI: Format I Keep
How I write EYLF learning stories with AI: the 5-block format I hand the model every time, and the child-voice lines I always rewrite by hand.
Read -
Five Curriculum Ideas From the AI — The One I Picked for Our Yard
I used AI for lesson planning in early childhood education to brainstorm five sustainability themes for 3-4s, then picked the one our yard could actually hold.
Read -
The 4 Things I Check Before Any AI Note Goes in the Record
A Sydney RN's 4-point routine for what to check before signing an AI-generated nursing note: identifiers, drug data, numbers and attribution — under NMBA accountability.
Read -
The Clinical Topic Where AI Was a Year Behind (and How I Caught It)
An AI handed me an outdated SABA-only asthma pathway. Is the AI safe to use for nursing? Here is the verification habit I now run on every AI clinical answer.
Read -
When Nurses Should Not Rely on AI: Cases That Need Hands, Not Prompts
A walk-in with central chest pain. I went to the ECG, not a prompt. When nurses should not rely on AI: the clinical layer where typing is the error.
Read -
Writing Better Patient Education Handouts With an AI Assistant
An AHPRA-registered RN shares the actual AI prompts I use to write patient education handouts at a 6th-grade reading level, plus the safety edits I add by hand.
Read -
Why I Stripped Every AI Brand Name Out of My Blog
I named one AI tool everywhere before launch, then deleted every mention. Whether you should name brand names in blog posts is a reader-trust call, not SEO.
Read -
Why I Killed My Nurse-Tooling Side Project Before Coding
An RN and indie dev on when to kill a side project before building: the privacy wall that ended my clinical-note tool before a single line of integration code.
Read -
Writing Child Observations With AI Without Losing Your Voice
I tried writing child observations with AI and deleted my first three. Here is the before-and-after, the guardrail prompt, and how I keep my teacher voice.
Read -
AI-assisted nursing notes: 9× faster, and one near-miss
An evening clinic. Twelve note types still to write. I built an AI-assisted tool to draft them. Then an iron-infusion record nearly went out incomplete.
Read -
AI wrote the parent newsletter — 4 prompts to sound like me
A Friday at 5:40pm. Blank newsletter template. Twenty minutes of blinking cursor. Four prompts later, the AI sounded like me — not a corporate chain.
Read -
Aura week 1: I killed my voice-first ECT app on day 14
Two weeks. 60 commits. One 4-hour bubble bug. Then I killed the project. What the AI coding tool got right, and what I should have checked first.
Read
Get a real post once a week
No filler. No hype. Just what we tried.
Drop your email and we'll send the new post the morning it goes up. Unsubscribe anytime.
Failures included. On purpose.
Most AI blogs sell you a workflow that worked. Ours show you the iron-infusion record that nearly shipped without its batch number, the four prompts before the AI stopped sounding corporate, and the day-14 project we killed.
Read the dev diary →