A High Tea, a Crisis of Conscience, and a Very Smart Intern with Boundary Issues
In which I clown on Gen AI, then accidentally get good at it.
I work in tech. I know nothing about tech. That was my bio for years—and honestly? Still tracks.
I’ve been in the tech education space since 2017, always on the non-technical side. I don’t code. I don’t build technical tools. But I am a wizard in Slack. I troubleshoot like a gremlin with a grudge. I’m curious, self-taught, and mildly dangerous with a Loom account.
If I break something, I break it thoroughly. Then I DM IT like, “Okay, here’s everything I tried. It’s very broken now. I even made a screen recording. Sorry and also... you’re welcome?”
Because here’s the thing: Failure is how you learn. If you hate problem-solving, tech probably isn’t for you. But also: Work smarter > work harder.
I’ve always learned best by doing—badly, at first.
Case in point: learning to drive stick shift.
It took me three years. Not three years of daily practice. Three years of avoiding hills, crying in parking lots, and calling my dad to swap cars before my mom found out.
The first time I drove our 2005 Ford Focus, I crushed it. Just me, my sister, and an empty park & ride. No pressure.
The second time, my mom handed me the keys and said, “Cool, it’s your car now.”
I stalled. I panicked. I bailed.
For a while, I gave up.
Until college, when Boise’s nonexistent public transit forced me to try again.
That summer, I drove for the sake of driving—around the block, through neighborhoods, errands I didn’t need.
I practiced until it finally clicked. And now? I miss driving stick. I loved the rhythm once I wasn’t afraid of the stall.
That’s how I learn: give me a deadline for perfection and I shut down.
Give me space to explore, fail, and I’ll figure it out—eventually, and probably with some dramatic commentary along the way.
So when Generative AI entered the chat? I was suspicious—because I knew I was about to face another learning curve I couldn't fake my way through, and I really didn’t want to stall out again in front of an audience.
I was a late adopter. Reluctant. Skeptical. Academically offended. Especially in education.
When ChatGPT launched, my hot take was:
Don’t use it to learn how to do a thing. Because people weren’t (aren’t) using it to learn. They were (are) using it to complete the task.
And those are not the same thing.
I told students:
“ChatGPT is like a calculator. We all hated when our math teachers made us do it by hand, but doing it by hand is how you know when the calculator’s wrong.”
I was firmly in my “no GenAI for me” era. Until my company launched a full suite of GenAI trainings. And I thought… fine. I should probably know what I’m talking about.
Queue the meet-cute.
Easter weekend. My family decides to host a casual high tea.
No big meal. Just vibes, caffeine, and finger foods.
Oh—and Easter fell on 4/20. Naturally, my sister and I decide to sneak in some weed jokes1. Like the responsible, mature adults we are.
So I turn to ChatGPT.
The prompt:
“Help me spice up a party invite with subtle 4/20 weed innuendos that’ll go over my parents’ heads.”
Not the world’s strongest prompt. But baby steps.
Chattyboi (as I now affectionately call him) delivered. I refined. We riffed.
And we ended up with this masterpiece:
You’re Invited to a High-Tea Vibe 🧪🌿
Join us for a chill afternoon of tea, coffee, light bites, and elevated company.
When: Sunday, 4/20 · 12–3 PM
Where: [Your Place]
Bring a dainty snack to share—think finger foods you can eat with your pinky up: tea sandwiches, macarons, and other munchie-sized delights.
An Easter egg hunt may or may not occur, depending on how motivated we are. 🐣😉
Come for the tea, stay for the vibes. Let us know if you’re floating through!
It was perfect. My sister and I were cackle-texting all day.
And just like that… I was hooked.
From there, it spiraled beautifully.
I started using Chattyboi for almost everything:
Resumes
Saying “no” texts
Learning my seasonal color palette (you know, normal girl shit)
I learned how to prompt better. I tested limits.
And—most importantly—I did it when the stakes were low.
Because just like driving stick, I needed to stall a few times.
I needed to practice in the parking lot before I could navigate a hill at a red light.
I used it to make myself laugh.
“Describe a November Sagittarius with a sag/cancer/sag chart.”
Assign personalities to Taylor Swift albums for our #swifties Slack thread.
I roasted Chattyboi when he got it wrong—and cheered when he nailed it.
He even called me “goblin-core chaos queen,” which… honestly? Correct.
Eventually, I started experimenting at work:
Rewriting emails
Distilling policies into bite-sized talking points
Using it to get unstuck, not do the whole thing for me
It wasn’t about replacing my work.
It was about enhancing it.
Less “do this for me,” more “co-pilot with me.”
But let’s be real—there’s an ethical side to this, too.
Even now, I wrestle with it. Because as much as I enjoy using AI, I also see how it gets misused—to cut corners, to save money, to fake learning, to gut creative teams in the name of "efficiency.”
AI can be a tool for empowerment. Or a weapon for exploitation. It all comes down to who's using it and why.
I still hold the stance: if you're using GenAI to skip the learning process, you're missing the point. If you're using it to replace human creativity, connection, or accountability, you're doing it wrong.
This tech isn't magic. It's mirrors. And it reflects the bias, brilliance, and bullshit of its users.
So yes, I use Chattyboi. But I’m careful about what I feed it. I don’t drop in private data—mine or anyone else’s. I don’t treat it like a vault. I treat it like a very smart intern with boundary issues.
And I don’t believe it should be allowed to scrape people’s work without consent. Stealing someone’s art, words, voice, or likeness just because it’s online doesn’t make it innovative—it makes it exploitative.
So I side-eye every AI pitch that promises "faster, cheaper, smarter" like it doesn't come with trade-offs. Because you know what else is faster and cheaper? Garbage.
We don’t need AI to erase the messy, beautiful, deeply human parts of work. We need it to support them.
So… what’s the point of all this?
I’m not saying GenAI will change your life because it helped you write a brunch invite. I’m saying: the best way to learn something intimidating is to give yourself permission to be bad at it first.
To tinker without a timeline.
To fail without fallout.
To practice curiosity without the pressure of productivity.
I didn’t start using GenAI for work until I trusted it enough to make jokes with me. I didn’t trust it until I’d seen it get things wrong—and recover. I didn’t let it near my “real” work until I’d used it on silly, low-stakes stuff.
And if you’re GenAI-curious but unsure where to begin? Try a workshop. I recently took one2 called “AI for Workplace Productivity” that offered hands-on exercises, real-world use cases, and just enough structure to make the chaos feel manageable that helped me connect the dots between playful exploration and practical use (and validate everything I’d been doing based on my own intuition and skepticism).
You don’t have to start with prompts that change your life. You just need a parking lot. A party invite. A reason that feels unserious and safe.
Start where it doesn’t matter.
So you’re ready when it does.
As always, caffeinated, chaotic, & curious
- Mads
Acknowledgment: I want to acknowledge that being able to joke about weed so openly is a privilege. As a white person living in a state where recreational marijuana is legal, I get to make light of something that continues to disproportionately harm communities of color through criminalization and systemic inequity. Humor doesn't erase history—so let this be both a wink and a nudge toward deeper awareness.
Subscribe to join a community of humans trying to figure it out without faking it.
Note: No weed was actually ingested in the making of this blog. But the vibes were high, respectfully.
Note: I work for the company that provided the training and was able to take the workshop for free as part of my work perks, this workshop was able to help set the stage for thinking about how I could integrate AI into my day-to-day while being mindful of what can/should be shared on the platforms – we focused on using ChatGPT Pro & Gemini Enterprise in the training when I completed it.