(2025-05-19) When O3 Plans Your Career Better Than You Do

Katie Parrott: When o3 Plans Your Career Better Than You Do. I’ve always hated the question, “Where do you see yourself in five years?”

So the Friday after o3 dropped, filled with equal measures of curiosity and dread, I let OpenAI’s newest model take a crack at earning a co-writing credit.

Six seconds later, I was staring at a scrollable epic, complete with revenue mix, milestones mapped to the quarter, and a five‑bullet case for why this future is “plausible.”

This model dreams bigger for me than I do. (ambition)

Six months ago, GPT‑4 helped me reorient after a layoff. Since then, my career has evolved—and ChatGPT has leveled up from reactive analyst to agenda‑setting aide‑de‑camp.

o3 presents a subtler puzzle: instead of waiting for prompts, it drafts possibilities. It proposes versions of you that you might actually want to become, and charts a path to get you there.

co-authoring a future with a machine unburdened by human baggage around ambition.

For months, I’ve fed ChatGPT the raw footage of my working life: half‑finished essays, dozens of style guide iterations, mini AI avatars that I’m building of my co-workers. Each upload teaches the model a little more about who “Katie the content strategist” supposedly is—and that version of “Katie” is apparently way bolder than the carbon‑based original. (single conversation context?)

What struck me about this, compared with past chats, was o3’s habit of preemptively offering a plan instead of waiting for prompts. GPT-4 answered questions; o3 l drafts a future, then waits for your redlines.

The system prompts back: o3 turns doubt into to‑dos

o3 treats it like a bug report: Log it, triage it, ship a fix before you refresh the tab.

When I told it I wasn’t sure I had the chops (or desire) to run a product studio, it didn’t drop the goal; it said “let’s unpack that.”

Then it laid out a four-step process for digging into the source of the self-doubt

Then it slid over a single Mad Lib:
If vibe coding vanished tomorrow, I’d still spend my time ______ because it lets me ______.

I filled in the blanks (”writing” and “understand what I think and connect with people authentically” were my answers) and hit enter. o3 redrew the flight plan yet again. Then it gave me an action item: Block off two hours each month for “agenda‑free writing.”

What used to float in “someday” now has a slot on my calendar, and that shift alone makes the climb feel real.

My implementation log

Here’s how I’m operationalizing the roadmap we built together

o3 made me a metrics tracker. I keep a simple spreadsheet where I collect post views, engagement stats, and content outputs

Run retrospectives. I’ve got slots on the calendar—weekly, monthly, quarterly—to bank wins and capture outcomes, and a thought partner to help me think through what matters

Recalibrate the client mix. Some clients energize me. Some clients... don’t.

Block off agenda-free playtime. Every other Friday, I have two hours booked for idea exploration with no deliverable attached.

Becoming the person o3 thinks I am

I open our chat whenever a new idea sparks or an old doubt resurfaces, because I know it will hand back something concrete to react to. Together, we are writing a story I want to be the protagonist of.

Try this at home: Build your own AI career coach

o3 offered up a tantalizing possibility: It could build a prompt version of itself for me to copy and paste into the custom instructions of a project, so our strategy sessions could span conversations. I had it do that—and so can you.

Set your intentions

• A few possible goals:

What to say to get started

Paste this text into a fresh chat window, being sure the model is set to o3:

Prompts that work well
“What’s a sustainable version of ambition for me right now?”

Optional extras (if you want the full experience)

Want to really nerd out? Ask your AI to:

• Reframe your to-do list each Monday into a focused plan
• Ask you a reflection question every Friday

When you feel like the bot has enough context (and if it hasn’t already volunteered to dot this already), give it the following prompt: Convert this conversation into a prompt I can paste into custom instructions of a project.

Models like o3 are extremely helpful—but they will mirror your ambition, your doubt, and your assumptions back at you. That’s the magic and the trap. Don’t be afraid to argue with them. That’s where the real insight lives.