Keep the thinking.
An AI field note
From everyday work to deeper learningAustin · Workshop edition
A personal perspective

A team
of one.
More within
reach.

Everything I produce at work has been touched by AI. ChatGPT and Codex help me check tone, get precise, and keep the thread from one task to the next.

Here’s what I’m learning

What I get from AI. What I want to carry forward.

A reflection in three parts
01 Everyday · 02 Limits · 03 Learners
01My everyday

AI is woven into
the whole day.

I use ChatGPT and Codex constantly. The organizational back-and-forth, emails, changing documents, and conference proposals all still land with me. AI helps me keep moving across them.

Professionally

A second read on the work.

I can check how an email will land, make a sentence more precise, or turn rough thoughts into a document or conference proposal. I appreciate having a place to think out loud and check that I’m saying what I mean.

TonePrecisionProposals
Personally

The everyday, too.

Travel planning, shopping, photo editing, help selling clothes, and keeping up with everyday life. I’m comfortable giving AI room to suggest things and handle more of the steps, then shaping the result as we go.

Plan a tripSell clothesEdit a photo
How it fits together

AI helps me keep the thread.

Having somewhere to bring the pieces back together helps me keep track of what I’m already trying to carry.

  1. Recover the context.

    Meeting recordings, transcripts, and notes help me return to what happened. AI helps me pull out decisions, follow-ups, and the next thing I need to do.

  2. Check the meaning.

    I ask for a clearer tone, a more precise phrase, or a check on an assumption. I correct what’s off and keep asking until the result fits what I mean.

  3. Make something I can use.

    I give AI room to make suggestions and take on more of the work. A conversation can become a usable draft, a conference proposal, or a working website.

One example: my weekly job roundup. I look for roles too. AI helps find and assemble options; I help decide what belongs in the post. It’s a recurring project we build together.

02The catch

When it misses,
I look at the ask.

Sometimes I’m disappointed. Often, I can trace it back to my prompting or to asking for too much at once. That makes using AI an ongoing practice in being clearer about what I need. I also correct assumptions and ask which details still need confirming.

Was I clear enough?

Have I given it the context and direction the task needs?

Was I asking for too much?

Would breaking the problem into smaller steps help?

I still have to read it, question it, and make the call.

03For learners

More AI.
More need for literacy.

What I want to carry forward is the ability to think critically, understand what’s in front of you, and work through the problem. That includes knowing when to invite AI’s suggestions, how to check them, and how to explain your own reasoning.

Read for meaning.

Understand what a response actually says. Have the literacy to ask a clear question and make sense of the answer.

Question what comes back.

Look at the reasoning. Notice assumptions. Ask whether the explanation makes sense and what would support it.

Work through the problem.

Stay involved in the steps. Decide what to delegate, try an approach, and explain why it works. Recognize where you need more help.

A small experiment · Try a different ask

What are we asking AI to do?

Scripted prompt examples
No live AI connection

Choose an approach

The prompt

“Solve this problem and give me the answer.”

What’s visible?A finished result. I still need a way to tell whether the learner understands how to get there.

These examples illustrate a learning intention. A prompt can invite reasoning; seeing and checking that reasoning is still part of the work.

↗A question for the room

How do we make understanding as visible as the finished work?

I want learners to leave with a stronger ability to read, question, and work things out. That’s what I want to keep building toward.

Revisit the prompt experiment
A 2–3 minute walkthrough

A few notes for the room.

  1. What I was trying to do · 30 secondsShow how I use AI as a team of one: to check tone and precision, recover context, and turn ideas into something usable. The human judgment illustration is the center of that story.
  2. How I worked with AI · 30 secondsI shared the workshop brief and my own notes, then asked Codex to look across selected examples of my AI use. It helped identify patterns and create the copy, design, illustration, and code. The experiences and priorities came from me.
  3. What the page brings into focus · 45 secondsMy weekly job roundup is a shared search-and-edit process. I give AI room to suggest and do more, while I still shape the result. In the experiment, switch from “Get an answer” to “Work through it” to show what I want learners to practice.
  4. The question I’m bringing · 30 secondsHow do we make understanding as visible as finished work? A better prompt can help, but I still want to know how we see the learner’s thinking.

Suggested talking points: adjust these to what you actually discovered. Before presenting, open this page once, try all three prompts, and keep the offline HTML and screenshots on your laptop.