Every week I talk with my friend Jonathan Goldman. He’s an artist in Woods Hole, Massachusetts. I’m in Santa Barbara. He’s used technology to make his work his whole life, and he picks up any tool that helps.
This Tuesday we talked for an hour. Before we hung up I told him I’d write about it on Sunday, and I asked for a picture of the two of us.
Then I told him about the six hours a week I give to conversations like this one. No agenda. We find thirty-five minutes of meat and potatoes. If I’d lived in Paris during cafe society, I would have been at the table. Probably as the waiter.
He laughed.
Through Another Lens is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
Three shocks
The iPhone arrived in 2007. By 2016, the leaders I worked with had all come up through the same school of thought. Plan the year. Run the quarter. Manage the people in the building.
Then three things hit inside ten years. Mobile put the internet in every pocket. COVID sent the building home. ChatGPT arrived in late 2022 and put a tutor, a writer, and an analyst on every desk.
None of the three stayed local. They reached everyone, everywhere.
Gallup, the research firm that tracks workplace engagement, reports that global manager engagement fell from 31 percent in 2022 to 22 percent in 2025. Nine points in three years. Jonathan asked whether this is partly a generational age out. Maybe it is. I’m 73, and by any count I’m one of the old guard. Where I see the split is in who spends an hour a day learning.
What held
Two things I learned long ago still work.
At Wavefront, the 3D software company I co-founded 43 years ago, we sold in product families. Group what you offer by who needs it. Tucker Howard, who builds IdeasOut with me, writes every feature as a user story in the Agile method. As a person, I want this, so that I can do that. Both come down to one question. Who is this for, and what do they need done?
That question survived the phone, the pandemic, and the chatbot. What changed is speed, and how a person learns.
During COVID I cooked for seven families. Each had its own needs. I wrote seven menus, built the shopping list from them, and put an ordering app in front of the families. Product families, user stories, a working tool. I did it in weeks, with software I couldn’t have touched five years earlier.
The contact sheet
Jonathan is a sculptor trained at MIT. His current project, Truth Lost at Sea, is a forty-eight-foot floating sculpture. I worked with him on it.
He also has tens of thousands of photographs from forty years of work. When he started using AI on that archive, he skipped the search box. He asked for a visual contact sheet, the way a photographer reads a roll of film. Each image gets a verdict. Keep, leave, or delete. The tools sort in the background while he looks.
He saw how he’d use the machine before most people finish asking what it does. That’s an artist’s leap.
See one
I learned see one, do one, teach one in kitchens. Watch a cook break down a chicken. Break one down yourself, badly. Then show the next person.
That loop works on AI too. See someone use it on a real problem. Do it on your own. Then teach it to somebody else, because teaching is where you find out what you don’t know.
The struggle in the middle matters. Researchers call it desirable difficulty. Learning that costs you effort sticks. If the machine hands you every answer, nothing stays in your head. Use it to push on the problem, and keep your hands on the work.
Five steps
AI is generative. Ask it to do something twice and it finds two different ways to do it. The models change underneath you, so the process changes too.
At Coastal Intelligence, we’ve come up with a way to approach AI and run the same five steps every time: Intent, Data, Envision, Validate, Apply. It’s called IDEVA.
Intent: why does this work exist, and who is it for? Data: what do I already have that the work should draw on? Envision: what does good look like, written down so it can be scored? Validate: where does the draft break when I push on it? Apply: put it to work, then write down what made it work.
IDEVA is new. It’s too soon to tell whether it holds up for anyone but us and our clients. I personally run it ten times a day, and the output has gotten steadier. When the tool changes, the process doesn’t.
The hour
When I think about a week, there are 168 hours. Call 50 of them your professional life. Ten percent of that is five hours, an hour a day that I can spend learning something new.
When I’m building something, I spend closer to thirty percent. You don’t need that much to start. Trade one status meeting for an hour of learning. Spend it on a real problem somebody handed you this week, with a real person waiting on the answer.
It’s an investment in your relevance. Five to ten percent of a work week, and the return compounds.
The table
Here’s what Tuesday taught me. You have to be human to learn all this. You need someone across from you who’ll laugh at the waiter line, challenge the age-out question, and say show me the contact sheet.
So I keep a standing conversation outside the building. An hour with Jonathan, no agenda. Neither of us brings a deliverable. We bring what we’re working on, and we leave with something neither of us had at the start.
Find your table. Show up at the same time each week. Bring a problem and an open mind.
I’ll be there Tuesday.
Through Another Lens is also a podcast. Find it wherever you listen, and subscribe here on Substack so the next one lands in your inbox.
Thanks for reading Through Another Lens! This post is public so feel free to share it.