Reading Time: 5 minutes

Hey Prompt Lover,

Yesterday I asked what skill you quietly stopped building the moment AI could do it for you.

I expected writing. I expected research. I expected people to talk about the soft skills — the ability to sit with uncertainty, to think through a problem without outsourcing the thinking.

Some of that came back. But the reply that stopped me was from Jeff Pogue. Jeff is the COO of Electric Avenue Yard Care and he replied with something that reframes the entire question.

He told me the skill he stopped building was learning Google Sheets formulas.

And then he said this.

"I now make even more complex spreadsheets than I have ever been capable of doing in the past. But I sort of have no idea how the formulas work. I can read them and just go, wow. I NEVER could have made that."

Then: "I am not learning new skills within Google Sheets. I am learning how to prompt it better. Jury is out on which is the better skill."

Read that twice. Because it breaks open something I hadn't fully named yet.

Jeff is not less capable because of AI. He is more capable. The spreadsheets are better than anything he could have produced before. The outputs are real. The dashboards work. The results are better than they were.

But he does not understand what he made.

And that is an entirely different kind of problem from the one I described yesterday.

2026 will redefine IT as a strategic driver of global growth. Automation, AI-driven support, unified platforms, and zero-trust security are becoming standard, especially for distributed teams. This toolkit helps IT and HR leaders assess readiness, define goals, and build a scalable, audit-ready IT strategy for the year ahead. Learn what’s changing and how to prepare.

Download the Toolkit

Yasir also replied. He said the skill that is diminishing for him is summarising.

"Now when I look at a long text, I put it into AI first."

That one is clean. A skill outsourced, slowly degrading. That is the story I told yesterday.

But Jeff's story is different and more interesting and more uncomfortable.

Because Jeff isn't describing a skill he lost. He is describing a capability he gained without the understanding that should come with it. He can produce something he cannot explain. He owns outputs he did not build in any meaningful sense. He operates at a level his actual knowledge cannot justify.

And here is the question that has been sitting with me since I read his reply.

If you can do something but cannot explain how it works, are you actually capable of that thing?

If AI makes you capable of things you don't understand, what does capability even mean?

This is not abstract. It is already happening across every profession that uses these tools.

A lawyer who uses AI to draft contracts they couldn't draft themselves. Are they a better lawyer or a more convincing one? A developer who ships code they can't fully read. Are they a better developer or a faster one? A marketer who produces analysis they couldn't produce manually. Are they more skilled or better resourced?

The traditional answer to "are you capable" was: can you do the thing? The AI answer is more complicated. Because you can do the thing but the thing is now being done by something else running inside your workflow.

Jeff said it himself. Awesome spreadsheets. No idea how the formulas work.

What is he, exactly? A more capable COO? Or a COO who has excellent tools and an increasingly thin understanding of how they produce what they produce?

Here is the version of this problem that most people haven't hit yet but will.

The day AI is unavailable. The system goes down. The tool changes. The model gets updated and produces something different. The spreadsheet breaks and you cannot fix it because you don't understand it. The contract needs to be amended and you can't parse what the original said.

What is your actual capability on that day?

This is not a hypothetical. It is a debt. Every time you produce something you couldn't explain, you accumulate a small amount of competence debt. Individually invisible. Collectively significant. The kind you only discover at the worst possible moment.

Jeff's jury is genuinely out on this one. He is more productive. His outputs are better. His business benefits from what AI produces for him. All of that is real.

But the gap between what he can produce and what he understands keeps growing. And a gap like that is fine until it isn't.

I think what the first two days of replies are showing is that there are two fundamentally different relationships people are building with these tools.

The first group uses AI to amplify things they already understand. They bring domain expertise to the tool, the tool accelerates their existing knowledge, and the output is a better version of what they could have produced anyway. They could explain every output even if they couldn't have produced it that fast. Theazaland, who replied on Day 1, is this person. They push back on Claude when it strips the human out. They understand the work well enough to know when the tool gets it wrong.

The second group uses AI to access capability they don't actually possess. The outputs are real. The results are valuable. But the understanding isn't there. Jeff is honest about this in a way most people wouldn't be. He can read the formula and go "wow, I never could have made that." That wow is worth paying attention to. It is the distance between what he produced and what he understands.

Neither group is doing something wrong. But they are accumulating very different things over time.

I am in both groups depending on the day.

For writing I am the first type. I understand every sentence even when Claude helps me find it. I could write a slower version of most things without the tool.

For technical tasks I am increasingly the second type. I build things I could not fully audit. I use outputs I couldn't fully recreate. The wow is happening for me too. And I am not always sure that is wisdom compounding. Sometimes it is dependency accumulating.

The question Jeff left me with is the right one for today.

What is the gap between what you can produce with AI and what you actually understand? And how wide is it getting?

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide.

Today's reply question is this.

Think about the most impressive thing you have produced using AI in the last month. Could you explain how it works to someone who needed to maintain it after you?

Reply and tell me what the gap looks like. The most honest answers shape where this series goes.

Day 4 arrives tomorrow.

— Prompt Guy

Day 3 of 30. One question every day that the AI industry is not asking. Forward this to someone who should be in this conversation.