Mount Stupid

The doctor flew out for a week.

The email sorter she’d built six weeks earlier, the one she’d been opening every morning before clinic to triage what mattered, stayed home with the desktop. Different laptop on the road. Different rhythm. Easier to just open Gmail.

She came back to a cobwebbed sorter and her staff still working their inboxes the old way. Not because the tool failed. Because nobody told them to keep going while she was gone.

That moment is what I want to talk about. The sorter didn’t break. The staff weren’t lazy. The owner didn’t backslide. Something else happened, and it’s the thing most operators miss the first time AI works for them.

She’d planted a flag. Then she’d stopped climbing.

Plant the flag, take the photo

Here’s the thing. The win was real. The sorter worked. Six weeks of clean mornings, less noise in front of patient care, calmer staff because she was calmer. That’s not a fake win. The flag was earned.

But the flag is not the summit.

There’s a phrase for the trap. Mount Stupid. I read it somewhere and can’t pin the source, so take it as observation rather than gospel. Mount Stupid is the mountain you climb where the view is real and the height is small. You set out to climb a mountain. Two hours in, you crest a ridge. Plant the flag. Take the photo. Beautiful view. Real summit, in the sense that you actually climbed it. But there’s a peak behind it that’s three times higher, and from where you’re standing you literally cannot see it. The hill you climbed is the only mountain you can see.

Most operators are on it right now with AI. Most don’t know.

The pull toward the flag-and-photo posture is stronger right now than it has ever been, because the wins are real and they come fast. The first useful prompt feels like a summit. The first custom GPT feels like a summit. The first agent that drafts a clean email feels like a summit. They’re all hills. The next peaks are invisible from where you’re standing, because you’re standing on the only mountain you can see.

The discipline isn’t gratitude. It’s looking up.

Here’s where the Four D’s start doing work.

The Four D’s name four stances an operator can take relative to a single piece of work. DOING. DISPATCHING. DIALING. DECIDING. They’re not a ladder you climb once for the whole business. They’re the change in stance you make for one dish at a time. The doctor with the email sorter had moved from DOING (sorting the inbox herself every morning) to something that felt like DECIDING (the AI did it, she checked the output, she moved on). It felt like she’d moved up. The flag said she had.

She hadn’t.

She’d just shifted DOING off her own hands and onto a tool she still had to babysit. The work was still tethered to her personal Monday rhythm. The recipe wasn’t written down. The staff hadn’t been onboarded into the new flow. The system existed in her head and in her desktop, and that was the whole of it. Take the desktop away, take her out of the building, and the system goes dormant.

DECIDING looks different. DECIDING is the stance where you’ve stepped back, the dish is being made by the station, you’re tasting it once in a while, and you’re asking the real question. Is this dish still on the menu? Is this the work we should be doing? DECIDING isn’t did the AI handle the email triage correctly today. DECIDING is was triaging the inbox the right job for that hour, or were there three other questions we should have been asking?

The gap between the doctor’s stance and DECIDING is Mount Stupid. The flag was real. The summit was not.

Yeah. That’s the trap.

The eight-minute version of the twenty-minute task

Here’s the cost of staying on the first hill.

You get better at the eight-minute version of the twenty-minute task. The email sorter shaved twelve minutes off the doctor’s morning. Six weeks in, she was very good at her twelve minutes. Better than she’d ever been at the original twenty. The output got crisper. The rules sharpened. The triage was more accurate than her own hand-sorting had ever been.

And the wrong question was getting answered faster.

The real question was never how do we make the morning triage twelve minutes shorter. The real question was why does the doctor sort the inbox at all. Twelve other questions sat under that one. Should the staff be sorting before she sees anything? Should patients be routed away from email entirely for the things that don’t belong there? Should the sorter be a station on the line, owned by an ops person, running whether the doctor is in the building or not?

She wasn’t asking any of those. Not because she didn’t know how to ask them. Because the flag was good. The photo looked great. And the climb you don’t take is invisible from the hill you already climbed.

The trip is the thing

This is why the trip mattered.

The trip didn’t break the tool. The trip exposed that the tool had never been a system. It had been a habit she ran on her desktop, dressed up in AI clothing. The minute she left the building, the system left with her. Habit, not system. Does that make sense?

I want to be careful here, because this isn’t a piece about the doctor being wrong. She wasn’t wrong. I am writing this piece because I see this pattern in client work constantly, and I see it in my own work too. I am on the same mountain. The first time I built something with AI that actually saved me real hours, I planted a flag for a month before I noticed I hadn’t built a system. I’d built a clever habit. The discipline of looking up is something I’m still trying to install in myself.

So the move isn’t the doctor should have known better. The move is the first AI win is a flag, not a summit, and the climb that matters starts the day after the flag goes up.

What looking up looks like

Right? OK. So what’s the actual discipline.

Three questions, and they all sit above the work the flag celebrates.

If I left for a week tomorrow, would this still be running when I got back? If the answer is no, the win was a habit. The system was never built. The climb you’re not taking is the one where the recipe gets written down and somebody on the staff is the one running it whether you’re in the building or not.

What is this freeing me up to do, and am I actually doing it? Twelve minutes a day for six weeks is six hours saved. If those six hours went into the same kind of work the AI was helping with, you haven’t moved. You’ve just gotten faster on a hill. The real climb starts when those six hours go into work the AI cannot do for you. Higher-altitude work. Pricing. Hiring. Strategy. The calls only you can make.

What did this win make obvious that I wasn’t seeing before? This is the question Mount Stupid hides. The first AI win usually exposes a half-dozen other things that need redesigning, and operators who do not look up never see them. The morning inbox triage being clean is a small win. The thing it should have revealed, that the inbox should not be carrying that much of the practice’s communication in the first place, is the next peak.

The Monday move

Pick the most recent AI win you posted about, told a peer about, or quietly congratulated yourself for.

Ask one question. If I left for a week tomorrow, would this still be running when I got back? If the answer is no, the win was a habit you ran on your desktop. The climb you weren’t taking was turning the habit into a system the team owns whether you’re in the building or not.

You don’t have to take the climb today. You just have to know which way the next peak is.

Plant the flag. Then look up.

Original framework. Distilled from client work.

Framework spine: The Four D’s, specifically the gap between DOING and DECIDING. Read the full framework.