Calvin's Updates

The Latchkey Club Daily Draft — August 2, 2026

Teleprompter / Blog Script

Welcome back to the channel, guys.

Today I wanted to talk about something that sounds like a work topic, but I think it is really a retirement topic.

I was looking at some of the things I have been trying to document at work. Processes, systems, why certain decisions were made, what to check when something looks wrong. And I had this uncomfortable thought.

If I put everything I know into a folder, and nobody can actually do the job when I leave, did I transfer anything?

Or did I just organize the evidence that I used to know it?

So, let’s get into it.

I’m 57, and retirement is no longer some imaginary event sitting way out past the horizon. I still have time. I’m not packing a cardboard box next Friday. But I can see the runway now.

And one of the things I’ve been thinking about is what I owe the people who will still be there after I’m gone.

For a long time, I thought knowledge transfer meant documentation. Write the instructions. Save the diagrams. Clean up the file names. Put everything in a place where somebody else can find it.

That matters. I have worked around enough old systems and mystery equipment to know that a useful diagram can save somebody hours. Sometimes the only documentation is a file called FINAL, then FINAL2, then FINAL2-USE-THIS-ONE. That is not a system. That is an archaeological site.

But documentation is only one layer.

The harder part is the stuff you do without realizing you are doing it.

You notice that one number looks technically valid but does not fit the property. You hear the way a customer describes a problem and realize the problem is probably somewhere else. You know which shortcut is safe, which shortcut creates a mess six months from now, and which “temporary” fix has already been temporary since 2019.

That knowledge does not always show up in a checklist.

It is pattern recognition. It is context. It is remembering consequences.

A lot of us in Gen X learned work that way. Somebody showed us part of the job, handed us the keys, and then disappeared into another meeting. We figured out the rest. Sometimes that built confidence. Sometimes it just meant we made avoidable mistakes quietly.

Now we are old enough to be the people with the keys.

And I think there is a temptation to believe that because we had to figure it out, the next person should have to figure it out too.

I understand that feeling. There is value in struggle. There is value in being allowed to solve a problem instead of being given every answer.

But there is no virtue in making somebody repeat thirty years of preventable mistakes just so they can earn the same scars.

The better question is: How do I transfer judgment without creating dependence?

That is where I think technology, including AI, can help. But probably not in the way the headline makes it sound.

There are companies talking about using AI to capture the knowledge of retiring workers. The basic idea makes sense. Let an AI system interview an experienced employee. Ask follow-up questions. Turn conversations, notes, old emails, and procedures into something searchable.

That could be useful.

An AI interviewer might keep asking after a human interviewer moves on. What do you mean by “check the usual things”? Which things? In what order? What makes you suspicious? What exception has burned you before?

Those are good questions.

But I do not think downloading a person into a chatbot solves the handoff.

Because knowing what I would say is not the same as knowing what to do.

A new person still has to face the customer. They have to decide with incomplete information. They have to live with the result. They need enough room to form judgment of their own.

And they may see something I missed.

That last part matters. If the goal is to create a digital version of me and train everybody to imitate it, then I have not built a successor. I have built a monument.

I do not need the next person to become me. I need them to understand the purpose behind the work well enough to make a better decision when the old answer no longer fits.

So I’ve been thinking that a real handoff needs three things.

First, yes, capture the facts.

Where is the equipment? What connects to what? Who owns the decision? What are the recurring problems? What did we try before? Why is the system designed this way?

Use documents. Use diagrams. Use video. Use AI to organize it if the information is approved and protected. Make it easy to search. Do not make the next person solve a scavenger hunt.

Second, capture the questions.

This may be more important than capturing the answers.

What do I look at first? What would make me stop? What assumption am I making? Who is affected if I am wrong? Is this a local fix that creates a larger problem? What does success look like a month from now, not just at the end of today?

Good judgment often sounds like a set of questions asked at the right time.

AI can help pull those questions out of us. It can look across past incidents and ask what changed the decision. It can draft scenarios. It can say, “Here are three cases that look similar. Why did you handle them differently?”

But then comes the third part.

Let somebody else drive.

Give them a real problem while you are still there. Let them do the discovery. Let them explain what they think is happening. Let them make the recommendation. Ask why. Do not rescue them the moment their approach differs from yours.

That part is hard, especially if you have been the person everybody calls when something gets complicated.

Being needed feels good. Even when we complain about it.

The phone rings, somebody has a problem, and you know what to do. That becomes part of your identity. Then you start training other people, and if it works, the phone rings less.

That can feel like becoming less important.

Maybe it actually means the work is working.

I have been trying to think about retirement less as the day I stop and more as the period when responsibility moves. If I wait until my last month to explain everything, I waited too long.

The test is not whether I can give a good presentation about what I know.

The test is whether I can step out of a decision and watch somebody else carry it well.

Can they explain the why, not just repeat the steps? Can they recognize when the standard process does not fit? Can they teach the next person? Can they disagree with me for a good reason?

That last one may be the real test.

For those of us in our fifties and early sixties, AI creates an interesting opportunity. It can help us sort decades of notes, surface forgotten decisions, record stories, build training cases, and make specialized knowledge easier to find.

But the hidden advantage is not that we can preserve ourselves forever.

It is that we still have time to use these tools while we are present. We can turn memory into a starting point, then spend the remaining years helping real people develop judgment.

Not protecting our little kingdom. Not making ourselves impossible to replace. Not leaving behind a beautiful library nobody knows how to use.

Just making sure the next person gets a fair chance to become capable.

Anyway, that’s what I’ve been thinking about.

If retirement is somewhere on your horizon, maybe ask yourself this: If I were gone for thirty days, what would break? And is the answer another document—or another person who needs the chance to lead?

I’d be curious how you are handling that in your own work. Leave me a note in the comments.

Thanks for listening.

Video Prompt Script — Questions to Answer Without Reading

Use these as prompts. Don’t read them on camera; answer them naturally.

1. Opening: What were you documenting when you realized that a folder full of instructions might not equal a real handoff?
   • Follow-up: What would still fail if you disappeared from work for thirty days?
2. Retirement horizon: How has turning 57 changed the way you think about responsibility to the people who remain after you retire?
   • Follow-up: Why is this more than a workplace-efficiency issue?
3. Facts versus judgment: What parts of your work can be documented, and what parts live in pattern recognition or context?
   • Follow-up: Give a generalized example of a technically valid answer that can still be the wrong answer.
4. Gen X formation: How did your generation usually learn a job—through formal training, or by being handed part of it and figuring out the rest?
   • Follow-up: Which parts of that approach built resilience, and which mistakes should not be repeated?
5. AI’s useful role: How could AI interview experienced workers, organize approved information, uncover assumptions, or build training scenarios?
   • Follow-up: Why is a searchable AI version of an expert still not a successor?
6. The three-part handoff: What does it mean to capture the facts, capture the questions, and then let somebody else drive?
   • Follow-up: Which of those three is hardest for you personally?
7. Identity and usefulness: Why can fewer calls and fewer rescue missions feel like becoming less important?
   • Follow-up: Could that actually be evidence that mentoring is working?
8. The real test: Can the next person explain the why, recognize exceptions, teach another person, and disagree with you for a sound reason?
   • Follow-up: What opportunity do people in their fifties still have if they begin now?
9. Close: If you were gone for thirty days, what would break—and does it need another document or another person with room to lead?

Title Options

1. Before You Retire, Let Somebody Try to Replace You
2. Your Retirement Handoff Is More Than a Folder
3. AI Can Save Your Notes. It Cannot Build Your Successor.

Thumbnail / Onscreen Text Options

• WHO CAN DO IT WITHOUT YOU?
• A FOLDER IS NOT A HANDOFF
• LET SOMEBODY ELSE DRIVE

Shorts / Reels Cutdowns

• “A folder is not a handoff” — The opening realization through: “Or did I just organize the evidence that I used to know it?”
• “Don’t build a monument” — The contrast between an AI imitation of an expert and a successor who can think independently.
• “Let somebody else drive” — The practical test: give someone a real problem, ask why, and resist rescuing them because their approach differs.

Viewer Question

If you were gone from work for thirty days, what would break—and is the answer better documentation or giving another person the chance to lead?