Calvin's Updates

The Latchkey Club Daily Draft — 2026-07-20

Teleprompter / Blog Script

I spend a lot of time around new technology, and even I have days when I open a piece of software and think, “Where did they move everything this time?”

The button I used yesterday is gone. There’s a new AI panel taking up half the screen. Something that used to take three clicks now takes one click, but only after twenty minutes of trying to figure out which click it is.

And when you’re 57, that can produce a thought you probably wouldn’t say out loud at work.

Maybe this is the signal.

Maybe the tools have finally moved past me. Maybe I should just get out before everybody else notices.

Welcome back to the channel, guys.

Today I wanted to talk about older workers, AI, and the danger of letting a bad technology transition make a retirement decision for us.

I saw a report from the Center for Retirement Research at Boston College that looked at workers 55 and older in jobs with different levels of AI exposure.

The researchers found that since ChatGPT arrived, older people in highly exposed jobs have become somewhat more likely to leave work and move into unemployment. The examples include jobs like programming and accounting, where a lot of daily tasks can now be done or changed by AI.

They were careful about the conclusion. This doesn’t prove that AI is simply replacing every older worker, and highly exposed office jobs can still have lower exit rates than more physical jobs.

But the pattern is worth paying attention to.

Some people may be pushed out. Some may decide they don’t want to learn another system this late in a career. Some employers may use AI as an excuse to reduce staff. And some workers may look at the whole mess and decide retirement sounds better than spending another year pretending to be excited about mandatory training.

I understand that feeling.

Part of the appeal of retirement is being done with all of it. No more reorganizations. No more passwords that expire right after you finally remember them. No more software rollout led by somebody who says the process is intuitive while sharing a 74-slide presentation about it.

But I also think we have to separate three different things.

Is the problem the job?

Is the problem the way the company is introducing the technology?

Or is the problem that I’m uncomfortable being a beginner again?

Those are not the same problem, and they shouldn’t automatically lead to the same answer.

A bad employer can make a useful tool feel threatening. If management announces that AI will “transform the workforce,” cuts the training budget, and expects everybody to become an expert by Friday, that’s not proof that older people can’t learn. That may just be poor management with newer vocabulary.

The tool itself may also be bad. Not every AI feature improves the work. Some of it creates more checking, more tabs, and more ways to produce the wrong answer faster.

Experience helps with that because after a few decades, you’ve seen enough expensive solutions looking for a problem.

But then there’s the third possibility.

Maybe the tool could help, and I just don’t like how it feels to be new at something again.

That one is harder to admit.

I’ve been programming on and off since college, mostly as a hobby. For years I could build small things, but anything larger required more time and energy than I had available.

AI-assisted coding changed that for me. I can describe a real problem, work through the logic, test what gets built, and turn knowledge that was sitting in my head into something useful.

I’m still responsible for whether it works. The AI can write confident nonsense just like a person can. It just does it without needing coffee.

But the tool lets me do work I probably would not have attempted a few years ago.

That doesn’t mean everybody needs to become a programmer or spend the weekend talking to a chatbot.

It means the same technology that makes one older worker feel replaceable may allow another older worker to stay useful longer, reduce the part of the job that wears them down, or finally build something from what they know.

The difference may not be age. It may be whether the person gets a fair chance to learn, whether the employer values judgment, and whether the tool is aimed at a real problem.

People our age do have something useful here.

We know where the exceptions live.

A younger employee may learn the new system faster, and that’s fine. Speed matters. But somebody still has to know why the customer’s “simple request” is not simple, which number in the report is usually wrong, what happened the last time the company tried this, and when an answer that looks efficient will create a mess three departments away.

That context is not visible in the software demo.

The mistake would be assuming context is enough by itself.

Experience that refuses to learn can become a story about how everything used to be better. New technology without experience can become a very fast way to repeat old mistakes.

The useful combination is knowing the work and staying willing to learn the tool.

So before somebody makes an early-retirement decision because AI arrived at work, I think there should be a small test.

Pick one task you know well. Not the whole job. Not a vague plan to “learn AI.” One task.

Maybe it’s summarizing a long document, preparing the first draft of a customer update, comparing two reports, organizing field notes, finding a pattern in a spreadsheet, or writing instructions for something you’ve explained fifty times.

Try the tool on that task for thirty days.

Keep the risk low. Don’t put private company information into an unapproved system. Check the output. Measure whether it saved time or merely moved the work around.

Then ask some plain questions.

Did this remove a part of the job I hate?

Did it make me better at a part I already understand?

Did it create more mistakes than it solved?

What did I know that the tool did not know?

And what training or access would make this genuinely useful?

That little experiment won’t solve ageism, and it won’t protect every job. Sometimes the employer really is cutting positions. Sometimes health, caregiving, burnout, or finances make retirement the right decision. Sometimes leaving is not fear. It’s wisdom.

I just don’t want embarrassment to dress itself up as wisdom.

There’s a difference between choosing retirement because the timing is right and retreating because a new screen made you feel old on a Tuesday afternoon.

If the job is wrong, leave the job when you can.

If the company is handling the change badly, name that honestly.

But if the real obstacle is having to become a beginner for a little while, maybe give yourself permission to be bad at the new thing before deciding you can’t do it.

At this age, we’re not starting from zero.

We’re adding a tool to thirty or forty years of seeing what happens after decisions get made.

That history still counts. The new software doesn’t erase it.

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

Has AI at work made you feel more capable, more replaceable, or some uncomfortable combination of both?

And if you’ve considered leaving because the tools changed, what would a small thirty-day test look like before you decide?

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. Cold open: What goes through your mind when software moves the button again or adds another AI panel?
   • Follow-up: Have you ever quietly wondered whether the change was a signal that it was time to leave?
2. The report: What did the Center for Retirement Research find about workers 55+ in highly AI-exposed jobs?
   • Follow-up: What did the researchers not prove, and why is the nuance important?
3. Three different problems: Is the real problem the job, the employer’s implementation, or the discomfort of being a beginner?
   • Follow-up: Why shouldn’t those three problems automatically produce the same retirement decision?
4. A bad rollout: How can weak training or cost-cutting make a useful tool feel threatening?
   • Follow-up: When is the AI feature itself simply not useful?
5. Your own experience: How did AI-assisted coding change what you could build from knowledge you already had?
   • Follow-up: What are you still responsible for checking?
6. The hidden 55+ advantage: What do experienced workers know about exceptions, customers, history, and downstream consequences that a demo cannot show?
   • Follow-up: Why is experience alone still not enough?
7. The thirty-day test: What one familiar task could somebody safely try with AI before making a larger judgment?
   • Follow-up: What should they measure, protect, and verify?
8. When leaving is wise: What real circumstances can make retirement or a job change the right decision?
   • Follow-up: How can somebody tell the difference between wisdom and embarrassment?
9. Closing: Has AI made you feel more capable, more replaceable, or both?
   • Follow-up: Invite viewers to describe one small test they could run before deciding the new tools have passed them by.

Title Options

1. Your Career Isn’t Over Because the Software Changed
2. When AI at Work Makes You Feel Obsolete
3. Before You Leave the Job, Test the Tool

Thumbnail / Onscreen Text Options

• AM I TOO OLD FOR THIS?
• DON’T LEAVE YET
• TEST THE TOOL FIRST

Shorts / Reels Cutdowns

• “Is it the job, the company, or being a beginner?” Use the three-question distinction and the point that each problem deserves a different response.
• “A bad rollout is not proof you can’t learn.” Cut from the mandatory-training example through poor management using newer vocabulary.
• “We’re not starting from zero.” Use the section about exceptions, customer history, downstream consequences, and adding a tool to decades of judgment.

Viewer Question

Has AI at work made you feel more capable, more replaceable, or both—and what one task could you test for thirty days before making a bigger career decision?