Calvin's Updates

The Latchkey Club Daily Draft — September 15, 2026

Teleprompter / Blog Script

I had one of those moments recently where a system changed, a familiar process stopped being familiar, and I could feel the thought forming before I had really examined it.

Maybe I’m too old for this.

Not too old to work. Not too old to understand the problem. Just too old to learn another interface, another vocabulary, another way of doing something that already worked yesterday.

And when retirement is close enough to see, that thought can turn into another one pretty quickly.

Maybe I should just leave.

Welcome back to the channel, guys. Today I wanted to talk about AI, late-career work, and why I don’t want a piece of software choosing my retirement date for me.

I’m 57. Retirement is not some distant concept anymore. I’m thinking about the money, but also health, family, timing, and what I want the next season to contain. There are days when work still feels meaningful. There are also days when one more new system arrives and retirement looks less like a plan and more like the emergency exit.

I don’t think I’m the only one feeling that.

A recent study from the Center for Retirement Research looked at workers 55 and older in jobs where many tasks could be affected by AI. Since ChatGPT arrived, older workers in highly exposed occupations became somewhat more likely to leave work and move into unemployment. The pattern showed up in work like programming, accounting, auditing, and management analysis.

The study is careful. It does not say AI caused every exit, and it says the long-term effect is still uncertain. Some jobs may disappear. Some may change. AI may also make certain jobs easier and allow people to work longer.

But one comment in AARP’s reporting stayed with me. If somebody has only a short time left before retirement, learning an entirely new technology may not feel worth the investment.

I understand that calculation.

If I’m leaving in six months, I probably don’t need to master every feature in a system that may be replaced again next year. But if I’m planning to work three, four, or five more years, a bad month with new software should not quietly make that decision for me.

Those years affect savings. They affect health insurance. They affect college and family plans. They affect the people I still want to teach at work. They also represent years of my life, so staying by default is not wise either.

The point is not that everybody should keep working.

The point is that retirement should remain a decision.

That can be hard when the workplace is sending a message that the future belongs to somebody younger. A new survey found that 76 percent of American adults believe people assume older workers are less comfortable with new technology. Among adults 55 and older, the number was even higher.

Once that stereotype gets into the room, it can affect who gets invited to test the new tool, who gets training, and who gets described as adaptable. It can also get into our own heads.

We start apologizing before we ask a question. We say, “I’m not very technical,” even when we have been adapting to workplace technology for thirty years. We see somebody move quickly through an interface and assume speed means understanding.

Sometimes it does. Sometimes they just found the button first.

Experience still has to ask whether the button should be pressed.

I use AI every day now, but I did not begin by understanding all of it. I started with a real problem. I was writing small Python scripts, got stuck, asked for help, copied an answer, broke something, and tried again. Eventually I realized the tool could help with more than one line of code. It could help build a system.

That shift did not happen because I decided to become an AI person. I had work I wanted to do, and the tool reduced the distance between the idea and something usable.

That is still the best way I know to approach a workplace change: begin with the job, not with the technology.

What is one task that keeps slowing me down? What part requires judgment I already have? Can the new tool prepare something without making the final decision? Can I compare the new method with the old one for a month?

That is a smaller question than, “Can I survive the AI revolution?”

I can work with a smaller question.

Before letting frustration move a retirement date, I think I would run a thirty-day test.

First, I would name exactly what is bothering me. Is the job itself no longer healthy? Is the company pushing people out? Has the work lost its meaning? Or am I mainly tired of feeling clumsy inside a new system?

Those are not the same problem.

A tool problem may improve with practice. A management problem may require a conversation or a different role. A financial problem needs real numbers. A health problem deserves more than another tutorial video.

Second, I would pick one ordinary task and use the approved AI tool on that task repeatedly. Not a flashy demonstration. Something that actually shows up on Tuesday afternoon. Draft the field notes. Compare two documents. Organize the questions before a meeting. Build the first version of a checklist.

Then check it with the knowledge I already have.

Did it save time? Did it create cleanup? Did it miss the exception that matters? Did it allow me to spend more attention on the customer, the decision, or the person I’m mentoring?

Third, I would ask for help without turning the request into an apology. Show me how you do this part. Let me watch once, then let me try it. In return, I can explain why the output looks polished but will cause trouble later.

That is not one generation rescuing the other. That is two people bringing different parts of the job.

Fourth, I would update the retirement plan for an earlier exit anyway.

Because sometimes the decision will not be ours. A reorganization, a health issue, or a family need can move the date. Knowing what one year earlier would require is not surrender. It is taking some fear out of the question.

I want a plan for staying and a plan for leaving.

Then I can make a choice without pretending either option is guaranteed.

After thirty days, maybe the conclusion is that the new system is useful. Maybe it handles the repetitive part and leaves more room for the work that still needs experience. That could make the last few working years better.

Maybe the conclusion is that the company is using technology as an excuse to remove people, reduce support, or demand an unhealthy pace. Learning the software will not solve that.

Or maybe the experiment confirms that I am ready to go—not because I failed a technology test, but because the life on the other side is planned, the numbers have been examined, and I know what I am moving toward.

That is a different kind of retirement.

I don’t want to stay just to prove an older person can keep up. I also don’t want to leave just because a rollout made me feel old for two weeks.

The software can change the work. It can reveal problems. It can even create a real threat to the job.

But as long as I still have a choice, it does not get to make the whole decision.

Anyway, that’s what I’ve been thinking about. If a new technology showed up at work tomorrow, would it change when you want to retire—or just change what you need to learn next? 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: When has a changed system made you briefly think, “Maybe I’m too old for this”?
   • Follow-up: How quickly can that become “Maybe I should just retire” when the runway is visible?
2. Why it matters personally: At 57, what makes retirement more than a number on an account statement?
   • Follow-up: Which family, health, timing, and mentoring questions make the date consequential?
3. What the research found: What happened to job exits among older workers in occupations highly exposed to AI?
   • Follow-up: Why is it important not to claim that AI caused every departure?
4. The investment question: When might learning a new system genuinely not be worth the effort?
   • Follow-up: How does the answer change with six months left versus several working years?
5. The stereotype: How can assumptions about older workers and technology affect training, assignments, and confidence?
   • Follow-up: When do we begin repeating the stereotype about ourselves?
6. Your own learning path: How did a real Python problem lead you from asking for one answer to building practical systems with AI?
   • Follow-up: Why was the problem more important than becoming an “AI person”?
7. The thirty-day test: What is the difference between testing AI on one real task and trying to master the entire technology?
   • Follow-up: What should you measure besides speed?
8. Name the real problem: Is the friction coming from the interface, the job, management, health, or finances?
   • Follow-up: Why does each problem require a different response?
9. Keep two plans: What belongs in a plan for adapting and staying?
   • Follow-up: What should an earlier-exit plan account for even if retirement is not the current choice?
10. Closing: What would make retirement a deliberate move toward something rather than an escape from two bad weeks with new software?

Title Options

1. Don’t Let New Software Pick Your Retirement Date
2. Is AI Pushing Gen X Out Before We’re Ready?
3. Before You Retire Because Work Changed, Try This

Thumbnail / Onscreen Text Options

• WHO PICKED YOUR RETIREMENT DATE?
• DON’T LET THE SOFTWARE DECIDE
• LEAVE BY CHOICE, NOT PANIC

Shorts / Reels Cutdowns

• “The emergency exit” — when retirement stops feeling like a plan and starts looking like the fastest way out of one more workplace rollout.
• “Sometimes they just found the button first” — why interface speed is not the same as understanding the consequences of using a tool.
• “Run the thirty-day test” — name the real problem, test one ordinary task, measure the cleanup as well as the speed, and keep plans for both staying and leaving.

Viewer Question

If a new technology showed up at work tomorrow, would it change when you want to retire—or only change what you need to learn next?