Calvin's Updates

The Latchkey Club Daily Draft — August 14, 2026

Teleprompter / Blog Script

I was reading a survey about AI at work, and one number stopped me.

Almost one out of three people said they had exaggerated their AI skills to coworkers.

I understood it immediately.

Nobody wants to be the person in the meeting who admits they don’t know what the new tool does, especially when everybody else is nodding like they’ve been using it for years.

Welcome back to the channel, guys.

Today I wanted to talk about the pressure to look good at AI before we’ve actually learned how to use it.

This feels especially complicated for people our age. If you’re in your fifties and somebody announces another major technology change at work, you know there may already be assumptions in the room.

Can the older employees keep up? Are they resistant? Will they need extra help? Are they close enough to retirement that training them is even worth it?

So you may be tempted to protect yourself. You use the vocabulary. You mention a tool you tried twice. You say you’re “working with AI” when what you really mean is that you opened it, typed something, got a strange answer, and quietly closed the tab.

But pretending creates a bad kind of pressure. Once you tell everybody you know what you’re doing, it becomes harder to ask the basic question you should have asked at the beginning.

The Ipsos survey was done in Canada, so I wouldn’t treat the percentages as a measurement of every workplace. But the pattern makes sense. Thirty-two percent admitted exaggerating their AI skills. Three quarters rated themselves at a C level or lower. Only 37 percent said their employer had given them enough training.

That’s a strange arrangement.

The expectation is arriving faster than the instruction.

Then we blame ourselves for not already knowing it.

My own path into AI did not begin with a strategy. I was working on a Python script and got stuck. I asked an AI tool how to do one specific thing, copied part of the answer, and tried it.

Then I asked for a little more.

Over time, I realized the tool could help with larger pieces of code. Later, coding agents became capable of building much more complete systems. But I didn’t wake up one morning as an AI expert. I kept bringing it real problems and learning where it helped, where it failed, and where I had given it poor instructions.

I still get things wrong. Tools change. Models change. The thing that worked last month gets renamed, moved, or placed behind another subscription. Sometimes the AI gives me a confident answer that falls apart as soon as I test it.

I think people over 55 may need a different definition of AI skill.

It probably isn’t knowing every product name. It isn’t memorizing a collection of clever prompts. And it definitely isn’t being able to speak for ten minutes in a meeting without admitting uncertainty.

AI skill at work may be much simpler.

Can you identify a problem worth solving?

Can you give the tool enough context to attempt it?

Can you tell whether the output fits the real situation?

Can you check the important parts before another person carries the consequence?

And can you improve the process the second time instead of repeating the same mistake?

A person with thirty years in a field may already have four of those five abilities. The unfamiliar part is the interface.

That matters because the internet makes AI learning look like another subject you have to master before you’re allowed to begin. There are courses, certifications, prompt libraries, daily newsletters, and videos telling you about hundreds of tools.

Some of that is helpful. Some of it is just another way to feel behind.

Current YouTube results tell you something about the demand. A beginner introduction to AI made specifically for seniors has more than two hundred thousand views. Another creator over 50 recently started a series by simply saying she is learning and bringing people along.

I like that approach.

There is no reason to pretend the beginner stage happened off camera.

And at work, I don’t think the best first step is “learn AI.” That is too large to be useful.

Start with one annoyance you understand.

Maybe you write the same summary every Friday. Maybe information is scattered across three documents. Maybe new employees keep asking the same question because the instructions begin halfway through the process. Maybe you need to compare a long vendor proposal with requirements you already know.

Use approved tools. Remove private or sensitive information. Then try the task on something low risk.

Did the tool save time?

Did it leave out something an experienced person would notice?

Did checking the answer take longer than doing the work yourself?

Would you trust this process with a customer, your money, or somebody’s health?

Those answers teach you more than pretending you understand a slide about AI transformation.

They also give you something honest to say in the next meeting.

“I tested it on this task. It handled the first draft well, but it missed these two exceptions. Here’s where I think we could use it safely.”

That is a much stronger contribution than acting fluent.

It also points to the hidden advantage of being older at work. We have seen enough software rollouts to know that the demonstration is not the daily job. We know the clean example in the presentation will eventually meet a customer with an unusual account, an old system nobody mentioned, or a deadline that changes the tradeoff.

Again, age doesn’t make us automatically right. Experience can harden into habit if we never question it.

But experience gives us test cases.

We know which failure is merely annoying and which one will have somebody calling at six in the morning. We know where the policy and the actual work stopped matching years ago. We may know that a beautifully written answer is still useless because it doesn’t fit how the job gets done.

That is AI skill too.

The younger person who knows the new tool and the older person who knows the consequences should not be trying to embarrass each other. Put them together. Let one person move quickly and let the other ask what happens after the demo.

Everybody learns faster when nobody has to perform expertise they do not have.

I also think managers have some responsibility here. If a company says AI is now essential but gives people no protected time, no approved tools, no examples tied to actual jobs, and no safe place to admit confusion, the company is teaching employees to bluff.

A training link is not the same as practice.

People need a small task, a boundary, somebody to ask, and permission to say the result was poor.

That last part matters. Failed tests are part of learning. Hidden failures become operational problems.

At this age, I don’t want to spend energy trying to look current. I would rather become useful with the tool in front of me.

That means I can say, “I don’t know this one yet.”

Then I need to take the next step and learn enough to test it. Humility without effort is just avoidance. But effort without humility can become a very polished mistake.

So if AI is making you feel behind, pick one real problem this week. Keep it small and low risk. Ask somebody for help if you need it. Check the result against what your experience tells you.

You do not need to become the AI person in the office by Friday.

You need one honest piece of evidence about where the tool helps and where it does not.

That is how confidence becomes real.

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

Have you ever felt pressure to pretend you understood a new tool at work? I’d be curious what helped you move from nodding along to actually using it.

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 did you feel when you saw that nearly one in three survey respondents admitted exaggerating their AI skills?
   • Follow-up: Have you ever nodded through a technology conversation because you did not want to look behind?
2. The Gen X pressure: What assumptions might an employee in their fifties feel when a company introduces another major tool?
   • Follow-up: Why can that pressure make basic questions harder to ask?
3. Your real starting point: How did one specific Python problem lead you gradually into larger AI-assisted projects?
   • Follow-up: What did you learn through testing that a general AI course could not have taught you?
4. Redefine AI skill: Which matters more at work: knowing every tool name or recognizing a real problem, judging the output, and checking the consequences?
   • Follow-up: Which of those abilities may an experienced worker already have?
5. One-task learning: What small, low-risk annoyance could someone test with an approved AI tool this week?
   • Follow-up: What information should stay out of the tool?
6. Useful evidence: How can someone report honestly that AI handled a first draft but missed important exceptions?
   • Follow-up: Why is that more valuable than sounding fluent in a meeting?
7. Cross-generational partnership: What happens when a tool-fluent younger worker and a consequence-aware older worker test the same workflow together?
8. Manager responsibility: What should employers provide besides a training link if they expect workers to use AI responsibly?
9. Closing: What is one honest sentence you can say when you do not know a tool yet?
   • Follow-up: How do you pair humility with the responsibility to keep learning?

Title Options

1. You Don’t Have to Pretend You’re Good at AI
2. Gen X, Stop Bluffing Your Way Through AI at Work
3. The Best Way to Learn AI After 50 Is One Real Problem

Thumbnail / Onscreen Text Options

• STOP PRETENDING YOU KNOW AI
• IT’S OKAY TO BE NEW AT THIS
• ONE REAL PROBLEM

Shorts / Reels Cutdowns

• The expectation arrived before the training: Use the Ipsos findings and the pressure older workers may feel to nod along rather than ask a basic question.
• A better definition of AI skill: Cut the question cluster about identifying a worthwhile problem, supplying context, checking the result, and owning the consequence.
• Bring one real problem: Use the practical workflow: choose a low-risk annoyance, protect sensitive information, run the test, and report honestly what the tool missed.

Viewer Question

Have you ever felt pressure to pretend you understood a new tool at work, and what helped you begin using it for real?